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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1249 - (show annotations)
Mon Feb 18 09:55:43 2013 UTC (6 years, 9 months ago) by zherczeg
File size: 242333 byte(s)
Error occurred while calculating annotation data.
Inlining subpatterns in recursions.
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 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
20 to unsigned int is reported to make a quite noticeable speed difference in
21 a specific Windows environment. Testing on Linux did also appear to show
22 some benefit (and it is clearly not harmful). Also fixed the definition of
23 Xop which should be unsigned.
24
25 5. Related to (4), changing the definition of the intermediate variable cc
26 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
27 improvements.
28
29 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
30 unnecessary spaces.
31
32 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
33 and later.
34
35 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
36
37 9. Optimizing fast_forward_start_bits in JIT.
38
39 10. Adding experimental support for callouts in JIT, and fixing some
40 issues revealed during this work. Namely:
41
42 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
43
44 (b) Minimum length was not checked before the matching is started.
45
46 11. The value of capture_last that is passed to callouts was incorrect in some
47 cases when there was a capture on one path that was subsequently abandoned
48 after a backtrack. Also, the capture_last value is now reset after a
49 recursion, since all captures are also reset in this case.
50
51 12. The interpreter no longer returns the "too many substrings" error in the
52 case when an overflowing capture is in a branch that is subsequently
53 abandoned after a backtrack.
54
55 13. In the pathological case when an offset vector of size 2 is used, pcretest
56 now prints out the matched string after a yield of 0 or 1.
57
58 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
59 Only supported by the JIT compiler at the moment.
60
61
62 Version 8.32 30-November-2012
63 -----------------------------
64
65 1. Improved JIT compiler optimizations for first character search and single
66 character iterators.
67
68 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
69 Patch by Daniel Richard G.
70
71 3. Single character iterator optimizations in the JIT compiler.
72
73 4. Improved JIT compiler optimizations for character ranges.
74
75 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
76 Reported by Giuseppe D'Angelo.
77
78 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
79 of a line, was being set incorrectly in cases where .* appeared inside
80 atomic brackets at the start of a pattern, or where there was a subsequent
81 *PRUNE or *SKIP.
82
83 7. Improved instruction cache flush for POWER/PowerPC.
84 Patch by Daniel Richard G.
85
86 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
87 grep:
88
89 (a) There is now no limit to the number of patterns to be matched.
90
91 (b) An error is given if a pattern is too long.
92
93 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
94 are now supported.
95
96 (d) --exclude-from and --include-from (multiple use) have been added.
97
98 (e) Exclusions and inclusions now apply to all files and directories, not
99 just to those obtained from scanning a directory recursively.
100
101 (f) Multiple uses of -f and --file-list are now supported.
102
103 (g) In a Windows environment, the default for -d has been changed from
104 "read" (the GNU grep default) to "skip", because otherwise the presence
105 of a directory in the file list provokes an error.
106
107 (h) The documentation has been revised and clarified in places.
108
109 9. Improve the matching speed of capturing brackets.
110
111 10. Changed the meaning of \X so that it now matches a Unicode extended
112 grapheme cluster.
113
114 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
115 out POSIX threads when JIT support is configured.
116
117 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
118
119 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
120 structure could go wrong in environments where size_t is not the same size
121 as int.
122
123 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
124 set.
125
126 15. The EBCDIC support had decayed; later updates to the code had included
127 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
128 general tidy up of EBCDIC-related issues, and the documentation was also
129 not quite right. There is now a test that can be run on ASCII systems to
130 check some of the EBCDIC-related things (but is it not a full test).
131
132 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
133 in a small tidy to the code.
134
135 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
136
137 18. If the --only-matching (-o) option in pcregrep is specified multiple
138 times, each one causes appropriate output. For example, -o1 -o2 outputs the
139 substrings matched by the 1st and 2nd capturing parentheses. A separating
140 string can be specified by --om-separator (default empty).
141
142 19. Improving the first n character searches.
143
144 20. Turn case lists for horizontal and vertical white space into macros so that
145 they are defined only once.
146
147 21. This set of changes together give more compatible Unicode case-folding
148 behaviour for characters that have more than one other case when UCP
149 support is available.
150
151 (a) The Unicode property table now has offsets into a new table of sets of
152 three or more characters that are case-equivalent. The MultiStage2.py
153 script that generates these tables (the pcre_ucd.c file) now scans
154 CaseFolding.txt instead of UnicodeData.txt for character case
155 information.
156
157 (b) The code for adding characters or ranges of characters to a character
158 class has been abstracted into a generalized function that also handles
159 case-independence. In UTF-mode with UCP support, this uses the new data
160 to handle characters with more than one other case.
161
162 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
163 whose other case is greater than 256 are now correctly matched
164 caselessly. Previously, the high codepoint matched the low one, but not
165 vice versa.
166
167 (d) The processing of \h, \H, \v, and \ in character classes now makes use
168 of the new class addition function, using character lists defined as
169 macros alongside the case definitions of 20 above.
170
171 (e) Caseless back references now work with characters that have more than
172 one other case.
173
174 (f) General caseless matching of characters with more than one other case
175 is supported.
176
177 22. Unicode character properties were updated from Unicode 6.2.0
178
179 23. Improved CMake support under Windows. Patch by Daniel Richard G.
180
181 24. Add support for 32-bit character strings, and UTF-32
182
183 25. Major JIT compiler update (code refactoring and bugfixing).
184 Experimental Sparc 32 support is added.
185
186 26. Applied a modified version of Daniel Richard G's patch to create
187 pcre.h.generic and config.h.generic by "make" instead of in the
188 PrepareRelease script.
189
190 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
191 pcre_compile.c when checking for a zero character.
192
193 28. Introducing a native interface for JIT. Through this interface, the compiled
194 machine code can be directly executed. The purpose of this interface is to
195 provide fast pattern matching, so several sanity checks are not performed.
196 However, feature tests are still performed. The new interface provides
197 1.4x speedup compared to the old one.
198
199 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
200 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
201 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
202
203 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
204 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
205 the "old" RFC 2279). Instead, it ended up passing a negative length to
206 pcre_exec().
207
208 31. Add support for GCC's visibility feature to hide internal functions.
209
210 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
211 "unknown -C option" after outputting 0 or 1.
212
213 33. There is now support for generating a code coverage report for the test
214 suite in environments where gcc is the compiler and lcov is installed. This
215 is mainly for the benefit of the developers.
216
217 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
218 unaddressable using valgrind annotations, allowing valgrind to detect
219 invalid memory accesses. This is mainly for the benefit of the developers.
220
221 25. (*UTF) can now be used to start a pattern in any of the three libraries.
222
223 26. Give configure error if --enable-cpp but no C++ compiler found.
224
225
226 Version 8.31 06-July-2012
227 -------------------------
228
229 1. Fixing a wrong JIT test case and some compiler warnings.
230
231 2. Removed a bashism from the RunTest script.
232
233 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
234 to unsigned type, result still unsigned" that was given by an MS compiler
235 on encountering the code "-sizeof(xxx)".
236
237 4. Partial matching support is added to the JIT compiler.
238
239 5. Fixed several bugs concerned with partial matching of items that consist
240 of more than one character:
241
242 (a) /^(..)\1/ did not partially match "aba" because checking references was
243 done on an "all or nothing" basis. This also applied to repeated
244 references.
245
246 (b) \R did not give a hard partial match if \r was found at the end of the
247 subject.
248
249 (c) \X did not give a hard partial match after matching one or more
250 characters at the end of the subject.
251
252 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
253 a partial match for the string "\r".
254
255 (e) When newline was set to CRLF, the metacharacter "." did not recognize
256 a partial match for a CR character at the end of the subject string.
257
258 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
259 running pcretest, the text "(JIT)" added to the output whenever JIT is
260 actually used to run the match.
261
262 7. Individual JIT compile options can be set in pcretest by following -s+[+]
263 or /S+[+] with a digit between 1 and 7.
264
265 8. OP_NOT now supports any UTF character not just single-byte ones.
266
267 9. (*MARK) control verb is now supported by the JIT compiler.
268
269 10. The command "./RunTest list" lists the available tests without actually
270 running any of them. (Because I keep forgetting what they all are.)
271
272 11. Add PCRE_INFO_MAXLOOKBEHIND.
273
274 12. Applied a (slightly modified) user-supplied patch that improves performance
275 when the heap is used for recursion (compiled with --disable-stack-for-
276 recursion). Instead of malloc and free for each heap frame each time a
277 logical recursion happens, frames are retained on a chain and re-used where
278 possible. This sometimes gives as much as 30% improvement.
279
280 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
281 call.
282
283 14. As documented, (*COMMIT) is now confined to within a positive assertion.
284
285 15. It is now possible to link pcretest with libedit as an alternative to
286 libreadline.
287
288 16. (*COMMIT) control verb is now supported by the JIT compiler.
289
290 17. The Unicode data tables have been updated to Unicode 6.1.0.
291
292 18. Added --file-list option to pcregrep.
293
294 19. Added binary file support to pcregrep, including the -a, --binary-files,
295 -I, and --text options.
296
297 20. The madvise function is renamed for posix_madvise for QNX compatibility
298 reasons. Fixed by Giuseppe D'Angelo.
299
300 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
301
302 22. Changed the option for creating symbolic links for 16-bit man pages from
303 -s to -sf so that re-installing does not cause issues.
304
305 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
306
307 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
308 matches in certain environments (the workspace was not being correctly
309 retained). Also added to pcre_dfa_exec() a simple plausibility check on
310 some of the workspace data at the beginning of a restart.
311
312 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
313 was not doing so when it should - probably a typo introduced by SVN 528
314 (change 8.10/14).
315
316 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
317 \w+ when the character tables indicated that \x{c4} was a word character.
318 There were several related cases, all because the tests for doing a table
319 lookup were testing for characters less than 127 instead of 255.
320
321 27. If a pattern contains capturing parentheses that are not used in a match,
322 their slots in the ovector are set to -1. For those that are higher than
323 any matched groups, this happens at the end of processing. In the case when
324 there were back references that the ovector was too small to contain
325 (causing temporary malloc'd memory to be used during matching), and the
326 highest capturing number was not used, memory off the end of the ovector
327 was incorrectly being set to -1. (It was using the size of the temporary
328 memory instead of the true size.)
329
330 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
331 ovector size, it uses memory at the end of the block that it has got.
332
333 29. Check for an overlong MARK name and give an error at compile time. The
334 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
335
336 30. JIT compiler update.
337
338 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
339 Rill for the patch.
340
341 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
342
343 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
344
345 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
346 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
347 was enabled.
348
349 35. Improve JIT code generation for greedy plus quantifier.
350
351 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
352 1 to "aa" instead of to an empty string. The bug affected repeated groups
353 that could potentially match an empty string.
354
355 37. Optimizing single character iterators in JIT.
356
357 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
358 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
359 codepoints that are too big for the mode are faulted, and in a UTF mode,
360 disallowed codepoints are also faulted.
361
362 39. If PCRE was compiled with UTF support, in three places in the DFA
363 matcher there was code that should only have been obeyed in UTF mode, but
364 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
365 processing when bytes with values greater than 127 were present. In 16-bit
366 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
367 both cases the values are those that cannot be the first data item in a UTF
368 character. The three items that might have provoked this were recursions,
369 possessively repeated groups, and atomic groups.
370
371 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
372 and pcregrep, because some OS require shared objects to be explicitly
373 passed to ld, causing the link step to fail if they are not.
374
375 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
376 mode, patterns that started with \h* or \R* might be incorrectly matched.
377
378
379 Version 8.30 04-February-2012
380 -----------------------------
381
382 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
383 name is defined in ctype.h.
384
385 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
386 only in quite long subpatterns.
387
388 3. Removed the function pcre_info(), which has been obsolete and deprecated
389 since it was replaced by pcre_fullinfo() in February 2000.
390
391 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
392 match a (*MARK), and the match failed at the start of the subject, a
393 reference to memory before the start of the subject could occur. This bug
394 was introduced by fix 17 of release 8.21.
395
396 5. A reference to an unset group with zero minimum repetition was giving
397 totally wrong answers (in non-JavaScript-compatibility mode). For example,
398 /(another)?(\1?)test/ matched against "hello world test". This bug was
399 introduced in release 8.13.
400
401 6. Add support for 16-bit character strings (a large amount of work involving
402 many changes and refactorings).
403
404 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
405 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
406 from a file.
407
408 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
409 rounding is not applied in this particular case).
410
411 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
412 if they appear, or are escaped, in patterns.
413
414 10. Get rid of a number of -Wunused-but-set-variable warnings.
415
416 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
417 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
418 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
419 also returns the mark "x". This bug applied to capturing parentheses,
420 non-capturing parentheses, and atomic parentheses. It also applied to some
421 assertions.
422
423 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
424 information out of configure.ac instead of relying on pcre.h.generic, which
425 is not stored in the repository.
426
427 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
428 -lreadline.
429
430 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
431
432 15. Applied Graycode's patch to put the top-level frame on the stack rather
433 than the heap when not using the stack for recursion. This gives a
434 performance improvement in many cases when recursion is not deep.
435
436 16. Experimental code added to "pcretest -C" to output the stack frame size.
437
438
439 Version 8.21 12-Dec-2011
440 ------------------------
441
442 1. Updating the JIT compiler.
443
444 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
445 are added as well.
446
447 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
448 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
449 calling _pcre_jit_exec. Some extra comments are added.
450
451 4. (*MARK) settings inside atomic groups that do not contain any capturing
452 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
453 was introduced by change 18 for 8.20.
454
455 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
456 ECMA-262 standard.
457
458 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
459 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
460 This bug was probably introduced by change 9 of 8.13.
461
462 7. While fixing 6 above, I noticed that a number of other items were being
463 incorrectly rejected as "not fixed length". This arose partly because newer
464 opcodes had not been added to the fixed-length checking code. I have (a)
465 corrected the bug and added tests for these items, and (b) arranged for an
466 error to occur if an unknown opcode is encountered while checking for fixed
467 length instead of just assuming "not fixed length". The items that were
468 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
469 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
470 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
471
472 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
473 being incorrectly compiled and would have given unpredicatble results.
474
475 9. A possessively repeated subpattern with minimum repeat count greater than
476 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
477 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
478 the first (A) could occur when it should not.
479
480 10. Add a cast and remove a redundant test from the code.
481
482 11. JIT should use pcre_malloc/pcre_free for allocation.
483
484 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
485 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
486 is anything other than /usr, -L is still shown).
487
488 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
489
490 14. Perl does not support \N without a following name in a [] class; PCRE now
491 also gives an error.
492
493 15. If a forward reference was repeated with an upper limit of around 2000,
494 it caused the error "internal error: overran compiling workspace". The
495 maximum number of forward references (including repeats) was limited by the
496 internal workspace, and dependent on the LINK_SIZE. The code has been
497 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
498 the default depends on LINK_SIZE. There is a new upper limit (for safety)
499 of around 200,000 forward references. While doing this, I also speeded up
500 the filling in of repeated forward references.
501
502 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
503 incorrectly expecting the subject to contain another "a" after the start.
504
505 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
506 in the match, the SKIP should be ignored. This was not happening; instead
507 the SKIP was being treated as NOMATCH. For patterns such as
508 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
509 tested.
510
511 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
512 now much more compatible with Perl, in particular in cases where the result
513 is a non-match for a non-anchored pattern. For example, if
514 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
515 "m", where previously it did not return a name. A side effect of this
516 change is that for partial matches, the last encountered mark name is
517 returned, as for non matches. A number of tests that were previously not
518 Perl-compatible have been moved into the Perl-compatible test files. The
519 refactoring has had the pleasing side effect of removing one argument from
520 the match() function, thus reducing its stack requirements.
521
522 19. If the /S+ option was used in pcretest to study a pattern using JIT,
523 subsequent uses of /S (without +) incorrectly behaved like /S+.
524
525 21. Retrieve executable code size support for the JIT compiler and fixing
526 some warnings.
527
528 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
529 not work when the shorter character appeared right at the end of the
530 subject string.
531
532 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
533 systems.
534
535 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
536 output it when the /M option is used in pcretest.
537
538 25. The CheckMan script was not being included in the distribution. Also, added
539 an explicit "perl" to run Perl scripts from the PrepareRelease script
540 because this is reportedly needed in Windows.
541
542 26. If study data was being save in a file and studying had not found a set of
543 "starts with" bytes for the pattern, the data written to the file (though
544 never used) was taken from uninitialized memory and so caused valgrind to
545 complain.
546
547 27. Updated RunTest.bat as provided by Sheri Pierce.
548
549 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
550
551 29. Computation of memory usage for the table of capturing group names was
552 giving an unnecessarily large value.
553
554
555 Version 8.20 21-Oct-2011
556 ------------------------
557
558 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
559 a POSIX class. After further experiments with Perl, which convinced me that
560 Perl has bugs and confusions, a closing square bracket is no longer allowed
561 in a POSIX name. This bug also affected patterns with classes that started
562 with full stops.
563
564 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
565 captured substring, but while checking the failing first alternative,
566 substring 1 is temporarily captured. If the output vector supplied to
567 pcre_exec() was not big enough for this capture, the yield of the function
568 was still zero ("insufficient space for captured substrings"). This cannot
569 be totally fixed without adding another stack variable, which seems a lot
570 of expense for a edge case. However, I have improved the situation in cases
571 such as /(a)(b)x|abc/ matched against "abc", where the return code
572 indicates that fewer than the maximum number of slots in the ovector have
573 been set.
574
575 3. Related to (2) above: when there are more back references in a pattern than
576 slots in the output vector, pcre_exec() uses temporary memory during
577 matching, and copies in the captures as far as possible afterwards. It was
578 using the entire output vector, but this conflicts with the specification
579 that only 2/3 is used for passing back captured substrings. Now it uses
580 only the first 2/3, for compatibility. This is, of course, another edge
581 case.
582
583 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
584 main code base, and can be used by building with --enable-jit. When this is
585 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
586 runtime --no-jit option is given.
587
588 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
589 ovector, the return from the function was zero, implying that there were
590 other matches that did not fit. The correct "exactly full" value is now
591 returned.
592
593 6. If a subpattern that was called recursively or as a subroutine contained
594 (*PRUNE) or any other control that caused it to give a non-standard return,
595 invalid errors such as "Error -26 (nested recursion at the same subject
596 position)" or even infinite loops could occur.
597
598 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
599 computing the minimum length on reaching *ACCEPT, and so ended up with the
600 wrong value of 1 rather than 0. Further investigation indicates that
601 computing a minimum subject length in the presence of *ACCEPT is difficult
602 (think back references, subroutine calls), and so I have changed the code
603 so that no minimum is registered for a pattern that contains *ACCEPT.
604
605 8. If (*THEN) was present in the first (true) branch of a conditional group,
606 it was not handled as intended. [But see 16 below.]
607
608 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
609 Sheri Pierce.
610
611 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
612 the first byte in a match must be "a".
613
614 11. Change 17 for 8.13 increased the recursion depth for patterns like
615 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
616 pattern contains any instances of (*THEN). If it does not, the old
617 optimizations are restored. It would be nice to do this on a per-group
618 basis, but at the moment that is not feasible.
619
620 12. In some environments, the output of pcretest -C is CRLF terminated. This
621 broke RunTest's code that checks for the link size. A single white space
622 character after the value is now allowed for.
623
624 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
625 For "fr", it uses the Windows-specific input and output files.
626
627 14. If (*THEN) appeared in a group that was called recursively or as a
628 subroutine, it did not work as intended. [But see next item.]
629
630 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
631 pattern fragments (but not containing any | characters). If A and B are
632 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
633 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
634 D. In other words, Perl considers parentheses that do not contain any |
635 characters to be part of a surrounding alternative, whereas PCRE was
636 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
637 differently. PCRE now behaves in the same way as Perl, except in the case
638 of subroutine/recursion calls such as (?1) which have in any case always
639 been different (but PCRE had them first :-).
640
641 16. Related to 15 above: Perl does not treat the | in a conditional group as
642 creating alternatives. Such a group is treated in the same way as an
643 ordinary group without any | characters when processing (*THEN). PCRE has
644 been changed to match Perl's behaviour.
645
646 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
647 RunGrepTest script failed.
648
649 18. Change 22 for version 13 caused atomic groups to use more stack. This is
650 inevitable for groups that contain captures, but it can lead to a lot of
651 stack use in large patterns. The old behaviour has been restored for atomic
652 groups that do not contain any capturing parentheses.
653
654 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
655 suppress the check for a minimum subject length at run time. (If it was
656 given to pcre_exec() or pcre_dfa_exec() it did work.)
657
658 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
659 fail to work when decoding hex characters in data strings in EBCDIC
660 environments.
661
662 21. It appears that in at least one Mac OS environment, the isxdigit() function
663 is implemented as a macro that evaluates to its argument more than once,
664 contravening the C 90 Standard (I haven't checked a later standard). There
665 was an instance in pcretest which caused it to go wrong when processing
666 \x{...} escapes in subject strings. The has been rewritten to avoid using
667 things like p++ in the argument of isxdigit().
668
669
670 Version 8.13 16-Aug-2011
671 ------------------------
672
673 1. The Unicode data tables have been updated to Unicode 6.0.0.
