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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1274 - (show annotations)
Fri Mar 8 11:35:41 2013 UTC (6 years, 8 months ago) by ph10
File size: 244590 byte(s)
Fix the case where there are two or more SKIPs that may have to be ignored.
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 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
62
63 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
64 offset of the starting point of the matching process, provided the offsets
65 vector is large enough.
66
67 17. The \A escape now records a lookbehind value of 1, though its execution
68 does not actually inspect the previous character. This is to ensure that,
69 in partial multi-segment matching, at least one character from the old
70 segment is retained when a new segment is processed. Otherwise, if there
71 are no lookbehinds in the pattern, \A might match incorrectly at the start
72 of a new segment.
73
74 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
75
76 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
77 gives some modest performance improvement in 8-bit mode.
78
79 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
80 be expressed in certain programming languages using Universal Character
81 Names.
82
83 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
84 which points out that "non characters" are not "characters that may not
85 appear in Unicode strings" but rather "characters that are reserved for
86 internal use and have only local meaning".
87
88 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
89 there was a conditional group that depended on an assertion, if the
90 assertion was false, the callout that immediately followed the alternation
91 in the condition was skipped when pcre_exec() was used for matching.
92
93 23. Allow an explicit callout to be inserted before an assertion that is the
94 condition for a conditional group, for compatibility with automatic
95 callouts, which always insert a callout at this point.
96
97 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
98 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
99
100 25. (*PRUNE) is now supported by the JIT compiler.
101
102 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
103
104 27. Fix the case where there are two or more SKIPs with arguments that may be
105 ignored.
106
107
108 Version 8.32 30-November-2012
109 -----------------------------
110
111 1. Improved JIT compiler optimizations for first character search and single
112 character iterators.
113
114 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
115 Patch by Daniel Richard G.
116
117 3. Single character iterator optimizations in the JIT compiler.
118
119 4. Improved JIT compiler optimizations for character ranges.
120
121 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
122 Reported by Giuseppe D'Angelo.
123
124 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
125 of a line, was being set incorrectly in cases where .* appeared inside
126 atomic brackets at the start of a pattern, or where there was a subsequent
127 *PRUNE or *SKIP.
128
129 7. Improved instruction cache flush for POWER/PowerPC.
130 Patch by Daniel Richard G.
131
132 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
133 grep:
134
135 (a) There is now no limit to the number of patterns to be matched.
136
137 (b) An error is given if a pattern is too long.
138
139 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
140 are now supported.
141
142 (d) --exclude-from and --include-from (multiple use) have been added.
143
144 (e) Exclusions and inclusions now apply to all files and directories, not
145 just to those obtained from scanning a directory recursively.
146
147 (f) Multiple uses of -f and --file-list are now supported.
148
149 (g) In a Windows environment, the default for -d has been changed from
150 "read" (the GNU grep default) to "skip", because otherwise the presence
151 of a directory in the file list provokes an error.
152
153 (h) The documentation has been revised and clarified in places.
154
155 9. Improve the matching speed of capturing brackets.
156
157 10. Changed the meaning of \X so that it now matches a Unicode extended
158 grapheme cluster.
159
160 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
161 out POSIX threads when JIT support is configured.
162
163 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
164
165 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
166 structure could go wrong in environments where size_t is not the same size
167 as int.
168
169 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
170 set.
171
172 15. The EBCDIC support had decayed; later updates to the code had included
173 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
174 general tidy up of EBCDIC-related issues, and the documentation was also
175 not quite right. There is now a test that can be run on ASCII systems to
176 check some of the EBCDIC-related things (but is it not a full test).
177
178 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
179 in a small tidy to the code.
180
181 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
182
183 18. If the --only-matching (-o) option in pcregrep is specified multiple
184 times, each one causes appropriate output. For example, -o1 -o2 outputs the
185 substrings matched by the 1st and 2nd capturing parentheses. A separating
186 string can be specified by --om-separator (default empty).
187
188 19. Improving the first n character searches.
189
190 20. Turn case lists for horizontal and vertical white space into macros so that
191 they are defined only once.
192
193 21. This set of changes together give more compatible Unicode case-folding
194 behaviour for characters that have more than one other case when UCP
195 support is available.