674
675 2. Two minor typos in pcre_internal.h have been fixed.
676
677 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
678 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
679 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
680
681 4. There were a number of related bugs in the code for matching backrefences
682 caselessly in UTF-8 mode when codes for the characters concerned were
683 different numbers of bytes. For example, U+023A and U+2C65 are an upper
684 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
685 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
686 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
687 2-byte code at the end of the subject (it thought there wasn't enough data
688 left).
689
690 5. Comprehensive information about what went wrong is now returned by
691 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
692 as the output vector has at least 2 elements. The offset of the start of
693 the failing character and a reason code are placed in the vector.
694
695 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
696 now returned is for the first byte of the failing character, instead of the
697 last byte inspected. This is an incompatible change, but I hope it is small
698 enough not to be a problem. It makes the returned offset consistent with
699 pcre_exec() and pcre_dfa_exec().
700
701 7. pcretest now gives a text phrase as well as the error number when
702 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
703 failure, the offset and reason code are output.
704
705 8. When \R was used with a maximizing quantifier it failed to skip backwards
706 over a \r\n pair if the subsequent match failed. Instead, it just skipped
707 back over a single character (\n). This seems wrong (because it treated the
708 two characters as a single entity when going forwards), conflicts with the
709 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
710 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
711 has been changed.
712
713 9. Some internal refactoring has changed the processing so that the handling
714 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
715 time (the PCRE_DOTALL option was changed this way some time ago: version
716 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
717 which was always a bit of a fudge. It also means that there is one less
718 argument for the match() function, which reduces its stack requirements
719 slightly. This change also fixes an incompatibility with Perl: the pattern
720 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
721
722 10. More internal refactoring has drastically reduced the number of recursive
723 calls to match() for possessively repeated groups such as (abc)++ when
724 using pcre_exec().
725
726 11. While implementing 10, a number of bugs in the handling of groups were
727 discovered and fixed:
728
729 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
730 (a|)*(?1) gave a compile-time internal error.
731 ((a|)+)+ did not notice that the outer group could match an empty string.
732 (^a|^)+ was not marked as anchored.
733 (.*a|.*)+ was not marked as matching at start or after a newline.
734
735 12. Yet more internal refactoring has removed another argument from the match()
736 function. Special calls to this function are now indicated by setting a
737 value in a variable in the "match data" data block.
738
739 13. Be more explicit in pcre_study() instead of relying on "default" for
740 opcodes that mean there is no starting character; this means that when new
741 ones are added and accidentally left out of pcre_study(), testing should
742 pick them up.
743
744 14. The -s option of pcretest has been documented for ages as being an old
745 synonym of -m (show memory usage). I have changed it to mean "force study
746 for every regex", that is, assume /S for every regex. This is similar to -i
747 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
748 using it. It makes it easier to run collections of tests with and without
749 study enabled, and thereby test pcre_study() more easily. All the standard
750 tests are now run with and without -s (but some patterns can be marked as
751 "never study" - see 20 below).
752
753 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
754 restoration of the capturing data to the outer values was not happening
755 correctly.
756
757 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
758 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
759 pattern had matched an empty string, and so incorrectly returned a no
760 match.
761
762 17. There was optimizing code for the last branch of non-capturing parentheses,
763 and also for the obeyed branch of a conditional subexpression, which used
764 tail recursion to cut down on stack usage. Unfortunately, now that there is
765 the possibility of (*THEN) occurring in these branches, tail recursion is
766 no longer possible because the return has to be checked for (*THEN). These
767 two optimizations have therefore been removed. [But see 8.20/11 above.]
768
769 18. If a pattern containing \R was studied, it was assumed that \R always
770 matched two bytes, thus causing the minimum subject length to be
771 incorrectly computed because \R can also match just one byte.
772
773 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
774 was incorrectly computed.
775
776 20. If /S is present twice on a test pattern in pcretest input, it now
777 *disables* studying, thereby overriding the use of -s on the command line
778 (see 14 above). This is necessary for one or two tests to keep the output
779 identical in both cases.
780
781 21. When (*ACCEPT) was used in an assertion that matched an empty string and
782 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
783
784 22. When an atomic group that contained a capturing parenthesis was
785 successfully matched, but the branch in which it appeared failed, the
786 capturing was not being forgotten if a higher numbered group was later
787 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
788 group 1 to "a", when in fact it should be unset. This applied to multi-
789 branched capturing and non-capturing groups, repeated or not, and also to
790 positive assertions (capturing in negative assertions does not happen
791 in PCRE) and also to nested atomic groups.
792
793 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
794 subject after a captured substring, to make it easier to tell which of a
795 number of identical substrings has been captured.
796
797 24. The way atomic groups are processed by pcre_exec() has been changed so that
798 if they are repeated, backtracking one repetition now resets captured
799 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
800 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
801 "aaa". Previously, it would have been "a". As part of this code
802 refactoring, the way recursive calls are handled has also been changed.
803
804 25. If an assertion condition captured any substrings, they were not passed
805 back unless some other capturing happened later. For example, if
806 (?(?=(a))a) was matched against "a", no capturing was returned.
807
808 26. When studying a pattern that contained subroutine calls or assertions,
809 the code for finding the minimum length of a possible match was handling
810 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
811 group 1 called group 2 while simultaneously a separate group 2 called group
812 1). A stack overflow occurred in this case. I have fixed this by limiting
813 the recursion depth to 10.
814
815 27. Updated RunTest.bat in the distribution to the version supplied by Tom
816 Fortmann. This supports explicit test numbers on the command line, and has
817 argument validation and error reporting.
818
819 28. An instance of \X with an unlimited repeat could fail if at any point the
820 first character it looked at was a mark character.
821
822 29. Some minor code refactoring concerning Unicode properties and scripts
823 should reduce the stack requirement of match() slightly.
824
825 30. Added the '=' option to pcretest to check the setting of unused capturing
826 slots at the end of the pattern, which are documented as being -1, but are
827 not included in the return count.
828
829 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
830 compiled something random. Now it gives a compile-time error (as does
831 Perl).
832
833 32. A *MARK encountered during the processing of a positive assertion is now
834 recorded and passed back (compatible with Perl).
835
836 33. If --only-matching or --colour was set on a pcregrep call whose pattern
837 had alternative anchored branches, the search for a second match in a line
838 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
839 matched the line "0102" twice. The same bug affected patterns that started
840 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
841 twice.
842
843 34. Previously, PCRE did not allow quantification of assertions. However, Perl
844 does, and because of capturing effects, quantifying parenthesized
845 assertions may at times be useful. Quantifiers are now allowed for
846 parenthesized assertions.
847
848 35. A minor code tidy in pcre_compile() when checking options for \R usage.
849
850 36. \g was being checked for fancy things in a character class, when it should
851 just be a literal "g".
852
853 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
854 appearance of a nested POSIX class supersedes an apparent external class.
855 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
856 unescaped square brackets may also appear as part of class names. For
857 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
858 more like Perl. (But see 8.20/1 above.)
859
860 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
861 was because it thought it was \N{name}, which is not supported).
862
863 39. Add minix to OS list not supporting the -S option in pcretest.
864
865 40. PCRE tries to detect cases of infinite recursion at compile time, but it
866 cannot analyze patterns in sufficient detail to catch mutual recursions
867 such as ((?1))((?2)). There is now a runtime test that gives an error if a
868 subgroup is called recursively as a subpattern for a second time at the
869 same position in the subject string. In previous releases this might have
870 been caught by the recursion limit, or it might have run out of stack.
871
872 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
873 happen only once. PCRE was, however incorrectly giving a compile time error
874 "recursive call could loop indefinitely" because it cannot analyze the
875 pattern in sufficient detail. The compile time test no longer happens when
876 PCRE is compiling a conditional subpattern, but actual runaway loops are
877 now caught at runtime (see 40 above).
878
879 42. It seems that Perl allows any characters other than a closing parenthesis
880 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
881 has been changed to be the same.
882
883 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
884 as not to get warnings when autogen.sh is called. Also changed
885 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
886
887 44. To help people who use pcregrep to scan files containing exceedingly long
888 lines, the following changes have been made:
889
890 (a) The default value of the buffer size parameter has been increased from
891 8K to 20K. (The actual buffer used is three times this size.)
892
893 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
894 PCRE is built.
895
896 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
897 to be set at run time.
898
899 (d) Numerical values in pcregrep options can be followed by K or M, for
900 example --buffer-size=50K.
901
902 (e) If a line being scanned overflows pcregrep's buffer, an error is now
903 given and the return code is set to 2.
904
905 45. Add a pointer to the latest mark to the callout data block.
906
907 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
908 partial match of an empty string instead of no match. This was specific to
909 the use of ".".
910
911 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
912 complete match instead of a partial match. This bug was dependent on both
913 the PCRE_UTF8 and PCRE_DOTALL options being set.
914
915 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
916 starting byte set, because \b was not being ignored.
917
918
919 Version 8.12 15-Jan-2011
920 ------------------------
921
922 1. Fixed some typos in the markup of the man pages, and wrote a script that
923 checks for such things as part of the documentation building process.
924
925 2. On a big-endian 64-bit system, pcregrep did not correctly process the
926 --match-limit and --recursion-limit options (added for 8.11). In
927 particular, this made one of the standard tests fail. (The integer value
928 went into the wrong half of a long int.)
929
930 3. If the --colour option was given to pcregrep with -v (invert match), it
931 did strange things, either producing crazy output, or crashing. It should,
932 of course, ignore a request for colour when reporting lines that do not
933 match.
934
935 4. Another pcregrep bug caused similar problems if --colour was specified with
936 -M (multiline) and the pattern match finished with a line ending.
937
938 5. In pcregrep, when a pattern that ended with a literal newline sequence was
939 matched in multiline mode, the following line was shown as part of the
940 match. This seems wrong, so I have changed it.
941
942 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
943 the check for further matches in the same line (so they could be coloured)
944 to overrun the end of the current line. If another match was found, it was
945 incorrectly shown (and then shown again when found in the next line).
946
947 7. If pcregrep was compiled under Windows, there was a reference to the
948 function pcregrep_exit() before it was defined. I am assuming this was
949 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
950 reported by a user. I've moved the definition above the reference.
951
952
953 Version 8.11 10-Dec-2010
954 ------------------------
955
956 1. (*THEN) was not working properly if there were untried alternatives prior
957 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
958 backtracked to try for "b" instead of moving to the next alternative branch
959 at the same level (in this case, to look for "c"). The Perl documentation
960 is clear that when (*THEN) is backtracked onto, it goes to the "next
961 alternative in the innermost enclosing group".
962
963 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
964 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
965 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
966 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
967 (*THEN).
968
969 3. If \s appeared in a character class, it removed the VT character from
970 the class, even if it had been included by some previous item, for example
971 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
972 of \s, but is part of the POSIX "space" class.)
973
974 4. A partial match never returns an empty string (because you can always
975 match an empty string at the end of the subject); however the checking for
976 an empty string was starting at the "start of match" point. This has been
977 changed to the "earliest inspected character" point, because the returned
978 data for a partial match starts at this character. This means that, for
979 example, /(?<=abc)def/ gives a partial match for the subject "abc"
980 (previously it gave "no match").
981
982 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
983 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
984 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
985 has an implication that the given string is incomplete (because a partial
986 match is preferred over a full match). For this reason, these items now
987 give a partial match in this situation. [Aside: previously, the one case
988 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
989 match rather than a full match, which was wrong by the old rules, but is
990 now correct.]
991
992 6. There was a bug in the handling of #-introduced comments, recognized when
993 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
994 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
995 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
996 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
997 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
998 places in pcre_compile().
999
1000 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1001 comments when looking ahead for named forward references to subpatterns,
1002 the only newline sequence it recognized was NL. It now handles newlines
1003 according to the set newline convention.
1004
1005 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1006 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1007 cater for a lack of strerror(). These oversights have been fixed.
1008
1009 9. Added --match-limit and --recursion-limit to pcregrep.
1010
1011 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1012
1013 11. When the -o option was used, pcregrep was setting a return code of 1, even
1014 when matches were found, and --line-buffered was not being honoured.
1015
1016 12. Added an optional parentheses number to the -o and --only-matching options
1017 of pcregrep.
1018
1019 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1020 can match an empty string. The code to do it in pcretest and pcredemo
1021 needed fixing:
1022
1023 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1024 only one byte after an empty string match just before CRLF (this case
1025 just got forgotten; "any" and "anycrlf" were OK).
1026
1027 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1028 mode when an empty string match preceded an ASCII character followed by
1029 a non-ASCII character. (The code for advancing by one character rather
1030 than one byte was nonsense.)
1031
1032 (c) The pcredemo.c sample program did not have any code at all to handle
1033 the cases when CRLF is a valid newline sequence.
1034
1035 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1036 as a starting offset was within the subject string. There is now a new
1037 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1038 negative or greater than the length of the string. In order to test this,
1039 pcretest is extended to allow the setting of negative starting offsets.
1040
1041 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1042 starting offset points to the beginning of a UTF-8 character was
1043 unnecessarily clumsy. I tidied it up.
1044
1045 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1046 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1047
1048 17. Nobody had reported that the --include_dir option, which was added in
1049 release 7.7 should have been called --include-dir (hyphen, not underscore)
1050 for compatibility with GNU grep. I have changed it to --include-dir, but
1051 left --include_dir as an undocumented synonym, and the same for
1052 --exclude-dir, though that is not available in GNU grep, at least as of
1053 release 2.5.4.
1054
1055 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1056 characters from a string of bytes) have been redefined so as not to use
1057 loops, in order to improve performance in some environments. At the same
1058 time, I abstracted some of the common code into auxiliary macros to save
1059 repetition (this should not affect the compiled code).
1060
1061 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1062 compile-time error is now given if \c is not followed by an ASCII
1063 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1064 different, and any byte value is allowed.)
1065
1066 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1067 START_OPTIMIZE option, which is now allowed at compile time - but just
1068 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1069 to pcregrep and other applications that have no direct access to PCRE
1070 options. The new /Y option in pcretest sets this option when calling
1071 pcre_compile().
1072
1073 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1074 back references. Groups containing recursive back references were forced to
1075 be atomic by that change, but in the case of named groups, the amount of
1076 memory required was incorrectly computed, leading to "Failed: internal
1077 error: code overflow". This has been fixed.
1078
1079 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1080 pcretest.c, to avoid build problems in some Borland environments.
1081
1082
1083 Version 8.10 25-Jun-2010
1084 ------------------------
1085
1086 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1087 THEN.
1088
1089 2. (*ACCEPT) was not working when inside an atomic group.
1090
1091 3. Inside a character class, \B is treated as a literal by default, but
1092 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1093 causes the error). The code is unchanged, but I tidied the documentation.
1094
1095 4. Inside a character class, PCRE always treated \R and \X as literals,
1096 whereas Perl faults them if its -w option is set. I have changed PCRE so
1097 that it faults them when PCRE_EXTRA is set.
1098
1099 5. Added support for \N, which always matches any character other than
1100 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1101
1102 6. When compiling pcregrep with newer versions of gcc which may have
1103 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1104 declared with attribute warn_unused_result" were given. Just casting the
1105 result to (void) does not stop the warnings; a more elaborate fudge is
1106 needed. I've used a macro to implement this.
1107
1108 7. Minor change to pcretest.c to avoid a compiler warning.
1109
1110 8. Added four artifical Unicode properties to help with an option to make
1111 \s etc use properties (see next item). The new properties are: Xan
1112 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1113
1114 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1115 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1116 this option. Modified pcretest to add /W to test this facility. Added
1117 REG_UCP to make it available via the POSIX interface.
1118
1119 10. Added --line-buffered to pcregrep.
1120
1121 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1122 studied, and the match started with a letter with a code point greater than
1123 127 whose first byte was different to the first byte of the other case of
1124 the letter, the other case of this starting letter was not recognized
1125 (#976).
1126
1127 12. If a pattern that was studied started with a repeated Unicode property
1128 test, for example, \p{Nd}+, there was the theoretical possibility of
1129 setting up an incorrect bitmap of starting bytes, but fortunately it could
1130 not have actually happened in practice until change 8 above was made (it
1131 added property types that matched character-matching opcodes).
1132
1133 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1134 possible starting bytes for non-anchored patterns.
1135
1136 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1137 \R, and also a number of cases that involve Unicode properties, both
1138 explicit and implicit when PCRE_UCP is set.
1139
1140 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1141 input, it could crash or give wrong results if characters with values
1142 greater than 0xc0 were present in the subject string. (Detail: it assumed
1143 UTF-8 input when processing these items.)
1144
1145 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1146 size_t is 64-bit (#991).
1147
1148 17. Added a check for running out of memory when PCRE is compiled with
1149 --disable-stack-for-recursion (#990).
1150
1151 18. If the last data line in a file for pcretest does not have a newline on
1152 the end, a newline was missing in the output.
1153
1154 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1155 less than 128) in its various bitmaps. However, there is a facility for
1156 generating tables according to the current locale when PCRE is compiled. It
1157 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1158 characters, are recognized by isspace() and therefore were getting set in
1159 these tables, and indeed these tables seem to approximate to ISO 8859. This
1160 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1161 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1162 which of course cannot start UTF-8 characters. I have changed the code so
1163 that only real ASCII characters (less than 128) and the correct starting
1164 bytes for UTF-8 encodings are set for characters greater than 127 when in
1165 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1166 altogether.)
1167
1168 20. Added the /T option to pcretest so as to be able to run tests with non-
1169 standard character tables, thus making it possible to include the tests
1170 used for 19 above in the standard set of tests.
1171
1172 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1173 reference to a subpattern the other side of a comment that contains an
1174 opening parenthesis caused either an internal compiling error, or a
1175 reference to the wrong subpattern.
1176
1177
1178 Version 8.02 19-Mar-2010
1179 ------------------------
1180
1181 1. The Unicode data tables have been updated to Unicode 5.2.0.
1182
1183 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1184 configured.
1185
1186 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1187 original author of that file, following a query about its status.
1188
1189 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1190 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1191
1192 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1193 quantifier applied to a forward-referencing subroutine call, could compile
1194 incorrect code or give the error "internal error: previously-checked
1195 referenced subpattern not found".
1196
1197 6. Both MS Visual Studio and Symbian OS have problems with initializing
1198 variables to point to external functions. For these systems, therefore,
1199 pcre_malloc etc. are now initialized to local functions that call the
1200 relevant global functions.
1201
1202 7. There were two entries missing in the vectors called coptable and poptable
1203 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1204 I've fixed the data, and added a kludgy way of testing at compile time that
1205 the lengths are correct (equal to the number of opcodes).
1206
1207 8. Following on from 7, I added a similar kludge to check the length of the
1208 eint vector in pcreposix.c.
1209
1210 9. Error texts for pcre_compile() are held as one long string to avoid too
1211 much relocation at load time. To find a text, the string is searched,
1212 counting zeros. There was no check for running off the end of the string,
1213 which could happen if a new error number was added without updating the
1214 string.
1215
1216 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1217
1218 11. \K was not working if it appeared in an atomic group or in a group that
1219 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1220 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1221 the assertion is positive, but not if it is negative.
1222
1223 12. Change 11 fortuitously reduced the size of the stack frame used in the
1224 "match()" function of pcre_exec.c by one pointer. Forthcoming
1225 implementation of support for (*MARK) will need an extra pointer on the
1226 stack; I have reserved it now, so that the stack frame size does not
1227 decrease.
1228
1229 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1230 item in branch that calls a recursion is a subroutine call - as in the
1231 second branch in the above example - was incorrectly given the compile-
1232 time error "recursive call could loop indefinitely" because pcre_compile()
1233 was not correctly checking the subroutine for matching a non-empty string.
1234
1235 14. The checks for overrunning compiling workspace could trigger after an
1236 overrun had occurred. This is a "should never occur" error, but it can be
1237 triggered by pathological patterns such as hundreds of nested parentheses.
1238 The checks now trigger 100 bytes before the end of the workspace.
1239
1240 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1241
1242
1243 Version 8.01 19-Jan-2010
1244 ------------------------
1245
1246 1. If a pattern contained a conditional subpattern with only one branch (in
1247 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1248 computed the wrong minimum data length (which is of course zero for such
1249 subpatterns). This could cause incorrect "no match" results.
1250
1251 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1252 the pattern is reset in the first branch, pcre_compile() failed with
1253 "internal error: code overflow at offset...". This happened only when
1254 the reset was to the original external option setting. (An optimization
1255 abstracts leading options settings into an external setting, which was the
1256 cause of this.)
1257
1258 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1259 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1260 assertion pattern did not match (meaning that the assertion was true), it
1261 was incorrectly treated as false if the SKIP had been reached during the
1262 matching. This also applied to assertions used as conditions.
1263
1264 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1265 assertion subpattern, including such a pattern used as a condition,
1266 unpredictable results occurred, instead of the error return
1267 PCRE_ERROR_DFA_UITEM.
1268
1269 5. The C++ GlobalReplace function was not working like Perl for the special
1270 situation when an empty string is matched. It now does the fancy magic
1271 stuff that is necessary.
1272
1273 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1274 removed. (These were left over from very, very early versions of PCRE.)
1275
1276 7. Some cosmetic changes to the code to make life easier when compiling it
1277 as part of something else:
1278
1279 (a) Change DEBUG to PCRE_DEBUG.
1280
1281 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1282 called "current" as "current_branch", to prevent a collision with the
1283 Linux macro when compiled as a kernel module.
1284
1285 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1286 prevent a collision with the Linux macro when compiled as a kernel
1287 module.
1288
1289 8. In pcre_compile() there are some checks for integer overflows that used to
1290 cast potentially large values to (double). This has been changed to that
1291 when building, a check for int64_t is made, and if it is found, it is used
1292 instead, thus avoiding the use of floating point arithmetic. (There is no
1293 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1294 double.
1295
1296 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1297 2005 (difference between two addresses compared to an unsigned value).
1298
1299 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1300 custom one, because of the following reported problem in Windows:
1301
1302 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1303 under Win32.
1304 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1305 therefore missing the function definition.
1306 - The compiler thus generates a "C" signature for the test function.
1307 - The linker fails to find the "C" function.
1308 - PCRE fails to configure if asked to do so against libbz2.
1309
1310 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1311 messages were output:
1312
1313 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1314 rerunning libtoolize, to keep the correct libtool macros in-tree.
1315 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1316
1317 I have done both of these things.
1318
1319 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1320 most of the time, it *can* run out if it is given a pattern that contains a
1321 runaway infinite recursion. I updated the discussion in the pcrestack man
1322 page.
1323
1324 13. Now that we have gone to the x.xx style of version numbers, the minor
1325 version may start with zero. Using 08 or 09 is a bad idea because users
1326 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1327 interpreted as invalid octal numbers. I've updated the previous comment in
1328 configure.ac, and also added a check that gives an error if 08 or 09 are
1329 used.
1330
1331 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1332 causing partial matching to fail when the end of the subject matched \W
1333 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1334
1335 15. There were some discrepancies between the declarations in pcre_internal.h
1336 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1337 their definitions. The declarations used "const uschar *" and the
1338 definitions used USPTR. Even though USPTR is normally defined as "const
1339 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1340 reported that: "This difference in casting confuses some C++ compilers, for
1341 example, SunCC recognizes above declarations as different functions and
1342 generates broken code for hbpcre." I have changed the declarations to use
1343 USPTR.
1344
1345 16. GNU libtool is named differently on some systems. The autogen.sh script now
1346 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1347 (FreeBSD).
1348
1349 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1350 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1351 comment: "Figure out how to create a longlong from a string: strtoll and
1352 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1353 instance, but it only takes 2 args instead of 3!"
1354
1355 18. A subtle bug concerned with back references has been fixed by a change of
1356 specification, with a corresponding code fix. A pattern such as
1357 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1358 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1359 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1360 same bug. Such groups have to be quantified to be useful, or contained
1361 inside another quantified group. (If there's no repetition, the reference
1362 can never match.) The problem arises because, having left the group and
1363 moved on to the rest of the pattern, a later failure that backtracks into
1364 the group uses the captured value from the final iteration of the group
1365 rather than the correct earlier one. I have fixed this in PCRE by forcing
1366 any group that contains a reference to itself to be an atomic group; that
1367 is, there cannot be any backtracking into it once it has completed. This is
1368 similar to recursive and subroutine calls.
1369
1370
1371 Version 8.00 19-Oct-09
1372 ----------------------
1373
1374 1. The table for translating pcre_compile() error codes into POSIX error codes
1375 was out-of-date, and there was no check on the pcre_compile() error code
1376 being within the table. This could lead to an OK return being given in
1377 error.
1378
1379 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1380 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1381 in a Windows environment.
1382
1383 3. The pcregrep --count option prints the count for each file even when it is
1384 zero, as does GNU grep. However, pcregrep was also printing all files when
1385 --files-with-matches was added. Now, when both options are given, it prints
1386 counts only for those files that have at least one match. (GNU grep just
1387 prints the file name in this circumstance, but including the count seems
1388 more useful - otherwise, why use --count?) Also ensured that the
1389 combination -clh just lists non-zero counts, with no names.
1390
1391 4. The long form of the pcregrep -F option was incorrectly implemented as
1392 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1393 but it seems right to fix it, and I didn't think it was worth preserving
1394 the old behaviour.
1395
1396 5. The command line items --regex=pattern and --regexp=pattern were not
1397 recognized by pcregrep, which required --regex pattern or --regexp pattern
1398 (with a space rather than an '='). The man page documented the '=' forms,
1399 which are compatible with GNU grep; these now work.
1400
1401 6. No libpcreposix.pc file was created for pkg-config; there was just
1402 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1403
1404 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1405 when UCP support is not needed, by modifying the Python script that
1406 generates it from Unicode data files. This should not matter if the module
1407 is correctly used as a library, but I received one complaint about 50K of
1408 unwanted data. My guess is that the person linked everything into his
1409 program rather than using a library. Anyway, it does no harm.
1410
1411 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1412 was a minimum greater than 1 for a wide character in a possessive
1413 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1414 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1415 character. Chaos in the form of incorrect output or a compiling loop could
1416 result.
1417
1418 9. The restrictions on what a pattern can contain when partial matching is
1419 requested for pcre_exec() have been removed. All patterns can now be
1420 partially matched by this function. In addition, if there are at least two
1421 slots in the offset vector, the offset of the earliest inspected character
1422 for the match and the offset of the end of the subject are set in them when
1423 PCRE_ERROR_PARTIAL is returned.
1424
1425 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1426 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1427 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1428 and may be more useful for multi-segment matching.
1429
1430 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1431 used to be given if ever the end of the subject was reached; now it is
1432 given only if matching could not proceed because another character was
1433 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1434 string "Z", which now yields "no match" instead of "partial match". In the
1435 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1436 final character ended with (*FAIL).
1437
1438 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1439 if the pattern had a "must contain" character that was already found in the
1440 earlier partial match, unless partial matching was again requested. For
1441 example, with the pattern /dog.(body)?/, the "must contain" character is
1442 "g". If the first part-match was for the string "dog", restarting with
1443 "sbody" failed. This bug has been fixed.
1444
1445 13. The string returned by pcre_dfa_exec() after a partial match has been
1446 changed so that it starts at the first inspected character rather than the
1447 first character of the match. This makes a difference only if the pattern
1448 starts with a lookbehind assertion or \b or \B (\K is not supported by
1449 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1450 matching functions compatible, and I think it's the right thing to do.
1451
1452 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1453 so that the demonstration program is easily available in environments where
1454 PCRE has not been installed from source.
1455
1456 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1457 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1458 library.
1459
1460 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1461 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1462 is not the first non-POSIX option to be added. Clearly some people find
1463 these options useful.
1464
1465 17. If a caller to the POSIX matching function regexec() passes a non-zero
1466 value for nmatch with a NULL value for pmatch, the value of
1467 nmatch is forced to zero.
1468
1469 18. RunGrepTest did not have a test for the availability of the -u option of
1470 the diff command, as RunTest does. It now checks in the same way as
1471 RunTest, and also checks for the -b option.
1472
1473 19. If an odd number of negated classes containing just a single character
1474 interposed, within parentheses, between a forward reference to a named
1475 subpattern and the definition of the subpattern, compilation crashed with
1476 an internal error, complaining that it could not find the referenced
1477 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1478 [The bug was that it was starting one character too far in when skipping
1479 over the character class, thus treating the ] as data rather than
1480 terminating the class. This meant it could skip too much.]
1481
1482 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1483 /g option in pcretest when the pattern contains \K, which makes it possible
1484 to have an empty string match not at the start, even when the pattern is
1485 anchored. Updated pcretest and pcredemo to use this option.
1486
1487 21. If the maximum number of capturing subpatterns in a recursion was greater
1488 than the maximum at the outer level, the higher number was returned, but
1489 with unset values at the outer level. The correct (outer level) value is
1490 now given.
1491
1492 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1493 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1494 make it do so. The string so far is captured, making this feature
1495 compatible with Perl.
1496
1497 23. The tests have been re-organized, adding tests 11 and 12, to make it
1498 possible to check the Perl 5.10 features against Perl 5.10.
1499
1500 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1501 pattern matches a fixed length string. PCRE did not allow this; now it
1502 does. Neither allows recursion.
1503
1504 25. I finally figured out how to implement a request to provide the minimum
1505 length of subject string that was needed in order to match a given pattern.
1506 (It was back references and recursion that I had previously got hung up
1507 on.) This code has now been added to pcre_study(); it finds a lower bound
1508 to the length of subject needed. It is not necessarily the greatest lower
1509 bound, but using it to avoid searching strings that are too short does give
1510 some useful speed-ups. The value is available to calling programs via
1511 pcre_fullinfo().
1512
1513 26. While implementing 25, I discovered to my embarrassment that pcretest had
1514 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1515 study optimizations had never been tested with that matching function.
1516 Oops. What is worse, even when it was passed study data, there was a bug in
1517 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1518 were also very few tests of studied patterns with pcre_dfa_exec().
1519
1520 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1521 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1522 on the other side of the coin, they are no longer allowed to have different
1523 names, because these cannot be distinguished in PCRE, and this has caused
1524 confusion. (This is a difference from Perl.)
1525
1526 28. When duplicate subpattern names are present (necessarily with different
1527 numbers, as required by 27 above), and a test is made by name in a
1528 conditional pattern, either for a subpattern having been matched, or for
1529 recursion in such a pattern, all the associated numbered subpatterns are
1530 tested, and the overall condition is true if the condition is true for any
1531 one of them. This is the way Perl works, and is also more like the way
1532 testing by number works.
1533
1534
1535 Version 7.9 11-Apr-09
1536 ---------------------
1537
1538 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1539 (pcretest), all targets were linked against these libraries. This included
1540 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1541 libraries. This caused unwanted dependencies to be created. This problem
1542 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1543 pcretest is linked with readline.
1544
1545 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1546 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1547 moved outside it again, because FALSE and TRUE are already defined in AIX,
1548 but BOOL is not.
1549
1550 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1551 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1552
1553 4. The pcregrep documentation said spaces were inserted as well as colons (or
1554 hyphens) following file names and line numbers when outputting matching
1555 lines. This is not true; no spaces are inserted. I have also clarified the
1556 wording for the --colour (or --color) option.
1557
1558 5. In pcregrep, when --colour was used with -o, the list of matching strings
1559 was not coloured; this is different to GNU grep, so I have changed it to be
1560 the same.
1561
1562 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1563 each matching line was coloured. Now it goes on to look for further matches
1564 of any of the test patterns, which is the same behaviour as GNU grep.
1565
1566 7. A pattern that could match an empty string could cause pcregrep to loop; it
1567 doesn't make sense to accept an empty string match in pcregrep, so I have
1568 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1569 seems to be how GNU grep behaves.
1570
1571 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1572 start or after a newline", because the conditional assertion was not being
1573 correctly handled. The rule now is that both the assertion and what follows
1574 in the first alternative must satisfy the test.
1575
1576 9. If auto-callout was enabled in a pattern with a conditional group whose
1577 condition was an assertion, PCRE could crash during matching, both with
1578 pcre_exec() and pcre_dfa_exec().
1579
1580 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1581 used for matching.
1582
1583 11. Unicode property support in character classes was not working for
1584 characters (bytes) greater than 127 when not in UTF-8 mode.
1585
1586 12. Added the -M command line option to pcretest.
1587
1588 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1589
1590 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1591
1592 16. Added comments and documentation about mis-use of no_arg in the C++
1593 wrapper.
1594
1595 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1596 from Martin Jerabek that uses macro names for all relevant character and
1597 string constants.
1598
1599 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1600 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1601 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1602 these, but not everybody uses configure.
1603
1604 19. A conditional group that had only one branch was not being correctly
1605 recognized as an item that could match an empty string. This meant that an
1606 enclosing group might also not be so recognized, causing infinite looping
1607 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1608 with the subject "ab", where knowledge that the repeated group can match
1609 nothing is needed in order to break the loop.
1610
1611 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1612 exec(), but without supplying a callout function, matching went wrong.
1613
1614 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1615 leak if the size of the offset vector was greater than 30. When the vector
1616 is smaller, the saved offsets during recursion go onto a local stack
1617 vector, but for larger vectors malloc() is used. It was failing to free
1618 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1619 error, in fact).
1620
1621 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1622 heapframe that is used only when UTF-8 support is enabled. This caused no
1623 problem, but was untidy.
1624
1625 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1626 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1627 included within another project.
1628
1629 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1630 slightly modified by me:
1631
1632 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1633 not building pcregrep.
1634
1635 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1636 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1637
1638 25. Forward references, both numeric and by name, in patterns that made use of
1639 duplicate group numbers, could behave incorrectly or give incorrect errors,
1640 because when scanning forward to find the reference group, PCRE was not
1641 taking into account the duplicate group numbers. A pattern such as
1642 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1643
1644 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1645 the feature of a custom pointer more persuasive (as requested by a user).
1646
1647 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1648 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1649 pre-defined.
1650
1651 28. Added support for (*UTF8) at the start of a pattern.
1652
1653 29. Arrange for flags added by the "release type" setting in CMake to be shown
1654 in the configuration summary.
1655
1656
1657 Version 7.8 05-Sep-08
1658 ---------------------
1659
1660 1. Replaced UCP searching code with optimized version as implemented for Ad
1661 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1662 stage table and inline lookup instead of a function, giving speed ups of 2
1663 to 5 times on some simple patterns that I tested. Permission was given to
1664 distribute the MultiStage2.py script that generates the tables (it's not in
1665 the tarball, but is in the Subversion repository).
1666
1667 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1668 scripts.
1669
1670 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1671 a group with a zero qualifier. The result of the study could be incorrect,
1672 or the function might crash, depending on the pattern.
1673
1674 4. Caseless matching was not working for non-ASCII characters in back
1675 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1676 It now works when Unicode Property Support is available.
1677
1678 5. In pcretest, an escape such as \x{de} in the data was always generating
1679 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1680 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1681 truncation.
1682
1683 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1684
1685 7. Added two (int) casts to pcregrep when printing the difference of two
1686 pointers, in case they are 64-bit values.
1687
1688 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1689 test 2 if it fails.
1690
1691 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1692 and a #define of that name to empty if it is not externally set. This is to
1693 allow users of MSVC to set it if necessary.
1694
1695 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1696 the convenience functions in the pcre_get.c source file.
1697
1698 11. An option change at the start of a pattern that had top-level alternatives
1699 could cause overwriting and/or a crash. This command provoked a crash in
1700 some environments:
1701
1702 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1703
1704 This potential security problem was recorded as CVE-2008-2371.
1705
1706 12. For a pattern where the match had to start at the beginning or immediately
1707 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1708 pcre_dfa_exec() could read past the end of the passed subject if there was
1709 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1710 pcretest so that it places the subject at the end of its malloc-ed buffer.
1711
1712 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1713 exec() might read past the end of the data buffer in UTF-8 mode.
1714
1715 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1716 the data contained the byte 0x85 as part of a UTF-8 character within its
1717 first line. This applied both to normal and DFA matching.
1718
1719 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1720 /^[^d]*?$/8 failed to match "abc".
1721
1722 16. Added a missing copyright notice to pcrecpp_internal.h.
1723
1724 17. Make it more clear in the documentation that values returned from
1725 pcre_exec() in ovector are byte offsets, not character counts.
1726
1727 18. Tidied a few places to stop certain compilers from issuing warnings.
1728
1729 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1730 supplied by Stefan Weber. I made a further small update for 7.8 because
1731 there is a change of source arrangements: the pcre_searchfuncs.c module is
1732 replaced by pcre_ucd.c.
1733
1734
1735 Version 7.7 07-May-08
1736 ---------------------
1737
1738 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1739 a string to a long long, pretend we don't even have a long long." This is
1740 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1741
1742 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1743 pre-7.6 versions, which defined a global no_arg variable instead of putting
1744 it in the RE class. (See also #8 below.)
1745
1746 3. Remove a line of dead code, identified by coverity and reported by Nuno
1747 Lopes.
1748
1749 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1750
1751 (1) The include/exclude patterns were being applied to the whole pathnames
1752 of files, instead of just to the final components.
1753
1754 (2) If there was more than one level of directory, the subdirectories were
1755 skipped unless they satisfied the include/exclude conditions. This is
1756 inconsistent with GNU grep (and could even be seen as contrary to the
1757 pcregrep specification - which I improved to make it absolutely clear).
1758 The action now is always to scan all levels of directory, and just
1759 apply the include/exclude patterns to regular files.
1760
1761 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1762 --exclude_dir in the tests to avoid scanning .svn directories.
1763
1764 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1765 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1766 doesn't support NULs in patterns.
1767
1768 7. Added some missing "const"s to declarations of static tables in
1769 pcre_compile.c and pcre_dfa_exec.c.
1770
1771 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1772 caused by fix #2 above. (Subsequently also a second patch to fix the
1773 first patch. And a third patch - this was a messy problem.)
1774
1775 9. Applied Craig's patch to remove the use of push_back().
1776
1777 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1778 matching function regexec().
1779
1780 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1781 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1782 references. PCRE supports relative numbers with this syntax (I don't think
1783 Oniguruma does).
1784
1785 12. Previously, a group with a zero repeat such as (...){0} was completely
1786 omitted from the compiled regex. However, this means that if the group
1787 was called as a subroutine from elsewhere in the pattern, things went wrong
1788 (an internal error was given). Such groups are now left in the compiled
1789 pattern, with a new opcode that causes them to be skipped at execution
1790 time.
1791
1792 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1793 to the way PCRE behaves:
1794
1795 (a) A lone ] character is dis-allowed (Perl treats it as data).
1796
1797 (b) A back reference to an unmatched subpattern matches an empty string
1798 (Perl fails the current match path).
1799
1800 (c) A data ] in a character class must be notated as \] because if the
1801 first data character in a class is ], it defines an empty class. (In
1802 Perl it is not possible to have an empty class.) The empty class []
1803 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1804 The negative empty class [^] matches any one character, independently
1805 of the DOTALL setting.
1806
1807 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1808 non-existent subpattern following a character class starting with ']' and
1809 containing () gave an internal compiling error instead of "reference to
1810 non-existent subpattern". Fortunately, when the pattern did exist, the
1811 compiled code was correct. (When scanning forwards to check for the
1812 existencd of the subpattern, it was treating the data ']' as terminating
1813 the class, so got the count wrong. When actually compiling, the reference
1814 was subsequently set up correctly.)
1815
1816 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1817 it was being rejected as not supported by pcre_dfa_exec(), even though
1818 other assertions are supported. I have made pcre_dfa_exec() support
1819 (*FAIL).
1820
1821 16. The implementation of 13c above involved the invention of a new opcode,
1822 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1823 cannot be changed at match time, I realized I could make a small
1824 improvement to matching performance by compiling OP_ALLANY instead of
1825 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1826 on the OP_ANY path.
1827
1828 17. Compiling pcretest on Windows with readline support failed without the
1829 following two fixes: (1) Make the unistd.h include conditional on
1830 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1831
1832 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1833 ncurses library to be included for pcretest when ReadLine support is
1834 requested, but also to allow for it to be overridden. This patch came from
1835 Daniel Bergström.
1836
1837 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1838 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1839 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1840 spotting this.
1841
1842
1843 Version 7.6 28-Jan-08
1844 ---------------------
1845
1846 1. A character class containing a very large number of characters with
1847 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1848 overflow.
1849
1850 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1851 HAVE_LONG_LONG is not defined.
1852
1853 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1854 bring them up to date and include new features. This patch includes:
1855
1856 - Fixed PH's badly added libz and libbz2 support.
1857 - Fixed a problem with static linking.
1858 - Added pcredemo. [But later removed - see 7 below.]
1859 - Fixed dftables problem and added an option.
1860 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1861 HAVE_LONG_LONG.
1862 - Added readline support for pcretest.
1863 - Added an listing of the option settings after cmake has run.
1864
1865 4. A user submitted a patch to Makefile that makes it easy to create
1866 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1867 Makefile.am that cause it to include this special target, without
1868 affecting anything else. Note that the same mingw target plus all
1869 the other distribution libraries and programs are now supported
1870 when configuring with CMake (see 6 below) instead of with
1871 Configure/Make.
1872
1873 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1874 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1875 exported in the Windows port". It has not yet been confirmed that the patch
1876 solves the problem, but it does no harm.
1877
1878 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1879 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1880 with CMake, and also correct the comment about stack recursion.
1881
1882 7. Remove the automatic building of pcredemo from the ./configure system and
1883 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1884 of a program that users should build themselves after PCRE is installed, so
1885 building it automatically is not really right. What is more, it gave
1886 trouble in some build environments.
1887
1888 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1889
1890
1891 Version 7.5 10-Jan-08
1892 ---------------------
1893
1894 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1895 values in parens when parsing an RE using the C++ wrapper."
1896
1897 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1898 Characters greater than 255 were excluded from the class instead of being
1899 included.
1900
1901 3. The same bug as (2) above applied to negated POSIX classes such as
1902 [:^space:].
1903
1904 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1905 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1906 I have changed it.
1907
1908 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1909 first named subpattern) and a construct such as (?&a) would reference the
1910 first named subpattern whose name started with "a" (in other words, the
1911 length check was missing). Both these problems are fixed. "Subpattern name
1912 expected" is now given for (?&) (a zero-length name), and this patch also
1913 makes it give the same error for \k'' (previously it complained that that
1914 was a reference to a non-existent subpattern).