196
197 (a) The Unicode property table now has offsets into a new table of sets of
198 three or more characters that are case-equivalent. The MultiStage2.py
199 script that generates these tables (the pcre_ucd.c file) now scans
200 CaseFolding.txt instead of UnicodeData.txt for character case
201 information.
202
203 (b) The code for adding characters or ranges of characters to a character
204 class has been abstracted into a generalized function that also handles
205 case-independence. In UTF-mode with UCP support, this uses the new data
206 to handle characters with more than one other case.
207
208 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
209 whose other case is greater than 256 are now correctly matched
210 caselessly. Previously, the high codepoint matched the low one, but not
211 vice versa.
212
213 (d) The processing of \h, \H, \v, and \ in character classes now makes use
214 of the new class addition function, using character lists defined as
215 macros alongside the case definitions of 20 above.
216
217 (e) Caseless back references now work with characters that have more than
218 one other case.
219
220 (f) General caseless matching of characters with more than one other case
221 is supported.
222
223 22. Unicode character properties were updated from Unicode 6.2.0
224
225 23. Improved CMake support under Windows. Patch by Daniel Richard G.
226
227 24. Add support for 32-bit character strings, and UTF-32
228
229 25. Major JIT compiler update (code refactoring and bugfixing).
230 Experimental Sparc 32 support is added.
231
232 26. Applied a modified version of Daniel Richard G's patch to create
233 pcre.h.generic and config.h.generic by "make" instead of in the
234 PrepareRelease script.
235
236 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
237 pcre_compile.c when checking for a zero character.
238
239 28. Introducing a native interface for JIT. Through this interface, the compiled
240 machine code can be directly executed. The purpose of this interface is to
241 provide fast pattern matching, so several sanity checks are not performed.
242 However, feature tests are still performed. The new interface provides
243 1.4x speedup compared to the old one.
244
245 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
246 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
247 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
248
249 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
250 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
251 the "old" RFC 2279). Instead, it ended up passing a negative length to
252 pcre_exec().
253
254 31. Add support for GCC's visibility feature to hide internal functions.
255
256 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
257 "unknown -C option" after outputting 0 or 1.
258
259 33. There is now support for generating a code coverage report for the test
260 suite in environments where gcc is the compiler and lcov is installed. This
261 is mainly for the benefit of the developers.
262
263 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
264 unaddressable using valgrind annotations, allowing valgrind to detect
265 invalid memory accesses. This is mainly for the benefit of the developers.
266
267 25. (*UTF) can now be used to start a pattern in any of the three libraries.
268
269 26. Give configure error if --enable-cpp but no C++ compiler found.
270
271
272 Version 8.31 06-July-2012
273 -------------------------
274
275 1. Fixing a wrong JIT test case and some compiler warnings.
276
277 2. Removed a bashism from the RunTest script.
278
279 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
280 to unsigned type, result still unsigned" that was given by an MS compiler
281 on encountering the code "-sizeof(xxx)".
282
283 4. Partial matching support is added to the JIT compiler.
284
285 5. Fixed several bugs concerned with partial matching of items that consist
286 of more than one character:
287
288 (a) /^(..)\1/ did not partially match "aba" because checking references was
289 done on an "all or nothing" basis. This also applied to repeated
290 references.
291
292 (b) \R did not give a hard partial match if \r was found at the end of the
293 subject.
294
295 (c) \X did not give a hard partial match after matching one or more
296 characters at the end of the subject.
297
298 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
299 a partial match for the string "\r".
300
301 (e) When newline was set to CRLF, the metacharacter "." did not recognize
302 a partial match for a CR character at the end of the subject string.
303
304 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
305 running pcretest, the text "(JIT)" added to the output whenever JIT is
306 actually used to run the match.
307
308 7. Individual JIT compile options can be set in pcretest by following -s+[+]
309 or /S+[+] with a digit between 1 and 7.
310
311 8. OP_NOT now supports any UTF character not just single-byte ones.
312
313 9. (*MARK) control verb is now supported by the JIT compiler.
314
315 10. The command "./RunTest list" lists the available tests without actually
316 running any of them. (Because I keep forgetting what they all are.)
317
318 11. Add PCRE_INFO_MAXLOOKBEHIND.