1915
1916 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1917 this is right because (?- can be followed by option settings as well as by
1918 digits. I have, however, made the messages clearer.
1919
1920 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1921 than the number used in the conditional) now cause a compile-time error.
1922 This is actually not compatible with Perl, which accepts such patterns, but
1923 treats the conditional as always being FALSE (as PCRE used to), but it
1924 seems to me that giving a diagnostic is better.
1925
1926 8. Change "alphameric" to the more common word "alphanumeric" in comments
1927 and messages.
1928
1929 9. Fix two occurrences of "backslash" in comments that should have been
1930 "backspace".
1931
1932 10. Remove two redundant lines of code that can never be obeyed (their function
1933 was moved elsewhere).
1934
1935 11. The program that makes PCRE's Unicode character property table had a bug
1936 which caused it to generate incorrect table entries for sequences of
1937 characters that have the same character type, but are in different scripts.
1938 It amalgamated them into a single range, with the script of the first of
1939 them. In other words, some characters were in the wrong script. There were
1940 thirteen such cases, affecting characters in the following ranges:
1941
1942 U+002b0 - U+002c1
1943 U+0060c - U+0060d
1944 U+0061e - U+00612
1945 U+0064b - U+0065e
1946 U+0074d - U+0076d
1947 U+01800 - U+01805
1948 U+01d00 - U+01d77
1949 U+01d9b - U+01dbf
1950 U+0200b - U+0200f
1951 U+030fc - U+030fe
1952 U+03260 - U+0327f
1953 U+0fb46 - U+0fbb1
1954 U+10450 - U+1049d
1955
1956 12. The -o option (show only the matching part of a line) for pcregrep was not
1957 compatible with GNU grep in that, if there was more than one match in a
1958 line, it showed only the first of them. It now behaves in the same way as
1959 GNU grep.
1960
1961 13. If the -o and -v options were combined for pcregrep, it printed a blank
1962 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1963 does the same. The return code can be used to tell if there were any
1964 non-matching lines.
1965
1966 14. Added --file-offsets and --line-offsets to pcregrep.
1967
1968 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1969 infinitely looping recursion. The bug was that positive lookaheads were not
1970 being skipped when checking for a possible empty match (negative lookaheads
1971 and both kinds of lookbehind were skipped).
1972
1973 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1974 inclusion of <windows.h> to before rather than after the definition of
1975 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1976
1977 17. Specifying a possessive quantifier with a specific limit for a Unicode
1978 character property caused pcre_compile() to compile bad code, which led at
1979 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1980 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1981 caused the error; without that there was no problem.
1982
1983 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1984
1985 19. Added --enable-pcretest-libreadline.
1986
1987 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1988 RE::GlobalReplace(). As a result, the number of replacements returned was
1989 double what it should be. I removed one of the increments, but Craig sent a
1990 later patch that removed the other one (the right fix) and added unit tests
1991 that check the return values (which was not done before).
1992
1993 21. Several CMake things:
1994
1995 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1996 the names libpcre and libpcreposix, not just pcre and pcreposix.
1997
1998 (2) The above change means that pcretest and pcregrep are now correctly
1999 linked with the newly-built libraries, not previously installed ones.
2000
2001 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2002
2003 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2004 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2005 UTF-8 newline character). The key issue is that the pattern starts .*;
2006 this means that the match must be either at the beginning, or after a
2007 newline. The bug was in the code for advancing after a failed match and
2008 checking that the new position followed a newline. It was not taking
2009 account of UTF-8 characters correctly.
2010
2011 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2012 character classes. PCRE was not treating the sequence [:...:] as a
2013 character class unless the ... were all letters. Perl, however, seems to
2014 allow any characters between [: and :], though of course it rejects as
2015 unknown any "names" that contain non-letters, because all the known class
2016 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2017 for example, whereas PCRE did not - it did not recognize a POSIX character
2018 class. This seemed a bit dangerous, so the code has been changed to be
2019 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2020 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2021 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2022 Perl does, and where it didn't before.
2023
2024 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2025 Windows environments %n is disabled by default.
2026
2027
2028 Version 7.4 21-Sep-07
2029 ---------------------
2030
2031 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2032 means that a class such as [\s] counted as "explicit reference to CR or
2033 LF". That isn't really right - the whole point of the change was to try to
2034 help when there was an actual mention of one of the two characters. So now
2035 the change happens only if \r or \n (or a literal CR or LF) character is
2036 encountered.
2037
2038 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2039 of both had grown to the point where there were only 3 bits left.
2040 Fortunately, there was spare space in the data structure, and so I have
2041 moved the internal flags into a new 16-bit field to free up more option
2042 bits.
2043
2044 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2045 but did not set the internal JCHANGED flag - either of these is enough to
2046 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2047 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2048 start sets both bits.
2049
2050 4. Added options (at build time, compile time, exec time) to change \R from
2051 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2052
2053 5. doc/pcresyntax.html was missing from the distribution.
2054
2055 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2056 compatibility, even though it is no longer used.
2057
2058 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2059 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2060 windows.h file is present (where different names are used). [This was
2061 reversed later after testing - see 16 below.]
2062
2063 8. Changed all #include <config.h> to #include "config.h". There were also
2064 some further <pcre.h> cases that I changed to "pcre.h".
2065
2066 9. When pcregrep was used with the --colour option, it missed the line ending
2067 sequence off the lines that it output.
2068
2069 10. It was pointed out to me that arrays of string pointers cause lots of
2070 relocations when a shared library is dynamically loaded. A technique of
2071 using a single long string with a table of offsets can drastically reduce
2072 these. I have refactored PCRE in four places to do this. The result is
2073 dramatic:
2074
2075 Originally: 290
2076 After changing UCP table: 187
2077 After changing error message table: 43
2078 After changing table of "verbs" 36
2079 After changing table of Posix names 22
2080
2081 Thanks to the folks working on Gregex for glib for this insight.
2082
2083 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2084 unicode-properties was also set.
2085
2086 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2087
2088 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2089 checked only for CRLF.
2090
2091 14. Added casts to pcretest.c to avoid compiler warnings.
2092
2093 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2094
2095 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2096 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2097 entirely. This removes changes made in 7 above.
2098
2099 17. The CMake files have been updated, and there is now more information about
2100 building with CMake in the NON-UNIX-USE document.
2101
2102
2103 Version 7.3 28-Aug-07
2104 ---------------------
2105
2106 1. In the rejigging of the build system that eventually resulted in 7.1, the
2107 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2108 brackets there is not right, since it causes compilers to look for an
2109 installed pcre.h, not the version that is in the source that is being
2110 compiled (which of course may be different). I have changed it back to:
2111
2112 #include "pcre.h"
2113
2114 I have a vague recollection that the change was concerned with compiling in
2115 different directories, but in the new build system, that is taken care of
2116 by the VPATH setting the Makefile.
2117
2118 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2119 when the subject happened to end in the byte 0x85 (e.g. if the last
2120 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2121 characters but of course it shouldn't be taken as a newline when it is part
2122 of another character. The bug was that, for an unlimited repeat of . in
2123 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2124 characters when looking for a newline.
2125
2126 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2127
2128 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2129 in debug output.
2130
2131 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2132 long printing in the pcrecpp unittest when running under MinGW.
2133
2134 6. ESC_K was left out of the EBCDIC table.
2135
2136 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2137 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2138 limit also applies to "virtual nesting" when a pattern is recursive, and in
2139 this case 1000 isn't so big. I have been able to remove this limit at the
2140 expense of backing off one optimization in certain circumstances. Normally,
2141 when pcre_exec() would call its internal match() function recursively and
2142 immediately return the result unconditionally, it uses a "tail recursion"
2143 feature to save stack. However, when a subpattern that can match an empty
2144 string has an unlimited repetition quantifier, it no longer makes this
2145 optimization. That gives it a stack frame in which to save the data for
2146 checking that an empty string has been matched. Previously this was taken
2147 from the 1000-entry workspace that had been reserved. So now there is no
2148 explicit limit, but more stack is used.
2149
2150 8. Applied Daniel's patches to solve problems with the import/export magic
2151 syntax that is required for Windows, and which was going wrong for the
2152 pcreposix and pcrecpp parts of the library. These were overlooked when this
2153 problem was solved for the main library.
2154
2155 9. There were some crude static tests to avoid integer overflow when computing
2156 the size of patterns that contain repeated groups with explicit upper
2157 limits. As the maximum quantifier is 65535, the maximum group length was
2158 set at 30,000 so that the product of these two numbers did not overflow a
2159 32-bit integer. However, it turns out that people want to use groups that
2160 are longer than 30,000 bytes (though not repeat them that many times).
2161 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2162 made it possible to implement the integer overflow checks in a much more
2163 dynamic way, which I have now done. The artificial limitation on group
2164 length has been removed - we now have only the limit on the total length of
2165 the compiled pattern, which depends on the LINK_SIZE setting.
2166
2167 10. Fixed a bug in the documentation for get/copy named substring when
2168 duplicate names are permitted. If none of the named substrings are set, the
2169 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2170 empty string.
2171
2172 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2173 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2174 because the ] is interpreted as the first data character and the
2175 terminating ] is not found. PCRE has been made compatible with Perl in this
2176 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2177 cause memory overwriting.
2178
2179 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2180 string has been matched (to stop an infinite loop). It was not recognizing
2181 a conditional subpattern that could match an empty string if that
2182 subpattern was within another subpattern. For example, it looped when
2183 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2184 condition was not nested. This bug has been fixed.
2185
2186 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2187 past the start of the subject in the presence of bytes with the top bit
2188 set, for example "\x8aBCD".
2189
2190 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2191 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2192
2193 14. Optimized (?!) to (*FAIL).
2194
2195 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2196 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2197 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2198 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2199 does: it's just the validity check that is more restrictive.
2200
2201 16. Inserted checks for integer overflows during escape sequence (backslash)
2202 processing, and also fixed erroneous offset values for syntax errors during
2203 backslash processing.
2204
2205 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2206 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2207
2208 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2209 caused an overrun.
2210
2211 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2212 something other than just ASCII characters) inside a group that had an
2213 unlimited repeat caused a loop at compile time (while checking to see
2214 whether the group could match an empty string).
2215
2216 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2217 [\P{Any}] did so. (Error in the code for printing property names.)
2218
2219 21. An orphan \E inside a character class could cause a crash.
2220
2221 22. A repeated capturing bracket such as (A)? could cause a wild memory
2222 reference during compilation.
2223
2224 23. There are several functions in pcre_compile() that scan along a compiled
2225 expression for various reasons (e.g. to see if it's fixed length for look
2226 behind). There were bugs in these functions when a repeated \p or \P was
2227 present in the pattern. These operators have additional parameters compared
2228 with \d, etc, and these were not being taken into account when moving along
2229 the compiled data. Specifically:
2230
2231 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2232 length.
2233
2234 (b) An item such as \pL+ within a repeated group could cause crashes or
2235 loops.
2236
2237 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2238 "reference to non-existent subpattern" error.
2239
2240 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2241
2242 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2243 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2244
2245 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2246
2247 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2248 character were causing crashes (broken optimization).
2249
2250 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2251 \p or \P) caused a compile-time loop.
2252
2253 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2254 break. For example, the unstudied pattern [\r\n]A does not match the string
2255 "\r\nA" because change 7.0/46 below moves the current point on by two
2256 characters after failing to match at the start. However, the pattern \nA
2257 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2258 the same is true. There doesn't seem any very clean way out of this, but
2259 what I have chosen to do makes the common cases work: PCRE now takes note
2260 of whether there can be an explicit match for \r or \n anywhere in the
2261 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2262 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2263 pattern has explicit CR or LF references.
2264
2265 29. Added (*CR) etc for changing newline setting at start of pattern.
2266
2267
2268 Version 7.2 19-Jun-07
2269 ---------------------
2270
2271 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2272 which is apparently normally available under Windows.
2273
2274 2. Re-jig the pcregrep tests with different newline settings in an attempt
2275 to make them independent of the local environment's newline setting.
2276
2277 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2278
2279 4. Some of the "internals" tests were previously cut out when the link size
2280 was not 2, because the output contained actual offsets. The recent new
2281 "Z" feature of pcretest means that these can be cut out, making the tests
2282 usable with all link sizes.
2283
2284 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2285 stack recursion. This gives a massive performance boost under BSD, but just
2286 a small improvement under Linux. However, it saves one field in the frame
2287 in all cases.
2288
2289 6. Added more features from the forthcoming Perl 5.10:
2290
2291 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2292 recursion call. It refers to the nth most recently opened parentheses.
2293
2294 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2295 to be opened parentheses.
2296
2297 (c) Conditions that refer to capturing parentheses can be specified
2298 relatively, for example, (?(-2)... or (?(+3)...
2299
2300 (d) \K resets the start of the current match so that everything before
2301 is not part of it.
2302
2303 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2304
2305 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2306 reference syntax.
2307
2308 (g) (?| introduces a group in which the numbering of parentheses in each
2309 alternative starts with the same number.
2310
2311 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2312
2313 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2314 PCRE_INFO_JCHANGED.
2315
2316 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2317 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2318 for detecting groups that can match an empty string.
2319
2320 9. A pattern with a very large number of alternatives (more than several
2321 hundred) was running out of internal workspace during the pre-compile
2322 phase, where pcre_compile() figures out how much memory will be needed. A
2323 bit of new cunning has reduced the workspace needed for groups with
2324 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2325 workspace instead of running out of the 4096 that are available.
2326
2327 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2328
2329 11. Applied patch from Google to remove an optimization that didn't quite work.
2330 The report of the bug said:
2331
2332 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2333 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2334 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2335
2336 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2337 it matched the wrong number of bytes.
2338
2339
2340 Version 7.1 24-Apr-07
2341 ---------------------
2342
2343 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2344 that is more "standard", making use of automake and other Autotools. There
2345 is some re-arrangement of the files and adjustment of comments consequent
2346 on this.
2347
2348 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2349 for recursive directory scanning broke on some systems because the files
2350 are not scanned in any specific order and on different systems the order
2351 was different. A call to "sort" has been inserted into RunGrepTest for the
2352 approprate test as a short-term fix. In the longer term there may be an
2353 alternative.
2354
2355 3. I had an email from Eric Raymond about problems translating some of PCRE's
2356 man pages to HTML (despite the fact that I distribute HTML pages, some
2357 people do their own conversions for various reasons). The problems
2358 concerned the use of low-level troff macros .br and .in. I have therefore
2359 removed all such uses from the man pages (some were redundant, some could
2360 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2361 HTML has been updated to handle .nf/.fi and to complain if it encounters
2362 .br or .in.
2363
2364 4. Updated comments in configure.ac that get placed in config.h.in and also
2365 arranged for config.h to be included in the distribution, with the name
2366 config.h.generic, for the benefit of those who have to compile without
2367 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2368
2369 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2370 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2371 makevp.bat for the current PCRE, using the additional files
2372 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2373
2374 6. A Windows user reported a minor discrepancy with test 2, which turned out
2375 to be caused by a trailing space on an input line that had got lost in his
2376 copy. The trailing space was an accident, so I've just removed it.
2377
2378 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2379 that is needed.
2380
2381 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2382 as "const" (a) because they are and (b) because it helps the PHP
2383 maintainers who have recently made a script to detect big data structures
2384 in the php code that should be moved to the .rodata section. I remembered
2385 to update Builducptable as well, so it won't revert if ucptable.h is ever
2386 re-created.
2387
2388 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2389 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2390 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2391 support is not required. This saves 1.5-2K of code, which is important in
2392 some applications.
2393
2394 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2395 so as not to refer to the tables, even though these functions will never be
2396 called when UTF-8 support is disabled. Otherwise there are problems with a
2397 shared library.
2398
2399 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2400
2401 (a) It was defining its arguments as char * instead of void *.
2402
2403 (b) It was assuming that all moves were upwards in memory; this was true
2404 a long time ago when I wrote it, but is no longer the case.
2405
2406 The emulated memove() is provided for those environments that have neither
2407 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2408 is clearly not the case, as these two bugs were recently reported.
2409
2410 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2411 and Detrail to create the HTML documentation, the .txt form of the man
2412 pages, and it removes trailing spaces from listed files. It also creates
2413 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2414 case, it wraps all the #defines with #ifndefs. This script should be run
2415 before "make dist".
2416
2417 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2418 with Unicode property support.
2419
2420 (a) For a maximizing quantifier, if the two different cases of the
2421 character were of different lengths in their UTF-8 codings (there are
2422 some cases like this - I found 11), and the matching function had to
2423 back up over a mixture of the two cases, it incorrectly assumed they
2424 were both the same length.
2425
2426 (b) When PCRE was configured to use the heap rather than the stack for
2427 recursion during matching, it was not correctly preserving the data for
2428 the other case of a UTF-8 character when checking ahead for a match
2429 while processing a minimizing repeat. If the check also involved
2430 matching a wide character, but failed, corruption could cause an
2431 erroneous result when trying to check for a repeat of the original
2432 character.
2433
2434 13. Some tidying changes to the testing mechanism:
2435
2436 (a) The RunTest script now detects the internal link size and whether there
2437 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2438 values substituted by "configure". (The RunGrepTest script already did
2439 this for UTF-8.) The configure.ac script no longer substitutes the
2440 relevant variables.
2441
2442 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2443 with length and offset values. This means that the output is different
2444 for different internal link sizes. Test 2 is skipped for link sizes
2445 other than 2 because of this, bypassing the problem. Unfortunately,
2446 there was also a test in test 3 (the locale tests) that used /B and
2447 failed for link sizes other than 2. Rather than cut the whole test out,
2448 I have added a new /Z option to pcretest that replaces the length and
2449 offset values with spaces. This is now used to make test 3 independent
2450 of link size. (Test 2 will be tidied up later.)
2451
2452 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2453 segmentation fault instead of returning the appropriate error message.
2454
2455 15. In multiline mode when the newline sequence was set to "any", the pattern
2456 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2457 This doesn't seem right; it now treats the CRLF combination as the line
2458 ending, and so does not match in that case. It's only a pattern such as ^$
2459 that would hit this one: something like ^ABC$ would have failed after \r
2460 and then tried again after \r\n.
2461
2462 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2463 in an attempt to make files that differ only in their line terminators
2464 compare equal. This works on Linux.
2465
2466 17. Under certain error circumstances pcregrep might try to free random memory
2467 as it exited. This is now fixed, thanks to valgrind.
2468
2469 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2470 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2471 was because its rules for how to advance for /g after matching an empty
2472 string at the end of a line did not allow for this case. They now check for
2473 it specially.
2474
2475 20. pcretest is supposed to handle patterns and data of any length, by
2476 extending its buffers when necessary. It was getting this wrong when the
2477 buffer for a data line had to be extended.
2478
2479 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2480 CRLF as a newline sequence.
2481
2482 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2483 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2484 I have nevertheless tidied it up.
2485
2486 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2487
2488 24. Added a man page for pcre-config.
2489
2490
2491 Version 7.0 19-Dec-06
2492 ---------------------
2493
2494 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2495 moving to gcc 4.1.1.
2496
2497 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2498 sys/time.h, which is documented as needed for this function. It doesn't
2499 seem to matter on Linux, but it showed up on some releases of OS X.
2500
2501 3. It seems that there are systems where bytes whose values are greater than
2502 127 match isprint() in the "C" locale. The "C" locale should be the
2503 default when a C program starts up. In most systems, only ASCII printing
2504 characters match isprint(). This difference caused the output from pcretest
2505 to vary, making some of the tests fail. I have changed pcretest so that:
2506
2507 (a) When it is outputting text in the compiled version of a pattern, bytes
2508 other than 32-126 are always shown as hex escapes.
2509
2510 (b) When it is outputting text that is a matched part of a subject string,
2511 it does the same, unless a different locale has been set for the match
2512 (using the /L modifier). In this case, it uses isprint() to decide.
2513
2514 4. Fixed a major bug that caused incorrect computation of the amount of memory
2515 required for a compiled pattern when options that changed within the
2516 pattern affected the logic of the preliminary scan that determines the
2517 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2518 that the computed length was too small. The symptoms of this bug were
2519 either the PCRE error "internal error: code overflow" from pcre_compile(),
2520 or a glibc crash with a message such as "pcretest: free(): invalid next
2521 size (fast)". Examples of patterns that provoked this bug (shown in
2522 pcretest format) are:
2523
2524 /(?-x: )/x
2525 /(?x)(?-x: \s*#\s*)/
2526 /((?i)[\x{c0}])/8
2527 /(?i:[\x{c0}])/8
2528
2529 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2530 is now done differently.
2531
2532 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2533 wrapper classes; (b) implement a new function in the C++ scanner that is
2534 more efficient than the old way of doing things because it avoids levels of
2535 recursion in the regex matching; (c) add a paragraph to the documentation
2536 for the FullMatch() function.
2537
2538 6. The escape sequence \n was being treated as whatever was defined as
2539 "newline". Not only was this contrary to the documentation, which states
2540 that \n is character 10 (hex 0A), but it also went horribly wrong when
2541 "newline" was defined as CRLF. This has been fixed.
2542
2543 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2544 was being set to -1 for the "end of line" case (supposedly a value that no
2545 character can have). Though this value is never used (the check for end of
2546 line is "zero bytes in current character"), it caused compiler complaints.
2547 I've changed it to 0xffffffff.
2548
2549 8. In pcre_version.c, the version string was being built by a sequence of
2550 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2551 string (as it is for production releases) called a macro with an empty
2552 argument. The C standard says the result of this is undefined. The gcc
2553 compiler treats it as an empty string (which was what was wanted) but it is
2554 reported that Visual C gives an error. The source has been hacked around to
2555 avoid this problem.
2556
2557 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2558 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2559 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2560 of them did).
2561
2562 10. Originally, pcretest opened its input and output without "b"; then I was
2563 told that "b" was needed in some environments, so it was added for release
2564 5.0 to both the input and output. (It makes no difference on Unix-like
2565 systems.) Later I was told that it is wrong for the input on Windows. I've
2566 now abstracted the modes into two macros, to make it easier to fiddle with
2567 them, and removed "b" from the input mode under Windows.
2568
2569 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2570
2571 12. Added -help and --help to pcretest as an official way of being reminded
2572 of the options.
2573
2574 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2575 and pcrecpp.cc because they annoy compilers at high warning levels.
2576
2577 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2578
2579 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2580 scripts are not C programs :-) and which was not noticed because it works
2581 on Linux.
2582
2583 16. pcretest is supposed to handle any length of pattern and data line (as one
2584 line or as a continued sequence of lines) by extending its input buffer if
2585 necessary. This feature was broken for very long pattern lines, leading to
2586 a string of junk being passed to pcre_compile() if the pattern was longer
2587 than about 50K.
2588
2589 17. I have done a major re-factoring of the way pcre_compile() computes the
2590 amount of memory needed for a compiled pattern. Previously, there was code
2591 that made a preliminary scan of the pattern in order to do this. That was
2592 OK when PCRE was new, but as the facilities have expanded, it has become
2593 harder and harder to keep it in step with the real compile phase, and there
2594 have been a number of bugs (see for example, 4 above). I have now found a
2595 cunning way of running the real compile function in a "fake" mode that
2596 enables it to compute how much memory it would need, while actually only
2597 ever using a few hundred bytes of working memory and without too many
2598 tests of the mode. This should make future maintenance and development
2599 easier. A side effect of this work is that the limit of 200 on the nesting
2600 depth of parentheses has been removed (though this was never a serious
2601 limitation, I suspect). However, there is a downside: pcre_compile() now
2602 runs more slowly than before (30% or more, depending on the pattern). I
2603 hope this isn't a big issue. There is no effect on runtime performance.
2604
2605 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2606 newline (only possible for the last line of a file) and it was a
2607 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2608
2609 19. Added additional timing features to pcretest. (1) The -tm option now times
2610 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2611 separate command line item, by a number that specifies the number of
2612 repeats to use when timing. The default is 50000; this gives better
2613 precision, but takes uncomfortably long for very large patterns.
2614
2615 20. Extended pcre_study() to be more clever in cases where a branch of a
2616 subpattern has no definite first character. For example, (a*|b*)[cd] would
2617 previously give no result from pcre_study(). Now it recognizes that the
2618 first character must be a, b, c, or d.
2619
2620 21. There was an incorrect error "recursive call could loop indefinitely" if
2621 a subpattern (or the entire pattern) that was being tested for matching an
2622 empty string contained only one non-empty item after a nested subpattern.
2623 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2624 incorrectly, because the \d was being skipped in the check.
2625
2626 22. The pcretest program now has a new pattern option /B and a command line
2627 option -b, which is equivalent to adding /B to every pattern. This causes
2628 it to show the compiled bytecode, without the additional information that
2629 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2630 is the same as /B/I).
2631
2632 23. A new optimization is now able automatically to treat some sequences such
2633 as a*b as a*+b. More specifically, if something simple (such as a character
2634 or a simple class like \d) has an unlimited quantifier, and is followed by
2635 something that cannot possibly match the quantified thing, the quantifier
2636 is automatically "possessified".
2637
2638 24. A recursive reference to a subpattern whose number was greater than 39
2639 went wrong under certain circumstances in UTF-8 mode. This bug could also
2640 have affected the operation of pcre_study().
2641
2642 25. Realized that a little bit of performance could be had by replacing
2643 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2644
2645 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2646
2647 27. Possessive quantifiers such as a++ were previously implemented by turning
2648 them into atomic groups such as ($>a+). Now they have their own opcodes,
2649 which improves performance. This includes the automatically created ones
2650 from 23 above.
2651
2652 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2653 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2654 the first matched character to be a colon. This applied both to named and
2655 numbered groups.
2656
2657 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2658
2659 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2660 building PCRE on Windows easier, so I have included it in the distribution.
2661
2662 31. There is now a check in pcretest against a ridiculously large number being
2663 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2664 loop, the loop is abandoned.
2665
2666 32. Forward references to subpatterns in conditions such as (?(2)...) where
2667 subpattern 2 is defined later cause pcre_compile() to search forwards in
2668 the pattern for the relevant set of parentheses. This search went wrong
2669 when there were unescaped parentheses in a character class, parentheses
2670 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2671
2672 33. "Subroutine" calls and backreferences were previously restricted to
2673 referencing subpatterns earlier in the regex. This restriction has now
2674 been removed.
2675
2676 34. Added a number of extra features that are going to be in Perl 5.10. On the
2677 whole, these are just syntactic alternatives for features that PCRE had
2678 previously implemented using the Python syntax or my own invention. The
2679 other formats are all retained for compatibility.
2680
2681 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2682 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2683 also .NET compatible.
2684
2685 (b) A recursion or subroutine call to a named group can now be defined as
2686 (?&name) as well as (?P>name).
2687
2688 (c) A backreference to a named group can now be defined as \k<name> or
2689 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2690 5.10, are also .NET compatible.
2691
2692 (d) A conditional reference to a named group can now use the syntax
2693 (?(<name>) or (?('name') as well as (?(name).
2694
2695 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2696 groups (named and numbered) that are never evaluated inline, but can be
2697 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2698 is always false. There may be only one alternative in such a group.
2699
2700 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2701 as the simple (?(R). The condition is true only if the most recent
2702 recursion is that of the given number or name. It does not search out
2703 through the entire recursion stack.
2704
2705 (g) The escape \gN or \g{N} has been added, where N is a positive or
2706 negative number, specifying an absolute or relative reference.
2707
2708 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2709 some "unreachable code" warnings.
2710
2711 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2712 things, this adds five new scripts.
2713
2714 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2715 There were also incompatibilities regarding the handling of \Q..\E inside
2716 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2717 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2718
2719 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2720 matches an empty string, and forcibly breaks the loop. There were bugs in
2721 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2722 against aaaa the result was just "a" rather than "aaaa", for example. Two
2723 separate and independent bugs (that affected different cases) have been
2724 fixed.
2725
2726 39. Refactored the code to abolish the use of different opcodes for small
2727 capturing bracket numbers. This is a tidy that I avoided doing when I
2728 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2729 The new approach is not only tidier, it makes it possible to reduce the
2730 memory needed to fix the previous bug (38).
2731
2732 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2733 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2734 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2735 mode.
2736
2737 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2738 report.
2739
2740 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2741 copy construction and assignment in the C++ wrapper.
2742
2743 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2744 couple of bugs that could have given wrong results in the "--newline=crlf"
2745 case.
2746
2747 44. Added a number of casts and did some reorganization of signed/unsigned int
2748 variables following suggestions from Dair Grant. Also renamed the variable
2749 "this" as "item" because it is a C++ keyword.
2750
2751 45. Arranged for dftables to add
2752
2753 #include "pcre_internal.h"
2754
2755 to pcre_chartables.c because without it, gcc 4.x may remove the array
2756 definition from the final binary if PCRE is built into a static library and
2757 dead code stripping is activated.
2758
2759 46. For an unanchored pattern, if a match attempt fails at the start of a
2760 newline sequence, and the newline setting is CRLF or ANY, and the next two
2761 characters are CRLF, advance by two characters instead of one.
2762
2763
2764 Version 6.7 04-Jul-06
2765 ---------------------
2766
2767 1. In order to handle tests when input lines are enormously long, pcretest has
2768 been re-factored so that it automatically extends its buffers when
2769 necessary. The code is crude, but this _is_ just a test program. The
2770 default size has been increased from 32K to 50K.
2771
2772 2. The code in pcre_study() was using the value of the re argument before
2773 testing it for NULL. (Of course, in any sensible call of the function, it
2774 won't be NULL.)
2775
2776 3. The memmove() emulation function in pcre_internal.h, which is used on
2777 systems that lack both memmove() and bcopy() - that is, hardly ever -
2778 was missing a "static" storage class specifier.
2779
2780 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2781 containing an extended class (one that cannot be represented by a bitmap
2782 because it contains high-valued characters or Unicode property items, e.g.
2783 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2784 pattern, but PCRE should not loop if you do not (it no longer does).
2785 [Detail: two cases were found: (a) a repeated subpattern containing an
2786 extended class; (b) a recursive reference to a subpattern that followed a
2787 previous extended class. It wasn't skipping over the extended class
2788 correctly when UTF-8 mode was not set.]
2789
2790 5. A negated single-character class was not being recognized as fixed-length
2791 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2792 compile error "lookbehind assertion is not fixed length".
2793
2794 6. The RunPerlTest auxiliary script was showing an unexpected difference
2795 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2796 write a Perl script that can interpret lines of an input file either as
2797 byte characters or as UTF-8, which is what "perltest" was being required to
2798 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2799 can't do is switch easily at run time between having the "use utf8;" pragma
2800 or not. In the end, I fudged it by using the RunPerlTest script to insert
2801 "use utf8;" explicitly for the UTF-8 tests.
2802
2803 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2804 the end of the subject string, contrary to the documentation and to what
2805 Perl does. This was true of both matching functions. Now it matches only at
2806 the start of the subject and immediately after *internal* newlines.
2807
2808 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2809 a pointer to an int instead of a pointer to an unsigned long int. This
2810 caused problems on 64-bit systems.
2811
2812 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2813 instance of the 'standard' template library not being so standard".
2814
2815 10. There was no check on the number of named subpatterns nor the maximum
2816 length of a subpattern name. The product of these values is used to compute
2817 the size of the memory block for a compiled pattern. By supplying a very
2818 long subpattern name and a large number of named subpatterns, the size
2819 computation could be caused to overflow. This is now prevented by limiting
2820 the length of names to 32 characters, and the number of named subpatterns
2821 to 10,000.
2822
2823 11. Subpatterns that are repeated with specific counts have to be replicated in
2824 the compiled pattern. The size of memory for this was computed from the
2825 length of the subpattern and the repeat count. The latter is limited to
2826 65535, but there was no limit on the former, meaning that integer overflow
2827 could in principle occur. The compiled length of a repeated subpattern is
2828 now limited to 30,000 bytes in order to prevent this.
2829
2830 12. Added the optional facility to have named substrings with the same name.
2831
2832 13. Added the ability to use a named substring as a condition, using the
2833 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2834 are numbers (not recommended). Forward references are permitted.
2835
2836 14. Added forward references in named backreferences (if you see what I mean).
2837
2838 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2839 pattern could run off the end of the subject. For example, the pattern
2840 "(?s)(.{1,5})"8 did this with the subject "ab".
2841
2842 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2843 PCRE_CASELESS was set when matching characters that were quantified with ?
2844 or *.
2845
2846 17. A character class other than a single negated character that had a minimum
2847 but no maximum quantifier - for example [ab]{6,} - was not handled
2848 correctly by pce_dfa_exec(). It would match only one character.
2849
2850 18. A valid (though odd) pattern that looked like a POSIX character
2851 class but used an invalid character after [ (for example [[,abc,]]) caused
2852 pcre_compile() to give the error "Failed: internal error: code overflow" or
2853 in some cases to crash with a glibc free() error. This could even happen if
2854 the pattern terminated after [[ but there just happened to be a sequence of
2855 letters, a binary zero, and a closing ] in the memory that followed.
2856
2857 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2858 over the years. Originally (before any Unicode support), just the bottom 8
2859 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2860 output from "man perlunicode" includes this:
2861
2862 The regular expression compiler produces polymorphic opcodes. That
2863 is, the pattern adapts to the data and automatically switches to
2864 the Unicode character scheme when presented with Unicode data--or
2865 instead uses a traditional byte scheme when presented with byte
2866 data.
2867
2868 Sadly, a wide octal escape does not cause a switch, and in a string with
2869 no other multibyte characters, these octal escapes are treated as before.
2870 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2871 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2872 Unicode string.
2873
2874 I have not perpetrated such confusion in PCRE. Up till now, it took just
2875 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2876 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2877 translate to the appropriate multibyte character.
2878
2879 29. Applied some refactoring to reduce the number of warnings from Microsoft
2880 and Borland compilers. This has included removing the fudge introduced
2881 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2882 a warning about an unused variable.
2883
2884 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2885 characters since release 4.0, because Perl (from release 5.004) does not.
2886 [Or at least, is documented not to: some releases seem to be in conflict
2887 with the documentation.] However, when a pattern was studied with
2888 pcre_study() and all its branches started with \s, PCRE still included VT
2889 as a possible starting character. Of course, this did no harm; it just
2890 caused an unnecessary match attempt.
2891
2892 22. Removed a now-redundant internal flag bit that recorded the fact that case
2893 dependency changed within the pattern. This was once needed for "required
2894 byte" processing, but is no longer used. This recovers a now-scarce options
2895 bit. Also moved the least significant internal flag bit to the most-
2896 significant bit of the word, which was not previously used (hangover from
2897 the days when it was an int rather than a uint) to free up another bit for
2898 the future.
2899
2900 23. Added support for CRLF line endings as well as CR and LF. As well as the
2901 default being selectable at build time, it can now be changed at runtime
2902 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2903 specify that it is scanning data with non-default line endings.
2904
2905 24. Changed the definition of CXXLINK to make it agree with the definition of
2906 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2907
2908 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2909 recursions. This makes a big different to stack usage for some patterns.
2910
2911 26. If a subpattern containing a named recursion or subroutine reference such
2912 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2913 the space required for the compiled pattern went wrong and gave too small a
2914 value. Depending on the environment, this could lead to "Failed: internal
2915 error: code overflow at offset 49" or "glibc detected double free or
2916 corruption" errors.
2917
2918 27. Applied patches from Google (a) to support the new newline modes and (b) to
2919 advance over multibyte UTF-8 characters in GlobalReplace.
2920
2921 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2922 difference for some implementation of PCRE in some Windows version.
2923
2924 29. Added some extra testing facilities to pcretest:
2925
2926 \q<number> in a data line sets the "match limit" value
2927 \Q<number> in a data line sets the "match recursion limt" value
2928 -S <number> sets the stack size, where <number> is in megabytes
2929
2930 The -S option isn't available for Windows.
2931
2932
2933 Version 6.6 06-Feb-06
2934 ---------------------
2935
2936 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2937 in pcreposix.h. I have copied the definition from pcre.h.
2938
2939 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2940 because pcre.h is no longer a built file.
2941
2942 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2943 not normally included in the compiled code.
2944
2945
2946 Version 6.5 01-Feb-06
2947 ---------------------
2948
2949 1. When using the partial match feature with pcre_dfa_exec(), it was not
2950 anchoring the second and subsequent partial matches at the new starting
2951 point. This could lead to incorrect results. For example, with the pattern
2952 /1234/, partially matching against "123" and then "a4" gave a match.
2953
2954 2. Changes to pcregrep:
2955
2956 (a) All non-match returns from pcre_exec() were being treated as failures
2957 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2958 error message is output. Some extra information is given for the
2959 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2960 probably the only errors that are likely to be caused by users (by
2961 specifying a regex that has nested indefinite repeats, for instance).
2962 If there are more than 20 of these errors, pcregrep is abandoned.
2963
2964 (b) A binary zero was treated as data while matching, but terminated the
2965 output line if it was written out. This has been fixed: binary zeroes
2966 are now no different to any other data bytes.
2967
2968 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2969 used to set a locale for matching. The --locale=xxxx long option has
2970 been added (no short equivalent) to specify a locale explicitly on the
2971 pcregrep command, overriding the environment variables.
2972
2973 (d) When -B was used with -n, some line numbers in the output were one less
2974 than they should have been.
2975
2976 (e) Added the -o (--only-matching) option.
2977
2978 (f) If -A or -C was used with -c (count only), some lines of context were
2979 accidentally printed for the final match.
2980
2981 (g) Added the -H (--with-filename) option.
2982
2983 (h) The combination of options -rh failed to suppress file names for files
2984 that were found from directory arguments.
2985
2986 (i) Added the -D (--devices) and -d (--directories) options.
2987
2988 (j) Added the -F (--fixed-strings) option.
2989
2990 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2991
2992 (l) Added the --colo(u)r option.
2993
2994 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2995 is not present by default.
2996
2997 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2998 items such as (?R) or (?1), when the recursion could match a number of
2999 alternatives. If it matched one of the alternatives, but subsequently,
3000 outside the recursion, there was a failure, the code tried to back up into
3001 the recursion. However, because of the way PCRE is implemented, this is not
3002 possible, and the result was an incorrect result from the match.
3003
3004 In order to prevent this happening, the specification of recursion has
3005 been changed so that all such subpatterns are automatically treated as
3006 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3007
3008 4. I had overlooked the fact that, in some locales, there are characters for
3009 which isalpha() is true but neither isupper() nor islower() are true. In
3010 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3011 and ordfeminine) are like this. This affected the treatment of \w and \W
3012 when they appeared in character classes, but not when they appeared outside
3013 a character class. The bit map for "word" characters is now created
3014 separately from the results of isalnum() instead of just taking it from the
3015 upper, lower, and digit maps. (Plus the underscore character, of course.)
3016
3017 5. The above bug also affected the handling of POSIX character classes such as
3018 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3019 permanent tables. Instead, the bit maps for such a class were previously
3020 created as the appropriate unions of the upper, lower, and digit bitmaps.
3021 Now they are created by subtraction from the [[:word:]] class, which has
3022 its own bitmap.
3023
3024 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3025 It is created by subtracting the vertical space characters (\x09, \x0a,
3026 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3027 subtraction was done in the overall bitmap for a character class, meaning
3028 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3029 be recognized. This bug has been fixed.
3030
3031 7. Patches from the folks at Google:
3032
3033 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3034 real life, but is still worth protecting against".
3035
3036 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3037 regular expressions".
3038
3039 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3040 have it.
3041
3042 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3043 "configure" and the latter not, in order to fix a problem somebody had
3044 with compiling the Arg class on HP-UX.
3045
3046 (e) Improve the error-handling of the C++ wrapper a little bit.
3047
3048 (f) New tests for checking recursion limiting.
3049
3050 8. The pcre_memmove() function, which is used only if the environment does not
3051 have a standard memmove() function (and is therefore rarely compiled),
3052 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3053 returning a result (though PCRE never actually uses the result).
3054
3055 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3056 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3057 returned instead of calling malloc() with an overflowing number that would
3058 most likely cause subsequent chaos.
3059
3060 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3061
3062 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3063 with this option is matched, the nmatch and pmatch options of regexec() are
3064 ignored.
3065
3066 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3067 provided in case anyone wants to the the POSIX interface with UTF-8
3068 strings.
3069
3070 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3071 C++ linking (needed for some HP-UX environments).
3072
3073 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3074 (unused parameter) and in the pcre_printint() function (omitted "default"
3075 switch label when the default is to do nothing).
3076
3077 15. Added some code to make it possible, when PCRE is compiled as a C++
3078 library, to replace subject pointers for pcre_exec() with a smart pointer
3079 class, thus making it possible to process discontinuous strings.
3080
3081 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3082 much the same function. They were added by different people who were trying
3083 to make PCRE easy to compile on non-Unix systems. It has been suggested
3084 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3085 for compiling on Windows systems. I have therefore replaced it with
3086 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3087 defaults to "extern" for C or "extern C" for C++, which works fine on
3088 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3089 SCOPE with something explicit in config.h. In addition:
3090
3091 (a) pcreposix.h still had just "extern" instead of either of these macros;
3092 I have replaced it with PCRE_DATA_SCOPE.
3093
3094 (b) Functions such as _pcre_xclass(), which are internal to the library,
3095 but external in the C sense, all had PCRE_EXPORT in their definitions.
3096 This is apparently wrong for the Windows case, so I have removed it.
3097 (It makes no difference on Unix-like systems.)
3098
3099 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3100 of recursive calls to match(). This is different to MATCH_LIMIT because
3101 that limits the total number of calls to match(), not all of which increase
3102 the depth of recursion. Limiting the recursion depth limits the amount of
3103 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3104 when PCRE is compiled, and changed at run time. A patch from Google adds
3105 this functionality to the C++ interface.
3106
3107 18. Changes to the handling of Unicode character properties:
3108
3109 (a) Updated the table to Unicode 4.1.0.
3110
3111 (b) Recognize characters that are not in the table as "Cn" (undefined).