319
320 12. Applied a (slightly modified) user-supplied patch that improves performance
321 when the heap is used for recursion (compiled with --disable-stack-for-
322 recursion). Instead of malloc and free for each heap frame each time a
323 logical recursion happens, frames are retained on a chain and re-used where
324 possible. This sometimes gives as much as 30% improvement.
325
326 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
327 call.
328
329 14. As documented, (*COMMIT) is now confined to within a positive assertion.
330
331 15. It is now possible to link pcretest with libedit as an alternative to
332 libreadline.
333
334 16. (*COMMIT) control verb is now supported by the JIT compiler.
335
336 17. The Unicode data tables have been updated to Unicode 6.1.0.
337
338 18. Added --file-list option to pcregrep.
339
340 19. Added binary file support to pcregrep, including the -a, --binary-files,
341 -I, and --text options.
342
343 20. The madvise function is renamed for posix_madvise for QNX compatibility
344 reasons. Fixed by Giuseppe D'Angelo.
345
346 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
347
348 22. Changed the option for creating symbolic links for 16-bit man pages from
349 -s to -sf so that re-installing does not cause issues.
350
351 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
352
353 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
354 matches in certain environments (the workspace was not being correctly
355 retained). Also added to pcre_dfa_exec() a simple plausibility check on
356 some of the workspace data at the beginning of a restart.
357
358 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
359 was not doing so when it should - probably a typo introduced by SVN 528
360 (change 8.10/14).
361
362 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
363 \w+ when the character tables indicated that \x{c4} was a word character.
364 There were several related cases, all because the tests for doing a table
365 lookup were testing for characters less than 127 instead of 255.
366
367 27. If a pattern contains capturing parentheses that are not used in a match,
368 their slots in the ovector are set to -1. For those that are higher than
369 any matched groups, this happens at the end of processing. In the case when
370 there were back references that the ovector was too small to contain
371 (causing temporary malloc'd memory to be used during matching), and the
372 highest capturing number was not used, memory off the end of the ovector
373 was incorrectly being set to -1. (It was using the size of the temporary
374 memory instead of the true size.)
375
376 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
377 ovector size, it uses memory at the end of the block that it has got.
378
379 29. Check for an overlong MARK name and give an error at compile time. The
380 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
381
382 30. JIT compiler update.
383
384 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
385 Rill for the patch.
386
387 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
388
389 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
390
391 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
392 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
393 was enabled.
394
395 35. Improve JIT code generation for greedy plus quantifier.
396
397 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
398 1 to "aa" instead of to an empty string. The bug affected repeated groups
399 that could potentially match an empty string.
400
401 37. Optimizing single character iterators in JIT.
402
403 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
404 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
405 codepoints that are too big for the mode are faulted, and in a UTF mode,
406 disallowed codepoints are also faulted.
407
408 39. If PCRE was compiled with UTF support, in three places in the DFA
409 matcher there was code that should only have been obeyed in UTF mode, but
410 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
411 processing when bytes with values greater than 127 were present. In 16-bit
412 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
413 both cases the values are those that cannot be the first data item in a UTF
414 character. The three items that might have provoked this were recursions,
415 possessively repeated groups, and atomic groups.
416
417 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
418 and pcregrep, because some OS require shared objects to be explicitly
419 passed to ld, causing the link step to fail if they are not.
420
421 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
422 mode, patterns that started with \h* or \R* might be incorrectly matched.
423
424
425 Version 8.30 04-February-2012
426 -----------------------------
427
428 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
429 name is defined in ctype.h.
430
431 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
432 only in quite long subpatterns.
433
434 3. Removed the function pcre_info(), which has been obsolete and deprecated
435 since it was replaced by pcre_fullinfo() in February 2000.
436
437 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
438 match a (*MARK), and the match failed at the start of the subject, a
439 reference to memory before the start of the subject could occur. This bug
440 was introduced by fix 17 of release 8.21.
441
442 5. A reference to an unset group with zero minimum repetition was giving
443 totally wrong answers (in non-JavaScript-compatibility mode). For example,
444 /(another)?(\1?)test/ matched against "hello world test". This bug was
445 introduced in release 8.13.
446
447 6. Add support for 16-bit character strings (a large amount of work involving
448 many changes and refactorings).
449
450 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
451 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
452 from a file.
453
454 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
455 rounding is not applied in this particular case).
456
457 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
458 if they appear, or are escaped, in patterns.
459
460 10. Get rid of a number of -Wunused-but-set-variable warnings.
461
462 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
463 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
464 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
465 also returns the mark "x". This bug applied to capturing parentheses,
466 non-capturing parentheses, and atomic parentheses. It also applied to some
467 assertions.
468
469 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
470 information out of configure.ac instead of relying on pcre.h.generic, which
471 is not stored in the repository.
472
473 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
474 -lreadline.
475
476 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
477
478 15. Applied Graycode's patch to put the top-level frame on the stack rather
479 than the heap when not using the stack for recursion. This gives a
480 performance improvement in many cases when recursion is not deep.
481
482 16. Experimental code added to "pcretest -C" to output the stack frame size.
483
484
485 Version 8.21 12-Dec-2011
486 ------------------------
487
488 1. Updating the JIT compiler.
489
490 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
491 are added as well.
492
493 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
494 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
495 calling _pcre_jit_exec. Some extra comments are added.
496
497 4. (*MARK) settings inside atomic groups that do not contain any capturing
498 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
499 was introduced by change 18 for 8.20.
500
501 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
502 ECMA-262 standard.
503
504 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
505 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
506 This bug was probably introduced by change 9 of 8.13.
507
508 7. While fixing 6 above, I noticed that a number of other items were being
509 incorrectly rejected as "not fixed length". This arose partly because newer
510 opcodes had not been added to the fixed-length checking code. I have (a)
511 corrected the bug and added tests for these items, and (b) arranged for an
512 error to occur if an unknown opcode is encountered while checking for fixed
513 length instead of just assuming "not fixed length". The items that were
514 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
515 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
516 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
517
518 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
519 being incorrectly compiled and would have given unpredicatble results.
520
521 9. A possessively repeated subpattern with minimum repeat count greater than
522 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
523 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
524 the first (A) could occur when it should not.
525
526 10. Add a cast and remove a redundant test from the code.
527
528 11. JIT should use pcre_malloc/pcre_free for allocation.
529
530 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
531 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
532 is anything other than /usr, -L is still shown).
533
534 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
535
536 14. Perl does not support \N without a following name in a [] class; PCRE now
537 also gives an error.
538
539 15. If a forward reference was repeated with an upper limit of around 2000,
540 it caused the error "internal error: overran compiling workspace". The
541 maximum number of forward references (including repeats) was limited by the
542 internal workspace, and dependent on the LINK_SIZE. The code has been
543 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
544 the default depends on LINK_SIZE. There is a new upper limit (for safety)
545 of around 200,000 forward references. While doing this, I also speeded up
546 the filling in of repeated forward references.
547
548 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
549 incorrectly expecting the subject to contain another "a" after the start.
550
551 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
552 in the match, the SKIP should be ignored. This was not happening; instead
553 the SKIP was being treated as NOMATCH. For patterns such as
554 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
555 tested.
556
557 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
558 now much more compatible with Perl, in particular in cases where the result
559 is a non-match for a non-anchored pattern. For example, if
560 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
561 "m", where previously it did not return a name. A side effect of this
562 change is that for partial matches, the last encountered mark name is
563 returned, as for non matches. A number of tests that were previously not
564 Perl-compatible have been moved into the Perl-compatible test files. The
565 refactoring has had the pleasing side effect of removing one argument from
566 the match() function, thus reducing its stack requirements.
567
568 19. If the /S+ option was used in pcretest to study a pattern using JIT,
569 subsequent uses of /S (without +) incorrectly behaved like /S+.
570
571 21. Retrieve executable code size support for the JIT compiler and fixing
572 some warnings.
573
574 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
575 not work when the shorter character appeared right at the end of the
576 subject string.
577
578 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
579 systems.
580
581 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
582 output it when the /M option is used in pcretest.
583
584 25. The CheckMan script was not being included in the distribution. Also, added
585 an explicit "perl" to run Perl scripts from the PrepareRelease script
586 because this is reportedly needed in Windows.
587
588 26. If study data was being save in a file and studying had not found a set of
589 "starts with" bytes for the pattern, the data written to the file (though
590 never used) was taken from uninitialized memory and so caused valgrind to
591 complain.
592
593 27. Updated RunTest.bat as provided by Sheri Pierce.