3112
3113 (c) I revised the way the table is implemented to a much improved format
3114 which includes recognition of ranges. It now supports the ranges that
3115 are defined in UnicodeData.txt, and it also amalgamates other
3116 characters into ranges. This has reduced the number of entries in the
3117 table from around 16,000 to around 3,000, thus reducing its size
3118 considerably. I realized I did not need to use a tree structure after
3119 all - a binary chop search is just as efficient. Having reduced the
3120 number of entries, I extended their size from 6 bytes to 8 bytes to
3121 allow for more data.
3122
3123 (d) Added support for Unicode script names via properties such as \p{Han}.
3124
3125 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3126 matching that character.
3127
3128 20. When matching a repeated Unicode property with a minimum greater than zero,
3129 (for example \pL{2,}), PCRE could look past the end of the subject if it
3130 reached it while seeking the minimum number of characters. This could
3131 happen only if some of the characters were more than one byte long, because
3132 there is a check for at least the minimum number of bytes.
3133
3134 21. Refactored the implementation of \p and \P so as to be more general, to
3135 allow for more different types of property in future. This has changed the
3136 compiled form incompatibly. Anybody with saved compiled patterns that use
3137 \p or \P will have to recompile them.
3138
3139 22. Added "Any" and "L&" to the supported property types.
3140
3141 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3142 but give a compile time error if the value is greater than 0xff.
3143
3144 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3145 accidentally not being installed or uninstalled.
3146
3147 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3148 made were to insert the current release number. This seemed silly, because
3149 it made things harder for people building PCRE on systems that don't run
3150 "configure". I have turned pcre.h into a distributed file, no longer built
3151 by "configure", with the version identification directly included. There is
3152 no longer a pcre.h.in file.
3153
3154 However, this change necessitated a change to the pcre-config script as
3155 well. It is built from pcre-config.in, and one of the substitutions was the
3156 release number. I have updated configure.ac so that ./configure now finds
3157 the release number by grepping pcre.h.
3158
3159 26. Added the ability to run the tests under valgrind.
3160
3161
3162 Version 6.4 05-Sep-05
3163 ---------------------
3164
3165 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3166 "--" to be printed when multiple files were scanned, even when none of the
3167 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3168 consider it to be a bug, and have restored the previous behaviour.
3169
3170 2. A couple of code tidies to get rid of compiler warnings.
3171
3172 3. The pcretest program used to cheat by referring to symbols in the library
3173 whose names begin with _pcre_. These are internal symbols that are not
3174 really supposed to be visible externally, and in some environments it is
3175 possible to suppress them. The cheating is now confined to including
3176 certain files from the library's source, which is a bit cleaner.
3177
3178 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3179 file's purpose clearer.
3180
3181 5. Reorganized pcre_ucp_findchar().
3182
3183
3184 Version 6.3 15-Aug-05
3185 ---------------------
3186
3187 1. The file libpcre.pc.in did not have general read permission in the tarball.
3188
3189 2. There were some problems when building without C++ support:
3190
3191 (a) If C++ support was not built, "make install" and "make test" still
3192 tried to test it.
3193
3194 (b) There were problems when the value of CXX was explicitly set. Some
3195 changes have been made to try to fix these, and ...
3196
3197 (c) --disable-cpp can now be used to explicitly disable C++ support.
3198
3199 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3200 backslash in a target when C++ was disabled. This confuses some
3201 versions of "make", apparently. Using an intermediate variable solves
3202 this. (Same for CPP_LOBJ.)
3203
3204 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3205 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3206 necessary on certain architectures.
3207
3208 4. Added a setting of -export-symbols-regex to the link command to remove
3209 those symbols that are exported in the C sense, but actually are local
3210 within the library, and not documented. Their names all begin with
3211 "_pcre_". This is not a perfect job, because (a) we have to except some
3212 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3213 available (and never for static libraries). I have made a note to try to
3214 find a way round (a) in the future.
3215
3216
3217 Version 6.2 01-Aug-05
3218 ---------------------
3219
3220 1. There was no test for integer overflow of quantifier values. A construction
3221 such as {1111111111111111} would give undefined results. What is worse, if
3222 a minimum quantifier for a parenthesized subpattern overflowed and became
3223 negative, the calculation of the memory size went wrong. This could have
3224 led to memory overwriting.
3225
3226 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3227
3228 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3229 operating environments where this matters.
3230
3231 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3232 PCRE options from within the C++ wrapper.
3233
3234 5. Named capturing subpatterns were not being correctly counted when a pattern
3235 was compiled. This caused two problems: (a) If there were more than 100
3236 such subpatterns, the calculation of the memory needed for the whole
3237 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3238 back references of the form \12, where the number was greater than 9, were
3239 not recognized as back references, even though there were sufficient
3240 previous subpatterns.
3241
3242 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3243 versions of gcc, e.g. 2.95.4.
3244
3245
3246 Version 6.1 21-Jun-05
3247 ---------------------
3248
3249 1. There was one reference to the variable "posix" in pcretest.c that was not
3250 surrounded by "#if !defined NOPOSIX".
3251
3252 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3253 the cross-check on the old pcre_info() function, for the benefit of the
3254 cut-down version of PCRE that is currently imported into Exim.
3255
3256 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3257 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3258 patterns that start (?i) but I don't think that matters. The use of (?i) is
3259 just an example; this all applies to the other options as well.
3260
3261 4. Since libtool seems to echo the compile commands it is issuing, the output
3262 from "make" can be reduced a bit by putting "@" in front of each libtool
3263 compile command.
3264
3265 5. Patch from the folks at Google for configure.in to be a bit more thorough
3266 in checking for a suitable C++ installation before trying to compile the
3267 C++ stuff. This should fix a reported problem when a compiler was present,
3268 but no suitable headers.
3269
3270 6. The man pages all had just "PCRE" as their title. I have changed them to
3271 be the relevant file name. I have also arranged that these names are
3272 retained in the file doc/pcre.txt, which is a concatenation in text format
3273 of all the man pages except the little individual ones for each function.
3274
3275 7. The NON-UNIX-USE file had not been updated for the different set of source
3276 files that come with release 6. I also added a few comments about the C++
3277 wrapper.
3278
3279
3280 Version 6.0 07-Jun-05
3281 ---------------------
3282
3283 1. Some minor internal re-organization to help with my DFA experiments.
3284
3285 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3286 didn't matter for the library itself when fully configured, but did matter
3287 when compiling without UCP support, or within Exim, where the ucp files are
3288 not imported.
3289
3290 3. Refactoring of the library code to split up the various functions into
3291 different source modules. The addition of the new DFA matching code (see
3292 below) to a single monolithic source would have made it really too
3293 unwieldy, quite apart from causing all the code to be include in a
3294 statically linked application, when only some functions are used. This is
3295 relevant even without the DFA addition now that patterns can be compiled in
3296 one application and matched in another.
3297
3298 The downside of splitting up is that there have to be some external
3299 functions and data tables that are used internally in different modules of
3300 the library but which are not part of the API. These have all had their
3301 names changed to start with "_pcre_" so that they are unlikely to clash
3302 with other external names.
3303
3304 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3305 a different (DFA) algorithm. Although it is slower than the original
3306 function, it does have some advantages for certain types of matching
3307 problem.
3308
3309 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3310 including restarting after a partial match.
3311
3312 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3313 defined when compiling for Windows was sent to me. I have put it into the
3314 code, though I have no means of testing or verifying it.
3315
3316 7. Added the pcre_refcount() auxiliary function.
3317
3318 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3319 match before or at the first newline in the subject string. In pcretest,
3320 the /f option on a pattern can be used to set this.
3321
3322 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3323 would behave wrongly. This has been present in PCRE since release 4.0.
3324
3325 10. A number of changes to the pcregrep command:
3326
3327 (a) Refactored how -x works; insert ^(...)$ instead of setting
3328 PCRE_ANCHORED and checking the length, in preparation for adding
3329 something similar for -w.
3330
3331 (b) Added the -w (match as a word) option.
3332
3333 (c) Refactored the way lines are read and buffered so as to have more
3334 than one at a time available.
3335
3336 (d) Implemented a pcregrep test script.
3337
3338 (e) Added the -M (multiline match) option. This allows patterns to match
3339 over several lines of the subject. The buffering ensures that at least
3340 8K, or the rest of the document (whichever is the shorter) is available
3341 for matching (and similarly the previous 8K for lookbehind assertions).
3342
3343 (f) Changed the --help output so that it now says
3344
3345 -w, --word-regex(p)
3346
3347 instead of two lines, one with "regex" and the other with "regexp"
3348 because that confused at least one person since the short forms are the
3349 same. (This required a bit of code, as the output is generated
3350 automatically from a table. It wasn't just a text change.)
3351
3352 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3353 option but starts with a hyphen. Could be a pattern or a path name
3354 starting with a hyphen, for instance.
3355
3356 (h) "-" can be given as a file name to represent stdin.
3357
3358 (i) When file names are being printed, "(standard input)" is used for
3359 the standard input, for compatibility with GNU grep. Previously
3360 "<stdin>" was used.
3361
3362 (j) The option --label=xxx can be used to supply a name to be used for
3363 stdin when file names are being printed. There is no short form.
3364
3365 (k) Re-factored the options decoding logic because we are going to add
3366 two more options that take data. Such options can now be given in four
3367 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3368
3369 (l) Added the -A, -B, and -C options for requesting that lines of context
3370 around matches be printed.
3371
3372 (m) Added the -L option to print the names of files that do not contain
3373 any matching lines, that is, the complement of -l.
3374
3375 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3376 continue to scan other files.
3377
3378 (o) The -s option was incorrectly implemented. For compatibility with other
3379 greps, it now suppresses the error message for a non-existent or non-
3380 accessible file (but not the return code). There is a new option called
3381 -q that suppresses the output of matching lines, which was what -s was
3382 previously doing.
3383
3384 (p) Added --include and --exclude options to specify files for inclusion
3385 and exclusion when recursing.
3386
3387 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3388 Hopefully, it now does.
3389
3390 12. Missing cast in pcre_study().
3391
3392 13. Added an "uninstall" target to the makefile.
3393
3394 14. Replaced "extern" in the function prototypes in Makefile.in with
3395 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3396 world, but is set differently for Windows.
3397
3398 15. Added a second compiling function called pcre_compile2(). The only
3399 difference is that it has an extra argument, which is a pointer to an
3400 integer error code. When there is a compile-time failure, this is set
3401 non-zero, in addition to the error test pointer being set to point to an
3402 error message. The new argument may be NULL if no error number is required
3403 (but then you may as well call pcre_compile(), which is now just a
3404 wrapper). This facility is provided because some applications need a
3405 numeric error indication, but it has also enabled me to tidy up the way
3406 compile-time errors are handled in the POSIX wrapper.
3407
3408 16. Added VPATH=.libs to the makefile; this should help when building with one
3409 prefix path and installing with another. (Or so I'm told by someone who
3410 knows more about this stuff than I do.)
3411
3412 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3413 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3414 match everything, including newlines. This is not POSIX-compatible, but
3415 somebody wanted the feature. From pcretest it can be activated by using
3416 both the P and the s flags.
3417
3418 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3419
3420 19. libpcre.pc was being incorrectly installed as executable.
3421
3422 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3423 it now also looks for '\r' so that it will work unmodified on Windows.
3424
3425 21. Added Google's contributed C++ wrapper to the distribution.
3426
3427 22. Added some untidy missing memory free() calls in pcretest, to keep
3428 Electric Fence happy when testing.
3429
3430
3431
3432 Version 5.0 13-Sep-04
3433 ---------------------
3434
3435 1. Internal change: literal characters are no longer packed up into items
3436 containing multiple characters in a single byte-string. Each character
3437 is now matched using a separate opcode. However, there may be more than one
3438 byte in the character in UTF-8 mode.
3439
3440 2. The pcre_callout_block structure has two new fields: pattern_position and
3441 next_item_length. These contain the offset in the pattern to the next match
3442 item, and its length, respectively.
3443
3444 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3445 insertion of callouts before each pattern item. Added the /C option to
3446 pcretest to make use of this.
3447
3448 4. On the advice of a Windows user, the lines
3449
3450 #if defined(_WIN32) || defined(WIN32)
3451 _setmode( _fileno( stdout ), 0x8000 );
3452 #endif /* defined(_WIN32) || defined(WIN32) */
3453
3454 have been added to the source of pcretest. This apparently does useful
3455 magic in relation to line terminators.
3456
3457 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3458 for the benefit of those environments where the "b" makes a difference.
3459
3460 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3461 to know about it. I have put a hack into configure.in that adds in code
3462 to set GCC=yes if CC=icc. This seems to end up at a point in the
3463 generated configure script that is early enough to affect the setting of
3464 compiler options, which is what is needed, but I have no means of testing
3465 whether it really works. (The user who reported this had patched the
3466 generated configure script, which of course I cannot do.)
3467
3468 LATER: After change 22 below (new libtool files), the configure script
3469 seems to know about icc (and also ecc). Therefore, I have commented out
3470 this hack in configure.in.
3471
3472 7. Added support for pkg-config (2 patches were sent in).
3473
3474 8. Negated POSIX character classes that used a combination of internal tables
3475 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3476 [[:^ascii]]. Typically, they would match almost any characters. The other
3477 POSIX classes were not broken in this way.
3478
3479 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3480 to find the match, as PCRE was deluded into thinking that the match had to
3481 start at the start point or following a newline. The same bug applied to
3482 patterns with negative forward assertions or any backward assertions
3483 preceding ".*" at the start, unless the pattern required a fixed first
3484 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3485
3486 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3487 starting at the last subject character, bytes beyond the end of the subject
3488 string were read.
3489
3490 11. Renamed the variable "class" as "classbits" to make life easier for C++
3491 users. (Previously there was a macro definition, but it apparently wasn't
3492 enough.)
3493
3494 12. Added the new field "tables" to the extra data so that tables can be passed
3495 in at exec time, or the internal tables can be re-selected. This allows
3496 a compiled regex to be saved and re-used at a later time by a different
3497 program that might have everything at different addresses.
3498
3499 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3500 -R library as well as a -L library.
3501
3502 14. The debugging options of pcretest (-d on the command line or D on a
3503 pattern) showed incorrect output for anything following an extended class
3504 that contained multibyte characters and which was followed by a quantifier.
3505
3506 15. Added optional support for general category Unicode character properties
3507 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3508 support. It adds about 90K to the size of the library. The meanings of the
3509 inbuilt class escapes such as \d and \s have NOT been changed.
3510
3511 16. Updated pcredemo.c to include calls to free() to release the memory for the
3512 compiled pattern.
3513
3514 17. The generated file chartables.c was being created in the source directory
3515 instead of in the building directory. This caused the build to fail if the
3516 source directory was different from the building directory, and was
3517 read-only.
3518
3519 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3520 file. No doubt somebody will tell me if they don't make sense... Also added
3521 Dan Mooney's comments about building on OpenVMS.
3522
3523 19. Added support for partial matching via the PCRE_PARTIAL option for
3524 pcre_exec() and the \P data escape in pcretest.
3525
3526 20. Extended pcretest with 3 new pattern features:
3527
3528 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3529 write the compiled pattern to the file whose name is "rest-of-line".
3530 This is a straight binary dump of the data, with the saved pointer to
3531 the character tables forced to be NULL. The study data, if any, is
3532 written too. After writing, pcretest reads a new pattern.
3533
3534 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3535 compiled pattern from the given file. There must not be any
3536 occurrences of "<" in the file name (pretty unlikely); if there are,
3537 pcretest will instead treat the initial "<" as a pattern delimiter.
3538 After reading in the pattern, pcretest goes on to read data lines as
3539 usual.
3540
3541 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3542 and 16-bit fields in a compiled pattern, to simulate a pattern that
3543 was compiled on a host of opposite endianness.
3544
3545 21. The pcre-exec() function can now cope with patterns that were compiled on
3546 hosts of opposite endianness, with this restriction:
3547
3548 As for any compiled expression that is saved and used later, the tables
3549 pointer field cannot be preserved; the extra_data field in the arguments
3550 to pcre_exec() should be used to pass in a tables address if a value
3551 other than the default internal tables were used at compile time.
3552
3553 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3554 now diagnosed as an error. Previously, most of the time, a negative number
3555 would have been treated as zero, but if in addition "ovector" was passed as
3556 NULL, a crash could occur.
3557
3558 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3559 new versions from the libtool 1.5 distribution (the last one is a copy of
3560 a file called libtool.m4). This seems to have fixed the need to patch
3561 "configure" to support Darwin 1.3 (which I used to do). However, I still
3562 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3563 workstation).
3564
3565 24. Changed the PCRE licence to be the more standard "BSD" licence.
3566
3567
3568 Version 4.5 01-Dec-03
3569 ---------------------
3570
3571 1. There has been some re-arrangement of the code for the match() function so
3572 that it can be compiled in a version that does not call itself recursively.
3573 Instead, it keeps those local variables that need separate instances for
3574 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3575 needs to "recurse". Keeping track of where control must go is done by means
3576 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3577 hide most of the details from the main code, and operates only if
3578 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3579 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3580 operating.
3581
3582 To make it easier for callers to provide specially tailored get/free
3583 functions for this usage, two new functions, pcre_stack_malloc, and
3584 pcre_stack_free, are used. They are always called in strict stacking order,
3585 and the size of block requested is always the same.
3586
3587 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3588 PCRE has been compiled to use the stack or the heap for recursion. The
3589 -C option of pcretest uses this to show which version is compiled.
3590
3591 A new data escape \S, is added to pcretest; it causes the amounts of store
3592 obtained and freed by both kinds of malloc/free at match time to be added
3593 to the output.
3594
3595 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3596 what's available on my current Linux desktop machine.
3597
3598 3. When matching a UTF-8 string, the test for a valid string at the start has
3599 been extended. If start_offset is not zero, PCRE now checks that it points
3600 to a byte that is the start of a UTF-8 character. If not, it returns
3601 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3602 this is necessary because there may be backward assertions in the pattern.
3603 When matching the same subject several times, it may save resources to use
3604 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3605
3606 4. The code for checking the validity of UTF-8 strings has been tightened so
3607 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3608 containing "overlong sequences".
3609
3610 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3611 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3612 should have been "&", but it just so happened that all the cases this let
3613 through by mistake were picked up later in the function.
3614
3615 6. I had used a variable called "isblank" - this is a C99 function, causing
3616 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3617
3618 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3619 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3620 so the version is shown; (c) stop "make test" echoing "./RunTest".
3621
3622 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3623
3624 9. The prototype for memmove() for systems that don't have it was using
3625 size_t, but the inclusion of the header that defines size_t was later. I've
3626 moved the #includes for the C headers earlier to avoid this.
3627
3628 10. Added some adjustments to the code to make it easier to compiler on certain
3629 special systems:
3630
3631 (a) Some "const" qualifiers were missing.
3632 (b) Added the macro EXPORT before all exported functions; by default this
3633 is defined to be empty.
3634 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3635 that it reads its output file name as an argument instead of writing
3636 to the standard output and assuming this can be redirected.
3637
3638 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3639 class containing characters with values greater than 255, PCRE compilation
3640 went into a loop.
3641
3642 12. A recursive reference to a subpattern that was within another subpattern
3643 that had a minimum quantifier of zero caused PCRE to crash. For example,
3644 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3645 recursion. If the recursively-called subpattern itself had a zero repeat,
3646 that was OK.
3647
3648 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3649 buffer into which it was copied (for escape processing) was still set at
3650 1024, so long lines caused crashes.
3651
3652 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3653 "internal error: code overflow...". This applied to any character class
3654 that was followed by a possessive quantifier.
3655
3656 15. Modified the Makefile to add libpcre.la as a prerequisite for
3657 libpcreposix.la because I was told this is needed for a parallel build to
3658 work.
3659
3660 16. If a pattern that contained .* following optional items at the start was
3661 studied, the wrong optimizing data was generated, leading to matching
3662 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3663 matching string must start with a or b or c. The correct conclusion for
3664 this pattern is that a match can start with any character.
3665
3666
3667 Version 4.4 13-Aug-03
3668 ---------------------
3669
3670 1. In UTF-8 mode, a character class containing characters with values between
3671 127 and 255 was not handled correctly if the compiled pattern was studied.
3672 In fixing this, I have also improved the studying algorithm for such
3673 classes (slightly).
3674
3675 2. Three internal functions had redundant arguments passed to them. Removal
3676 might give a very teeny performance improvement.
3677
3678 3. Documentation bug: the value of the capture_top field in a callout is *one
3679 more than* the number of the hightest numbered captured substring.
3680
3681 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3682 in incorrectly linking with a previously installed version. They now link
3683 explicitly with libpcre.la.
3684
3685 5. configure.in no longer needs to recognize Cygwin specially.
3686
3687 6. A problem in pcre.in for Windows platforms is fixed.
3688
3689 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3690 pcretest, it used to include the size of the study block as part of its
3691 output. Unfortunately, the structure contains a field that has a different
3692 size on different hardware architectures. This meant that the tests that
3693 showed this size failed. As the block is currently always of a fixed size,
3694 this information isn't actually particularly useful in pcretest output, so
3695 I have just removed it.
3696
3697 8. Three pre-processor statements accidentally did not start in column 1.
3698 Sadly, there are *still* compilers around that complain, even though
3699 standard C has not required this for well over a decade. Sigh.