594
595 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
596
597 29. Computation of memory usage for the table of capturing group names was
598 giving an unnecessarily large value.
599
600
601 Version 8.20 21-Oct-2011
602 ------------------------
603
604 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
605 a POSIX class. After further experiments with Perl, which convinced me that
606 Perl has bugs and confusions, a closing square bracket is no longer allowed
607 in a POSIX name. This bug also affected patterns with classes that started
608 with full stops.
609
610 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
611 captured substring, but while checking the failing first alternative,
612 substring 1 is temporarily captured. If the output vector supplied to
613 pcre_exec() was not big enough for this capture, the yield of the function
614 was still zero ("insufficient space for captured substrings"). This cannot
615 be totally fixed without adding another stack variable, which seems a lot
616 of expense for a edge case. However, I have improved the situation in cases
617 such as /(a)(b)x|abc/ matched against "abc", where the return code
618 indicates that fewer than the maximum number of slots in the ovector have
619 been set.
620
621 3. Related to (2) above: when there are more back references in a pattern than
622 slots in the output vector, pcre_exec() uses temporary memory during
623 matching, and copies in the captures as far as possible afterwards. It was
624 using the entire output vector, but this conflicts with the specification
625 that only 2/3 is used for passing back captured substrings. Now it uses
626 only the first 2/3, for compatibility. This is, of course, another edge
627 case.
628
629 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
630 main code base, and can be used by building with --enable-jit. When this is
631 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
632 runtime --no-jit option is given.
633
634 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
635 ovector, the return from the function was zero, implying that there were
636 other matches that did not fit. The correct "exactly full" value is now
637 returned.
638
639 6. If a subpattern that was called recursively or as a subroutine contained
640 (*PRUNE) or any other control that caused it to give a non-standard return,
641 invalid errors such as "Error -26 (nested recursion at the same subject
642 position)" or even infinite loops could occur.
643
644 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
645 computing the minimum length on reaching *ACCEPT, and so ended up with the
646 wrong value of 1 rather than 0. Further investigation indicates that
647 computing a minimum subject length in the presence of *ACCEPT is difficult
648 (think back references, subroutine calls), and so I have changed the code
649 so that no minimum is registered for a pattern that contains *ACCEPT.
650
651 8. If (*THEN) was present in the first (true) branch of a conditional group,
652 it was not handled as intended. [But see 16 below.]
653
654 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
655 Sheri Pierce.
656
657 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
658 the first byte in a match must be "a".
659
660 11. Change 17 for 8.13 increased the recursion depth for patterns like
661 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
662 pattern contains any instances of (*THEN). If it does not, the old
663 optimizations are restored. It would be nice to do this on a per-group
664 basis, but at the moment that is not feasible.
665
666 12. In some environments, the output of pcretest -C is CRLF terminated. This
667 broke RunTest's code that checks for the link size. A single white space
668 character after the value is now allowed for.
669
670 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
671 For "fr", it uses the Windows-specific input and output files.
672
673 14. If (*THEN) appeared in a group that was called recursively or as a
674 subroutine, it did not work as intended. [But see next item.]
675
676 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
677 pattern fragments (but not containing any | characters). If A and B are
678 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
679 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
680 D. In other words, Perl considers parentheses that do not contain any |
681 characters to be part of a surrounding alternative, whereas PCRE was
682 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
683 differently. PCRE now behaves in the same way as Perl, except in the case
684 of subroutine/recursion calls such as (?1) which have in any case always
685 been different (but PCRE had them first :-).
686
687 16. Related to 15 above: Perl does not treat the | in a conditional group as
688 creating alternatives. Such a group is treated in the same way as an
689 ordinary group without any | characters when processing (*THEN). PCRE has
690 been changed to match Perl's behaviour.
691
692 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
693 RunGrepTest script failed.
694
695 18. Change 22 for version 13 caused atomic groups to use more stack. This is
696 inevitable for groups that contain captures, but it can lead to a lot of
697 stack use in large patterns. The old behaviour has been restored for atomic
698 groups that do not contain any capturing parentheses.
699
700 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
701 suppress the check for a minimum subject length at run time. (If it was
702 given to pcre_exec() or pcre_dfa_exec() it did work.)
703
704 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
705 fail to work when decoding hex characters in data strings in EBCDIC
706 environments.