3700
3701 9. In pcretest, the code for checking callouts passed small integers in the
3702 callout_data field, which is a void * field. However, some picky compilers
3703 complained about the casts involved for this on 64-bit systems. Now
3704 pcretest passes the address of the small integer instead, which should get
3705 rid of the warnings.
3706
3707 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3708 both compile and run time, and gives an error if an invalid UTF-8 sequence
3709 is found. There is a option for disabling this check in cases where the
3710 string is known to be correct and/or the maximum performance is wanted.
3711
3712 11. In response to a bug report, I changed one line in Makefile.in from
3713
3714 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3715 to
3716 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3717
3718 to look similar to other lines, but I have no way of telling whether this
3719 is the right thing to do, as I do not use Windows. No doubt I'll get told
3720 if it's wrong...
3721
3722
3723 Version 4.3 21-May-03
3724 ---------------------
3725
3726 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3727 Makefile.
3728
3729 2. Some refactoring to improve the quality of the code:
3730
3731 (i) The utf8_table... variables are now declared "const".
3732
3733 (ii) The code for \cx, which used the "case flipping" table to upper case
3734 lower case letters, now just substracts 32. This is ASCII-specific,
3735 but the whole concept of \cx is ASCII-specific, so it seems
3736 reasonable.
3737
3738 (iii) PCRE was using its character types table to recognize decimal and
3739 hexadecimal digits in the pattern. This is silly, because it handles
3740 only 0-9, a-f, and A-F, but the character types table is locale-
3741 specific, which means strange things might happen. A private
3742 table is now used for this - though it costs 256 bytes, a table is
3743 much faster than multiple explicit tests. Of course, the standard
3744 character types table is still used for matching digits in subject
3745 strings against \d.
3746
3747 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3748 ending in _t are). So I've renamed it as ESC_tee.
3749
3750 3. The first argument for regexec() in the POSIX wrapper should have been
3751 defined as "const".
3752
3753 4. Changed pcretest to use malloc() for its buffers so that they can be
3754 Electric Fenced for debugging.
3755
3756 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3757 to read one or more bytes before the start of the subject string. Often this
3758 had no effect on PCRE's behaviour, but in some circumstances it could
3759 provoke a segmentation fault.
3760
3761 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3762 to try to read one or more bytes before the start of the subject string.
3763
3764 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3765 UTF-8 support could misbehave in various ways if the subject string
3766 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3767 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3768 back over UTF-8 characters.)
3769
3770
3771 Version 4.2 14-Apr-03
3772 ---------------------
3773
3774 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3775
3776 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3777 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3778 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3779 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3780 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3781 and BUILD_EXEEXT
3782 Note: automatic setting of the BUILD variables is not yet working
3783 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3784 compile-time but not at link-time
3785 [LINK]: use for linking executables only
3786 make different versions for Windows and non-Windows
3787 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3788 libraries
3789 [LINK_FOR_BUILD]: new variable
3790 [OBJEXT]: use throughout
3791 [EXEEXT]: use throughout
3792 <winshared>: new target
3793 <wininstall>: new target
3794 <dftables.o>: use native compiler
3795 <dftables>: use native linker
3796 <install>: handle Windows platform correctly
3797 <clean>: ditto
3798 <check>: ditto
3799 copy DLL to top builddir before testing
3800
3801 As part of these changes, -no-undefined was removed again. This was reported
3802 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3803 in any case.
3804
3805 3. Some tidies to get rid of compiler warnings:
3806
3807 . In the match_data structure, match_limit was an unsigned long int, whereas
3808 match_call_count was an int. I've made them both unsigned long ints.
3809
3810 . In pcretest the fact that a const uschar * doesn't automatically cast to
3811 a void * provoked a warning.
3812
3813 . Turning on some more compiler warnings threw up some "shadow" variables
3814 and a few more missing casts.
3815
3816 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3817 option, a class that contained a single character with a value between 128
3818 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3819
3820 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3821 option, a class that contained several characters, but with at least one
3822 whose value was between 128 and 255 caused PCRE to crash.
3823
3824
3825 Version 4.1 12-Mar-03
3826 ---------------------
3827
3828 1. Compiling with gcc -pedantic found a couple of places where casts were
3829 needed, and a string in dftables.c that was longer than standard compilers are
3830 required to support.
3831
3832 2. Compiling with Sun's compiler found a few more places where the code could
3833 be tidied up in order to avoid warnings.
3834
3835 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3836 first of these names is deprecated in the latest Autoconf in favour of the name
3837 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3838 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3839 analogy I have changed it to CFLAGS_FOR_BUILD.
3840
3841 4. Added -no-undefined to the linking command in the Makefile, because this is
3842 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3843 linking step for the pcreposix library.
3844
3845 5. PCRE was failing to diagnose the case of two named groups with the same
3846 name.
3847
3848 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3849 literal character that is needed in the subject for a match, and scans along to
3850 ensure that it is present before embarking on the full matching process. This
3851 saves time in cases of nested unlimited repeats that are never going to match.
3852 Problem: the scan can take a lot of time if the subject is very long (e.g.
3853 megabytes), thus penalizing straightforward matches. It is now done only if the
3854 amount of subject to be scanned is less than 1000 bytes.
3855
3856 7. A lesser problem with the same optimization is that it was recording the
3857 first character of an anchored pattern as "needed", thus provoking a search
3858 right along the subject, even when the first match of the pattern was going to
3859 fail. The "needed" character is now not set for anchored patterns, unless it
3860 follows something in the pattern that is of non-fixed length. Thus, it still
3861 fulfils its original purpose of finding quick non-matches in cases of nested
3862 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3863
3864
3865 Version 4.0 17-Feb-03
3866 ---------------------
3867
3868 1. If a comment in an extended regex that started immediately after a meta-item
3869 extended to the end of string, PCRE compiled incorrect data. This could lead to
3870 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3871
3872 2. Moved to autoconf 2.53 and libtool 1.4.2.
3873
3874 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3875 the special perltest8 script is no longer needed - all the tests can be run
3876 from a single perltest script.
3877
3878 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3879 by \s. It has now been removed in PCRE. This means it isn't recognized as
3880 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3881 class [:space:] *does* include VT, thereby creating a mess.
3882
3883 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3884 space and tab.
3885
3886 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3887 its new features into the main test script, reducing the number of scripts.
3888
3889 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3890 were backward compatible, and made the (?i) apply to the whole pattern, as if
3891 /i were given. Now it behaves more logically, and applies the option setting
3892 only to what follows. PCRE has been changed to follow suit. However, if it
3893 finds options settings right at the start of the pattern, it extracts them into
3894 the global options, as before. Thus, they show up in the info data.
3895
3896 8. Added support for the \Q...\E escape sequence. Characters in between are
3897 treated as literals. This is slightly different from Perl in that $ and @ are
3898 also handled as literals inside the quotes. In Perl, they will cause variable
3899 interpolation. Note the following examples:
3900
3901 Pattern PCRE matches Perl matches
3902
3903 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3904 \Qabc\$xyz\E abc\$xyz abc\$xyz
3905 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3906
3907 For compatibility with Perl, \Q...\E sequences are recognized inside character
3908 classes as well as outside them.
3909
3910 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3911 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3912 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3913 signed/unsigned warnings.
3914
3915 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3916 option for pcretest, so I've replaced it by a simple function that does just
3917 that job.
3918
3919 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3920 "pcregrep -".
3921
3922 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3923 Java package. This provides some syntactic sugar for simple cases of what my
3924 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3925 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3926 item, you can use this simplified notation. Note that only makes sense with
3927 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3928 greediness, whatever the setting of the PCRE_UNGREEDY option.
3929
3930 13. A change of greediness default within a pattern was not taking effect at
3931 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3932 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3933 was abstracted outside.
3934
3935 14. PCRE now supports the \G assertion. It is true when the current matching
3936 position is at the start point of the match. This differs from \A when the
3937 starting offset is non-zero. Used with the /g option of pcretest (or similar
3938 code), it works in the same way as it does for Perl's /g option. If all
3939 alternatives of a regex begin with \G, the expression is anchored to the start
3940 match position, and the "anchored" flag is set in the compiled expression.
3941
3942 15. Some bugs concerning the handling of certain option changes within patterns
3943 have been fixed. These applied to options other than (?ims). For example,
3944 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3945 been the other way round. Some of this was related to change 7 above.
3946
3947 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3948 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3949 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3950 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3951
3952 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3953 mode. Unlike ".", it always matches newline, whatever the setting of
3954 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3955 assertions. Perl allows it, but it doesn't (in general) work because it can't
3956 calculate the length of the lookbehind. At least, that's the case for Perl
3957 5.8.0 - I've been told they are going to document that it doesn't work in
3958 future.
3959
3960 18. Added an error diagnosis for escapes that PCRE does not support: these are
3961 \L, \l, \N, \P, \p, \U, \u, and \X.
3962
3963 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3964 reading past the end of the pattern in cases such as /[abcd/.
3965
3966 20. PCRE was getting more memory than necessary for patterns with classes that
3967 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3968
3969 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3970 compiling PCRE for use with Virtual Pascal.
3971
3972 22. Small fix to the Makefile to make it work properly if the build is done
3973 outside the source tree.
3974
3975 23. Added a new extension: a condition to go with recursion. If a conditional
3976 subpattern starts with (?(R) the "true" branch is used if recursion has
3977 happened, whereas the "false" branch is used only at the top level.
3978
3979 24. When there was a very long string of literal characters (over 255 bytes
3980 without UTF support, over 250 bytes with UTF support), the computation of how
3981 much memory was required could be incorrect, leading to segfaults or other
3982 strange effects.
3983
3984 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3985 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3986 there was a subsequent back reference to those brackets. This meant that, for
3987 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3988 possible to check for precisely this case. All we can do is abandon the
3989 optimization if .* occurs inside capturing brackets when there are any back
3990 references whatsoever. (See below for a better fix that came later.)
3991
3992 26. The handling of the optimization for finding the first character of a
3993 non-anchored pattern, and for finding a character that is required later in the
3994 match were failing in some cases. This didn't break the matching; it just
3995 failed to optimize when it could. The way this is done has been re-implemented.
3996
3997 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3998
3999 28. Added a new feature that provides some of the functionality that Perl
4000 provides with (?{...}). The facility is termed a "callout". The way it is done
4001 in PCRE is for the caller to provide an optional function, by setting
4002 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4003 global variable. By default it is unset, which disables all calling out. To get
4004 the function called, the regex must include (?C) at appropriate points. This
4005 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4006 This provides a means of identifying different callout points. When PCRE
4007 reaches such a point in the regex, if pcre_callout has been set, the external
4008 function is called. It is provided with data in a structure called
4009 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4010 matching continues; if it returns a non-zero value, the match at the current
4011 point fails. However, backtracking will occur if possible. [This was changed
4012 later and other features added - see item 49 below.]
4013
4014 29. pcretest is upgraded to test the callout functionality. It provides a
4015 callout function that displays information. By default, it shows the start of
4016 the match and the current position in the text. There are some new data escapes
4017 to vary what happens:
4018
4019 \C+ in addition, show current contents of captured substrings
4020 \C- do not supply a callout function
4021 \C!n return 1 when callout number n is reached
4022 \C!n!m return 1 when callout number n is reached for the mth time
4023
4024 30. If pcregrep was called with the -l option and just a single file name, it
4025 output "<stdin>" if a match was found, instead of the file name.
4026
4027 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4028 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4029 pcre_exec(). This saves a malloc/free per call. The default value of
4030 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4031 when configuring.
4032
4033 32. The default maximum size of a compiled pattern is 64K. There have been a
4034 few cases of people hitting this limit. The code now uses macros to handle the
4035 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4036 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4037 configuring. Tests 2 and 5 work only with 2-byte links because they output
4038 debugging information about compiled patterns.
4039
4040 33. Internal code re-arrangements:
4041
4042 (a) Moved the debugging function for printing out a compiled regex into
4043 its own source file (printint.c) and used #include to pull it into
4044 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4045 separate copies.
4046
4047 (b) Defined the list of op-code names for debugging as a macro in
4048 internal.h so that it is next to the definition of the opcodes.
4049
4050 (c) Defined a table of op-code lengths for simpler skipping along compiled
4051 code. This is again a macro in internal.h so that it is next to the
4052 definition of the opcodes.
4053
4054 34. Added support for recursive calls to individual subpatterns, along the
4055 lines of Robin Houston's patch (but implemented somewhat differently).
4056
4057 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4058 allow it to read and process whole directories in Win32. This code was
4059 contributed by Lionel Fourquaux; it has not been tested by me.
4060
4061 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4062 used to name a group. Names consist of alphanumerics and underscores, and must
4063 be unique. Back references use the syntax (?P=name) and recursive calls use
4064 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4065 numbers. The function pcre_fullinfo() can be used after compilation to extract
4066 a name/number map. There are three relevant calls:
4067
4068 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4069 PCRE_INFO_NAMECOUNT yields the number of entries
4070 PCRE_INFO_NAMETABLE yields a pointer to the map.
4071
4072 The map is a vector of fixed-size entries. The size of each entry depends on
4073 the length of the longest name used. The first two bytes of each entry are the
4074 group number, most significant byte first. There follows the corresponding
4075 name, zero terminated. The names are in alphabetical order.
4076
4077 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4078 case instead of 255. Making it the same both with and without UTF-8 support
4079 means that the same test output works with both.
4080
4081 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4082 calling malloc() with a zero argument.
4083
4084 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4085 optimization. I've improved things by keeping a bitmap of backreferences with
4086 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4087 fact referenced, the optimization can be applied. It is unlikely that a
4088 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4089 the match to follow \n) will appear inside brackets with a number greater than
4090 31, but if it does, any back reference > 31 suppresses the optimization.
4091
4092 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4093 of disabling numbered capturing parentheses. Any opening parenthesis that is
4094 not followed by ? behaves as if it were followed by ?: but named parentheses
4095 can still be used for capturing (and they will acquire numbers in the usual
4096 way).
4097
4098 41. Redesigned the return codes from the match() function into yes/no/error so
4099 that errors can be passed back from deep inside the nested calls. A malloc
4100 failure while inside a recursive subpattern call now causes the
4101 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4102
4103 42. It is now possible to set a limit on the number of times the match()
4104 function is called in a call to pcre_exec(). This facility makes it possible to
4105 limit the amount of recursion and backtracking, though not in a directly
4106 obvious way, because the match() function is used in a number of different
4107 circumstances. The count starts from zero for each position in the subject
4108 string (for non-anchored patterns). The default limit is, for compatibility, a
4109 large number, namely 10 000 000. You can change this in two ways:
4110
4111 (a) When configuring PCRE before making, you can use --with-match-limit=n
4112 to set a default value for the compiled library.
4113
4114 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4115 a different value is set. See 45 below.
4116
4117 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4118
4119 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4120 of things that can be changed at compile time. The first argument specifies
4121 what is wanted and the second points to where the information is to be placed.
4122 The current list of available information is:
4123
4124 PCRE_CONFIG_UTF8
4125
4126 The output is an integer that is set to one if UTF-8 support is available;
4127 otherwise it is set to zero.
4128
4129 PCRE_CONFIG_NEWLINE
4130
4131 The output is an integer that it set to the value of the code that is used for
4132 newline. It is either LF (10) or CR (13).
4133
4134 PCRE_CONFIG_LINK_SIZE
4135
4136 The output is an integer that contains the number of bytes used for internal
4137 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4138
4139 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4140
4141 The output is an integer that contains the threshold above which the POSIX
4142 interface uses malloc() for output vectors. See item 31 above.
4143
4144 PCRE_CONFIG_MATCH_LIMIT
4145
4146 The output is an unsigned integer that contains the default limit of the number
4147 of match() calls in a pcre_exec() execution. See 42 above.
4148
4149 44. pcretest has been upgraded by the addition of the -C option. This causes it
4150 to extract all the available output from the new pcre_config() function, and to
4151 output it. The program then exits immediately.
4152
4153 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4154 order to support additional features. One way would have been to define
4155 pcre_exec2() (for example) with extra arguments, but this would not have been
4156 extensible, and would also have required all calls to the original function to
4157 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4158 is used for passing in "extra" data from pcre_study().
4159
4160 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4161 contains the following fields:
4162
4163 flags a bitmap indicating which of the following fields are set
4164 study_data opaque data from pcre_study()
4165 match_limit a way of specifying a limit on match() calls for a specific
4166 call to pcre_exec()
4167 callout_data data for callouts (see 49 below)
4168
4169 The flag bits are also defined in pcre.h, and are
4170
4171 PCRE_EXTRA_STUDY_DATA
4172 PCRE_EXTRA_MATCH_LIMIT
4173 PCRE_EXTRA_CALLOUT_DATA
4174
4175 The pcre_study() function now returns one of these new pcre_extra blocks, with
4176 the actual study data pointed to by the study_data field, and the
4177 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4178 before. That is, this change is entirely upwards-compatible and requires no
4179 change to existing code.
4180
4181 If you want to pass in additional data to pcre_exec(), you can either place it
4182 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4183 block.
4184
4185 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4186 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4187 times with different match limits, until it finds the minimum value needed for
4188 pcre_exec() to complete. The value is then output. This can be instructive; for
4189 most simple matches the number is quite small, but for pathological cases it
4190 gets very large very quickly.
4191
4192 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4193 returns the size of the data block pointed to by the study_data field in a
4194 pcre_extra block, that is, the value that was passed as the argument to
4195 pcre_malloc() when PCRE was getting memory in which to place the information
4196 created by pcre_study(). The fourth argument should point to a size_t variable.
4197 pcretest has been extended so that this information is shown after a successful
4198 pcre_study() call when information about the compiled regex is being displayed.
4199
4200 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4201 because what follows is always an absolute path. (Later: it turns out that this
4202 is more than cosmetic for MinGW, because it doesn't like empty path
4203 components.)
4204
4205 49. Some changes have been made to the callout feature (see 28 above):
4206
4207 (i) A callout function now has three choices for what it returns:
4208
4209 0 => success, carry on matching
4210 > 0 => failure at this point, but backtrack if possible
4211 < 0 => serious error, return this value from pcre_exec()
4212
4213 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4214 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4215 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4216 use by callout functions. It will never be used by PCRE itself.
4217
4218 (ii) The pcre_extra structure (see 45 above) has a void * field called
4219 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4220 pcre_callout_block structure has a field of the same name. The contents of
4221 the field passed in the pcre_extra structure are passed to the callout
4222 function in the corresponding field in the callout block. This makes it
4223 easier to use the same callout-containing regex from multiple threads. For
4224 testing, the pcretest program has a new data escape
4225
4226 \C*n pass the number n (may be negative) as callout_data
4227
4228 If the callout function in pcretest receives a non-zero value as
4229 callout_data, it returns that value.
4230
4231 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4232 there were some redundant $(CFLAGS) in commands that are now specified as
4233 $(LINK), which already includes $(CFLAGS).
4234
4235 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4236 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4237 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4238 one-byte characters throughout. Note that case-insensitive matching applies
4239 only to characters whose values are less than 256. PCRE doesn't support the
4240 notion of cases for higher-valued characters.
4241
4242 (i) A character class whose characters are all within 0-255 is handled as
4243 a bit map, and the map is inverted for negative classes. Previously, a
4244 character > 255 always failed to match such a class; however it should
4245 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4246
4247 (ii) A negated character class with a single character < 255 is coded as
4248 "not this character" (OP_NOT). This wasn't working properly when the test
4249 character was multibyte, either singly or repeated.
4250
4251 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4252 mode, for example: \x{100}{2,3}.
4253
4254 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4255 singly or repeated) now correctly test multibyte characters. However,
4256 PCRE doesn't recognize any characters with values greater than 255 as
4257 digits, spaces, or word characters. Such characters always match \D, \S,
4258 and \W, and never match \d, \s, or \w.
4259
4260 (v) Classes may now contain characters and character ranges with values
4261 greater than 255. For example: [ab\x{100}-\x{400}].
4262
4263 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4264 PCRE in UTF-8 mode.
4265
4266 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4267 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4268 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4269 value.)
4270
4271 53. The single man page has become too large. I have therefore split it up into
4272 a number of separate man pages. These also give rise to individual HTML pages;
4273 these are now put in a separate directory, and there is an index.html page that
4274 lists them all. Some hyperlinking between the pages has been installed.
4275
4276 54. Added convenience functions for handling named capturing parentheses.
4277
4278 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4279 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4280 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4281 are faulted.
4282
4283 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4284 calling configure. These values are used when compiling the dftables.c program
4285 which is run to generate the source of the default character tables. They
4286 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4287 you will need to set these values.
4288
4289 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4290
4291
4292 Version 3.9 02-Jan-02
4293 ---------------------
4294
4295 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4296
4297 2. If --disable-static was given, the building process failed when trying to
4298 build pcretest and pcregrep. (For some reason it was using libtool to compile
4299 them, which is not right, as they aren't part of the library.)
4300
4301
4302 Version 3.8 18-Dec-01
4303 ---------------------
4304
4305 1. The experimental UTF-8 code was completely screwed up. It was packing the
4306 bytes in the wrong order. How dumb can you get?
4307
4308
4309 Version 3.7 29-Oct-01
4310 ---------------------
4311
4312 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4313 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4314 this didn't happen under Solaris 8, where I normally test things.
4315
4316 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4317 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4318 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4319 this fix an hour or so after the initial 3.7 release.)