707
708 21. It appears that in at least one Mac OS environment, the isxdigit() function
709 is implemented as a macro that evaluates to its argument more than once,
710 contravening the C 90 Standard (I haven't checked a later standard). There
711 was an instance in pcretest which caused it to go wrong when processing
712 \x{...} escapes in subject strings. The has been rewritten to avoid using
713 things like p++ in the argument of isxdigit().
714
715
716 Version 8.13 16-Aug-2011
717 ------------------------
718
719 1. The Unicode data tables have been updated to Unicode 6.0.0.
720
721 2. Two minor typos in pcre_internal.h have been fixed.
722
723 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
724 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
725 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
726
727 4. There were a number of related bugs in the code for matching backrefences
728 caselessly in UTF-8 mode when codes for the characters concerned were
729 different numbers of bytes. For example, U+023A and U+2C65 are an upper
730 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
731 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
732 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
733 2-byte code at the end of the subject (it thought there wasn't enough data
734 left).
735
736 5. Comprehensive information about what went wrong is now returned by
737 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
738 as the output vector has at least 2 elements. The offset of the start of
739 the failing character and a reason code are placed in the vector.
740
741 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
742 now returned is for the first byte of the failing character, instead of the
743 last byte inspected. This is an incompatible change, but I hope it is small
744 enough not to be a problem. It makes the returned offset consistent with
745 pcre_exec() and pcre_dfa_exec().
746
747 7. pcretest now gives a text phrase as well as the error number when
748 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
749 failure, the offset and reason code are output.
750
751 8. When \R was used with a maximizing quantifier it failed to skip backwards
752 over a \r\n pair if the subsequent match failed. Instead, it just skipped
753 back over a single character (\n). This seems wrong (because it treated the
754 two characters as a single entity when going forwards), conflicts with the
755 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
756 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
757 has been changed.
758
759 9. Some internal refactoring has changed the processing so that the handling
760 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
761 time (the PCRE_DOTALL option was changed this way some time ago: version
762 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
763 which was always a bit of a fudge. It also means that there is one less
764 argument for the match() function, which reduces its stack requirements
765 slightly. This change also fixes an incompatibility with Perl: the pattern
766 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
767
768 10. More internal refactoring has drastically reduced the number of recursive
769 calls to match() for possessively repeated groups such as (abc)++ when
770 using pcre_exec().
771
772 11. While implementing 10, a number of bugs in the handling of groups were
773 discovered and fixed:
774
775 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
776 (a|)*(?1) gave a compile-time internal error.
777 ((a|)+)+ did not notice that the outer group could match an empty string.
778 (^a|^)+ was not marked as anchored.
779 (.*a|.*)+ was not marked as matching at start or after a newline.
780
781 12. Yet more internal refactoring has removed another argument from the match()
782 function. Special calls to this function are now indicated by setting a
783 value in a variable in the "match data" data block.
784
785 13. Be more explicit in pcre_study() instead of relying on "default" for
786 opcodes that mean there is no starting character; this means that when new
787 ones are added and accidentally left out of pcre_study(), testing should
788 pick them up.
789
790 14. The -s option of pcretest has been documented for ages as being an old
791 synonym of -m (show memory usage). I have changed it to mean "force study
792 for every regex", that is, assume /S for every regex. This is similar to -i
793 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
794 using it. It makes it easier to run collections of tests with and without
795 study enabled, and thereby test pcre_study() more easily. All the standard
796 tests are now run with and without -s (but some patterns can be marked as
797 "never study" - see 20 below).
798
799 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
800 restoration of the capturing data to the outer values was not happening
801 correctly.
802
803 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
804 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
805 pattern had matched an empty string, and so incorrectly returned a no
806 match.
807
808 17. There was optimizing code for the last branch of non-capturing parentheses,
809 and also for the obeyed branch of a conditional subexpression, which used
810 tail recursion to cut down on stack usage. Unfortunately, now that there is
811 the possibility of (*THEN) occurring in these branches, tail recursion is
812 no longer possible because the return has to be checked for (*THEN). These
813 two optimizations have therefore been removed. [But see 8.20/11 above.]
814
815 18. If a pattern containing \R was studied, it was assumed that \R always
816 matched two bytes, thus causing the minimum subject length to be
817 incorrectly computed because \R can also match just one byte.