4320
4321
4322 Version 3.6 23-Oct-01
4323 ---------------------
4324
4325 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4326 offsets passed as NULL with zero offset count.
4327
4328 2. The config.guess and config.sub files had not been updated when I moved to
4329 the latest autoconf.
4330
4331
4332 Version 3.5 15-Aug-01
4333 ---------------------
4334
4335 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4336 had been forgotten.
4337
4338 2. By using declared but undefined structures, we can avoid using "void"
4339 definitions in pcre.h while keeping the internal definitions of the structures
4340 private.
4341
4342 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4343 user point of view, this means that both static and shared libraries are built
4344 by default, but this can be individually controlled. More of the work of
4345 handling this static/shared cases is now inside libtool instead of PCRE's make
4346 file.
4347
4348 4. The pcretest utility is now installed along with pcregrep because it is
4349 useful for users (to test regexs) and by doing this, it automatically gets
4350 relinked by libtool. The documentation has been turned into a man page, so
4351 there are now .1, .txt, and .html versions in /doc.
4352
4353 5. Upgrades to pcregrep:
4354 (i) Added long-form option names like gnu grep.
4355 (ii) Added --help to list all options with an explanatory phrase.
4356 (iii) Added -r, --recursive to recurse into sub-directories.
4357 (iv) Added -f, --file to read patterns from a file.
4358
4359 6. pcre_exec() was referring to its "code" argument before testing that
4360 argument for NULL (and giving an error if it was NULL).
4361
4362 7. Upgraded Makefile.in to allow for compiling in a different directory from
4363 the source directory.
4364
4365 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4366 options bits, the pointer it was passed was to an int instead of to an unsigned
4367 long int. This mattered only on 64-bit systems.
4368
4369 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4370 generated) instead of pcre.in, which it its source. Also made the same change
4371 in several of the .c files.
4372
4373 10. A new release of gcc defines printf() as a macro, which broke pcretest
4374 because it had an ifdef in the middle of a string argument for printf(). Fixed
4375 by using separate calls to printf().
4376
4377 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4378 script, to force use of CR or LF instead of \n in the source. On non-Unix
4379 systems, the value can be set in config.h.
4380
4381 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4382 absolute limit. Changed the text of the error message to make this clear, and
4383 likewise updated the man page.
4384
4385 13. The limit of 99 on the number of capturing subpatterns has been removed.
4386 The new limit is 65535, which I hope will not be a "real" limit.
4387
4388
4389 Version 3.4 22-Aug-00
4390 ---------------------
4391
4392 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4393
4394 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4395
4396
4397 Version 3.3 01-Aug-00
4398 ---------------------
4399
4400 1. If an octal character was given, but the value was greater than \377, it
4401 was not getting masked to the least significant bits, as documented. This could
4402 lead to crashes in some systems.
4403
4404 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4405 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4406
4407 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4408 These just pass their arguments on to (pcre_free)(), but they are provided
4409 because some uses of PCRE bind it to non-C systems that can call its functions,
4410 but cannot call free() or pcre_free() directly.
4411
4412 4. Add "make test" as a synonym for "make check". Corrected some comments in
4413 the Makefile.
4414
4415 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4416 Makefile.
4417
4418 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4419 command called pgrep for grepping around the active processes.
4420
4421 7. Added the beginnings of support for UTF-8 character strings.
4422
4423 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4424 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4425 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4426 out for the ar command.)
4427
4428
4429 Version 3.2 12-May-00
4430 ---------------------
4431
4432 This is purely a bug fixing release.
4433
4434 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4435 of ZA. This was just one example of several cases that could provoke this bug,
4436 which was introduced by change 9 of version 2.00. The code for breaking
4437 infinite loops after an iteration that matches an empty string was't working
4438 correctly.
4439
4440 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4441 when matched against abbab (for example). After matching an empty string, it
4442 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4443 caused it to match further down the string than it should.
4444
4445 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4446 was there because it doesn't seem to be needed, and it causes trouble on some
4447 systems, as it is not a Standard C header. It has been removed.
4448
4449 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4450 were reported on the Macintosh. The changes were from
4451
4452 while ((c = *(++ptr)) != 0 && c != '\n');
4453 to
4454 while ((c = *(++ptr)) != 0 && c != '\n') ;
4455
4456 Totally extraordinary, but if that's what it takes...
4457
4458 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4459 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4460 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4461 assumes the way PCRE uses memmove() (always moving upwards).
4462
4463 6. PCRE is being used in one environment where strchr() is not available. There
4464 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4465 faster code anyway.
4466
4467
4468 Version 3.1 09-Feb-00
4469 ---------------------
4470
4471 The only change in this release is the fixing of some bugs in Makefile.in for
4472 the "install" target:
4473
4474 (1) It was failing to install pcreposix.h.
4475
4476 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4477
4478
4479 Version 3.0 01-Feb-00
4480 ---------------------
4481
4482 1. Add support for the /+ modifier to perltest (to output $` like it does in
4483 pcretest).
4484
4485 2. Add support for the /g modifier to perltest.
4486
4487 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4488 matches null strings.
4489
4490 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4491 pattern. Perl treats empty patterns specially - it reuses the most recent
4492 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4493 effect.
4494
4495 5. The POSIX interface was broken in that it was just handing over the POSIX
4496 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4497 required a bigger vector, with some working space on the end. This means that
4498 the POSIX wrapper now has to get and free some memory, and copy the results.
4499
4500 6. Added some simple autoconf support, placing the test data and the
4501 documentation in separate directories, re-organizing some of the
4502 information files, and making it build pcre-config (a GNU standard). Also added
4503 libtool support for building PCRE as a shared library, which is now the
4504 default.
4505
4506 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4507 09 are not valid octal constants. Single digits will be used for minor values
4508 less than 10.
4509
4510 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4511 existing programs that set these in the POSIX interface can use PCRE without
4512 modification.
4513
4514 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4515 return all that pcre_info() returns, plus additional data. The pcre_info()
4516 function is retained for compatibility, but is considered to be obsolete.
4517
4518 10. Added experimental recursion feature (?R) to handle one common case that
4519 Perl 5.6 will be able to do with (?p{...}).
4520
4521 11. Added support for POSIX character classes like [:alpha:], which Perl is
4522 adopting.
4523
4524
4525 Version 2.08 31-Aug-99
4526 ----------------------
4527
4528 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4529 trying to match at the startoffset position, but instead was moving forward to
4530 the next newline as if a previous match had failed.
4531
4532 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4533 and could get into a loop if a null string was matched other than at the start
4534 of the subject.
4535
4536 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4537 be distinguished at compile time, and for completeness also added PCRE_DATE.
4538
4539 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4540 in GnuWin32 environments.
4541
4542
4543 Version 2.07 29-Jul-99
4544 ----------------------
4545
4546 1. The documentation is now supplied in plain text form and HTML as well as in
4547 the form of man page sources.
4548
4549 2. C++ compilers don't like assigning (void *) values to other pointer types.
4550 In particular this affects malloc(). Although there is no problem in Standard
4551 C, I've put in casts to keep C++ compilers happy.
4552
4553 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4554 should be (const char *).
4555
4556 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4557 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4558 However, I haven't made this a standard facility. The documentation doesn't
4559 mention it, and the Makefile doesn't support it.
4560
4561 5. The Makefile now contains an "install" target, with editable destinations at
4562 the top of the file. The pcretest program is not installed.
4563
4564 6. pgrep -V now gives the PCRE version number and date.
4565
4566 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4567 causing the entire string to be ignored, instead of just the last character.
4568
4569 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4570 non-matching string, it can take a very, very long time, even for strings of
4571 quite modest length, because of the nested recursion. PCRE now does better in
4572 some of these cases. It does this by remembering the last required literal
4573 character in the pattern, and pre-searching the subject to ensure it is present
4574 before running the real match. In other words, it applies a heuristic to detect
4575 some types of certain failure quickly, and in the above example, if presented
4576 with a string that has no trailing " it gives "no match" very quickly.
4577
4578 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4579 other alternatives are tried instead.
4580
4581
4582 Version 2.06 09-Jun-99
4583 ----------------------
4584
4585 1. Change pcretest's output for amount of store used to show just the code
4586 space, because the remainder (the data block) varies in size between 32-bit and
4587 64-bit systems.
4588
4589 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4590 start matching at. This allows lookbehinds to work when searching for multiple
4591 occurrences in a string.
4592
4593 3. Added additional options to pcretest for testing multiple occurrences:
4594
4595 /+ outputs the rest of the string that follows a match
4596 /g loops for multiple occurrences, using the new startoffset argument
4597 /G loops for multiple occurrences by passing an incremented pointer
4598
4599 4. PCRE wasn't doing the "first character" optimization for patterns starting
4600 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4601 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4602 the letter 'x'. On long subject strings, this gives a significant speed-up.
4603
4604
4605 Version 2.05 21-Apr-99
4606 ----------------------
4607
4608 1. Changed the type of magic_number from int to long int so that it works
4609 properly on 16-bit systems.
4610
4611 2. Fixed a bug which caused patterns starting with .* not to work correctly
4612 when the subject string contained newline characters. PCRE was assuming
4613 anchoring for such patterns in all cases, which is not correct because .* will
4614 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4615 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4616 must be retried after every newline in the subject.
4617
4618
4619 Version 2.04 18-Feb-99
4620 ----------------------
4621
4622 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4623 computation of the store needed to hold the pattern was incorrect (too large).
4624 If such patterns were nested a few deep, this could multiply and become a real
4625 problem.
4626
4627 2. Added /M option to pcretest to show the memory requirement of a specific
4628 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4629
4630 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4631 compiled in such a way that the backtracking after subsequent failure was
4632 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4633 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4634
4635
4636 Version 2.03 02-Feb-99
4637 ----------------------
4638
4639 1. Fixed typo and small mistake in man page.
4640
4641 2. Added 4th condition (GPL supersedes if conflict) and created separate
4642 LICENCE file containing the conditions.
4643
4644 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4645 Perl, that is the internal \ allows the delimiter to be included in the
4646 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4647 the final delimiter, add \ to the end of the pattern (to test the error).
4648
4649 4. Added the convenience functions for extracting substrings after a successful
4650 match. Updated pcretest to make it able to test these functions.
4651
4652
4653 Version 2.02 14-Jan-99
4654 ----------------------
4655
4656 1. Initialized the working variables associated with each extraction so that
4657 their saving and restoring doesn't refer to uninitialized store.
4658
4659 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4660 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4661 fix the problem.
4662
4663 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4664 calls, and wasn't printing the correct value for compiling calls. Increased the
4665 default value of LOOPREPEAT, and the number of significant figures in the
4666 times.
4667
4668 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4669
4670 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4671 a building problem on Windows NT with a FAT file system.
4672
4673
4674 Version 2.01 21-Oct-98
4675 ----------------------
4676
4677 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4678 to character tables built by pcre_maketables() in the current locale. If NULL
4679 is passed, the default tables are used.
4680
4681
4682 Version 2.00 24-Sep-98
4683 ----------------------
4684
4685 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4686 it any more.
4687
4688 2. Allow quantification of (?>) groups, and make it work correctly.
4689
4690 3. The first character computation wasn't working for (?>) groups.
4691
4692 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4693 end of the subject) and add 5.005's \z, which really does match only at the
4694 very end of the subject.
4695
4696 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4697
4698 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4699 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4700 localized options. All options to pcre_study() were also removed.
4701
4702 7. Add other new features from 5.005:
4703
4704 $(?<= positive lookbehind
4705 $(?<! negative lookbehind
4706 (?imsx-imsx) added the unsetting capability
4707 such a setting is global if at outer level; local otherwise
4708 (?imsx-imsx:) non-capturing groups with option setting
4709 (?(cond)re|re) conditional pattern matching
4710
4711 A backreference to itself in a repeated group matches the previous
4712 captured string.
4713
4714 8. General tidying up of studying (both automatic and via "study")
4715 consequential on the addition of new assertions.
4716
4717 9. As in 5.005, unlimited repeated groups that could match an empty substring
4718 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4719 runtime if any iteration does actually match an empty substring.
4720
4721 10. Include the RunTest script in the distribution.
4722
4723 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4724 discrepancies, some of which were old and were also with respect to 5.004. They
4725 have now been fixed.
4726
4727
4728 Version 1.09 28-Apr-98
4729 ----------------------
4730
4731 1. A negated single character class followed by a quantifier with a minimum
4732 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4733 program crashes, or just wrong answers. This did not apply to negated classes
4734 containing more than one character, or to minima other than one.
4735
4736
4737 Version 1.08 27-Mar-98
4738 ----------------------
4739
4740 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4741
4742 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4743 latter must appear before anything that relies on it in the pattern.
4744
4745
4746 Version 1.07 16-Feb-98
4747 ----------------------
4748
4749 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4750 repeat of a potentially empty string).
4751
4752
4753 Version 1.06 23-Jan-98
4754 ----------------------
4755
4756 1. Added Markus Oberhumer's little patches for C++.
4757
4758 2. Literal strings longer than 255 characters were broken.
4759
4760
4761 Version 1.05 23-Dec-97
4762 ----------------------
4763
4764 1. Negated character classes containing more than one character were failing if
4765 PCRE_CASELESS was set at run time.
4766
4767
4768 Version 1.04 19-Dec-97
4769 ----------------------
4770
4771 1. Corrected the man page, where some "const" qualifiers had been omitted.
4772
4773 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4774 input syntax.
4775
4776 3. Fixed memory leak which occurred when a regex with back references was
4777 matched with an offsets vector that wasn't big enough. The temporary memory
4778 that is used in this case wasn't being freed if the match failed.
4779
4780 4. Tidied pcretest to ensure it frees memory that it gets.
4781
4782 5. Temporary memory was being obtained in the case where the passed offsets
4783 vector was exactly big enough.
4784
4785 6. Corrected definition of offsetof() from change 5 below.
4786
4787 7. I had screwed up change 6 below and broken the rules for the use of
4788 setjmp(). Now fixed.
4789
4790
4791 Version 1.03 18-Dec-97
4792 ----------------------
4793
4794 1. A erroneous regex with a missing opening parenthesis was correctly
4795 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4796 on some systems.
4797
4798 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4799 it was reported that one broken compiler failed on the former because "code" is
4800 also an independent variable.
4801
4802 3. The erroneous regex a[]b caused an array overrun reference.
4803
4804 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4805 fail on data ending with that character. (It was going on too far, and checking
4806 the next character, typically a binary zero.) This was specific to the
4807 optimized code for single-character negative classes.
4808
4809 5. Added a contributed patch from the TIN world which does the following:
4810
4811 + Add an undef for memmove, in case the the system defines a macro for it.
4812
4813 + Add a definition of offsetof(), in case there isn't one. (I don't know
4814 the reason behind this - offsetof() is part of the ANSI standard - but
4815 it does no harm).
4816
4817 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4818 most of the places where whitespace preceded '#'. I have given up and
4819 allowed the remaining 2 cases to be at the margin.
4820
4821 + Rename some variables in pcre to eliminate shadowing. This seems very
4822 pedantic, but does no harm, of course.
4823
4824 6. Moved the call to setjmp() into its own function, to get rid of warnings
4825 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4826
4827 7. Constructs such as \d{8,} were compiling into the equivalent of
4828 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4829 outcome, but in this particular case used more store than had been allocated,
4830 which caused the bug to be discovered because it threw up an internal error.
4831
4832 8. The debugging code in both pcre and pcretest for outputting the compiled
4833 form of a regex was going wrong in the case of back references followed by
4834 curly-bracketed repeats.
4835
4836
4837 Version 1.02 12-Dec-97
4838 ----------------------
4839
4840 1. Typos in pcre.3 and comments in the source fixed.
4841
4842 2. Applied a contributed patch to get rid of places where it used to remove
4843 'const' from variables, and fixed some signed/unsigned and uninitialized
4844 variable warnings.
4845
4846 3. Added the "runtest" target to Makefile.
4847
4848 4. Set default compiler flag to -O2 rather than just -O.
4849
4850
4851 Version 1.01 19-Nov-97
4852 ----------------------
4853
4854 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4855 like /([ab]*)*/, that is, for classes with more than one character in them.
4856
4857 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4858 as /((?>a*))*/ (a PCRE_EXTRA facility).
4859
4860
4861 Version 1.00 18-Nov-97
4862 ----------------------
4863
4864 1. Added compile-time macros to support systems such as SunOS4 which don't have
4865 memmove() or strerror() but have other things that can be used instead.
4866
4867 2. Arranged that "make clean" removes the executables.
4868
4869
4870 Version 0.99 27-Oct-97
4871 ----------------------
4872
4873 1. Fixed bug in code for optimizing classes with only one character. It was
4874 initializing a 32-byte map regardless, which could cause it to run off the end
4875 of the memory it had got.
4876
4877 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4878
4879
4880 Version 0.98 22-Oct-97
4881 ----------------------
4882
4883 1. Fixed bug in code for handling temporary memory usage when there are more
4884 back references than supplied space in the ovector. This could cause segfaults.
4885
4886
4887 Version 0.97 21-Oct-97
4888 ----------------------
4889
4890 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4891
4892 2. Optimized negated single characters not to use a bit map.
4893
4894 3. Brought error texts together as macro definitions; clarified some of them;
4895 fixed one that was wrong - it said "range out of order" when it meant "invalid
4896 escape sequence".
4897
4898 4. Changed some char * arguments to const char *.
4899
4900 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4901
4902 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4903 pcretest.
4904
4905
4906 Version 0.96 16-Oct-97
4907 ----------------------
4908
4909 1. Added a simple "pgrep" utility to the distribution.
4910
4911 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4912 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4913 where "ddd" means "one or more decimal digits".
4914
4915 3. Fixed serious bug. If a pattern had a back reference, but the call to
4916 pcre_exec() didn't supply a large enough ovector to record the related
4917 identifying subpattern, the match always failed. PCRE now remembers the number
4918 of the largest back reference, and gets some temporary memory in which to save
4919 the offsets during matching if necessary, in order to ensure that
4920 backreferences always work.
4921
4922 4. Increased the compatibility with Perl in a number of ways:
4923
4924 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4925 to request this handling. The option can be set at compile or exec time.
4926
4927 (b) $ matches before a terminating newline by default; an option
4928 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4929 mode). The option can be set at compile or exec time.
4930
4931 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4932 the same as Perl's. If the decimal number it represents is less than 10
4933 or there aren't that many previous left capturing parentheses, an octal
4934 escape is read. Inside a character class, it's always an octal escape,
4935 even if it is a single digit.
4936
4937 (d) An escaped but undefined alphabetic character is taken as a literal,
4938 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4939 escapes.
4940
4941 (e) {0} is now permitted. (The previous item is removed from the compiled
4942 pattern).
4943
4944 5. Changed all the names of code files so that the basic parts are no longer
4945 than 10 characters, and abolished the teeny "globals.c" file.
4946
4947 6. Changed the handling of character classes; they are now done with a 32-byte
4948 bit map always.
4949
4950 7. Added the -d and /D options to pcretest to make it possible to look at the
4951 internals of compilation without having to recompile pcre.
4952
4953
4954 Version 0.95 23-Sep-97
4955 ----------------------
4956
4957 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4958 \x20 at the start of a run of normal characters. These were being treated as
4959 real characters, instead of the source characters being re-checked.
4960
4961
4962 Version 0.94 18-Sep-97
4963 ----------------------
4964
4965 1. The functions are now thread-safe, with the caveat that the global variables
4966 containing pointers to malloc() and free() or alternative functions are the
4967 same for all threads.
4968
4969 2. Get pcre_study() to generate a bitmap of initial characters for non-
4970 anchored patterns when this is possible, and use it if passed to pcre_exec().
4971
4972
4973 Version 0.93 15-Sep-97
4974 ----------------------
4975
4976 1. /(b)|(:+)/ was computing an incorrect first character.
4977
4978 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4979 but not actually doing anything yet.
4980
4981 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4982 as Perl does (e.g. [-az] or [az-]).
4983
4984 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4985 all possible positions.
4986
4987 5. Split up into different modules to avoid including unneeded functions in a
4988 compiled binary. However, compile and exec are still in one module. The "study"
4989 function is split off.
4990
4991 6. The character tables are now in a separate module whose source is generated
4992 by an auxiliary program - but can then be edited by hand if required. There are
4993 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4994 toupper() in the code.
4995
4996 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4997 make them global. Abolish the function for setting them, as the caller can now
4998 set them directly.
4999
5000
5001 Version 0.92 11-Sep-97
5002 ----------------------
5003
5004 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5005 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5006
5007 2. Caseless matching was not working in character classes if the characters in
5008 the pattern were in upper case.
5009
5010 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5011
5012 4. Make PCRE_ANCHORED public and accept as a compile option.
5013
5014 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5015 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5016 pass them.
5017
5018 6. Give an error if bad option bits passed at compile or run time.
5019
5020 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5021 pcretest to cause it to pass that flag.
5022
5023 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5024 options, and the first character, if set.
5025
5026 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5027
5028
5029 Version 0.91 10-Sep-97
5030 ----------------------
5031
5032 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5033 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5034
5035 2. PCRE was looping on encountering an indefinitely repeated back reference to
5036 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5037 Perl does - treats the match as successful.
5038
5039 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5