818
819 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
820 was incorrectly computed.
821
822 20. If /S is present twice on a test pattern in pcretest input, it now
823 *disables* studying, thereby overriding the use of -s on the command line
824 (see 14 above). This is necessary for one or two tests to keep the output
825 identical in both cases.
826
827 21. When (*ACCEPT) was used in an assertion that matched an empty string and
828 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
829
830 22. When an atomic group that contained a capturing parenthesis was
831 successfully matched, but the branch in which it appeared failed, the
832 capturing was not being forgotten if a higher numbered group was later
833 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
834 group 1 to "a", when in fact it should be unset. This applied to multi-
835 branched capturing and non-capturing groups, repeated or not, and also to
836 positive assertions (capturing in negative assertions does not happen
837 in PCRE) and also to nested atomic groups.
838
839 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
840 subject after a captured substring, to make it easier to tell which of a
841 number of identical substrings has been captured.
842
843 24. The way atomic groups are processed by pcre_exec() has been changed so that
844 if they are repeated, backtracking one repetition now resets captured
845 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
846 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
847 "aaa". Previously, it would have been "a". As part of this code
848 refactoring, the way recursive calls are handled has also been changed.
849
850 25. If an assertion condition captured any substrings, they were not passed
851 back unless some other capturing happened later. For example, if
852 (?(?=(a))a) was matched against "a", no capturing was returned.
853
854 26. When studying a pattern that contained subroutine calls or assertions,
855 the code for finding the minimum length of a possible match was handling
856 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
857 group 1 called group 2 while simultaneously a separate group 2 called group
858 1). A stack overflow occurred in this case. I have fixed this by limiting
859 the recursion depth to 10.
860
861 27. Updated RunTest.bat in the distribution to the version supplied by Tom
862 Fortmann. This supports explicit test numbers on the command line, and has
863 argument validation and error reporting.
864
865 28. An instance of \X with an unlimited repeat could fail if at any point the
866 first character it looked at was a mark character.
867
868 29. Some minor code refactoring concerning Unicode properties and scripts
869 should reduce the stack requirement of match() slightly.
870
871 30. Added the '=' option to pcretest to check the setting of unused capturing
872 slots at the end of the pattern, which are documented as being -1, but are
873 not included in the return count.
874
875 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
876 compiled something random. Now it gives a compile-time error (as does
877 Perl).
878
879 32. A *MARK encountered during the processing of a positive assertion is now
880 recorded and passed back (compatible with Perl).
881
882 33. If --only-matching or --colour was set on a pcregrep call whose pattern
883 had alternative anchored branches, the search for a second match in a line
884 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
885 matched the line "0102" twice. The same bug affected patterns that started
886 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
887 twice.
888
889 34. Previously, PCRE did not allow quantification of assertions. However, Perl
890 does, and because of capturing effects, quantifying parenthesized
891 assertions may at times be useful. Quantifiers are now allowed for
892 parenthesized assertions.
893
894 35. A minor code tidy in pcre_compile() when checking options for \R usage.
895
896 36. \g was being checked for fancy things in a character class, when it should
897 just be a literal "g".
898
899 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
900 appearance of a nested POSIX class supersedes an apparent external class.
901 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
902 unescaped square brackets may also appear as part of class names. For
903 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
904 more like Perl. (But see 8.20/1 above.)
905
906 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
907 was because it thought it was \N{name}, which is not supported).
908
909 39. Add minix to OS list not supporting the -S option in pcretest.
910
911 40. PCRE tries to detect cases of infinite recursion at compile time, but it
912 cannot analyze patterns in sufficient detail to catch mutual recursions
913 such as ((?1))((?2)). There is now a runtime test that gives an error if a
914 subgroup is called recursively as a subpattern for a second time at the
915 same position in the subject string. In previous releases this might have
916 been caught by the recursion limit, or it might have run out of stack.
917
918 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
919 happen only once. PCRE was, however incorrectly giving a compile time error
920 "recursive call could loop indefinitely" because it cannot analyze the
921 pattern in sufficient detail. The compile time test no longer happens when
922 PCRE is compiling a conditional subpattern, but actual runaway loops are
923 now caught at runtime (see 40 above).
924
925 42. It seems that Perl allows any characters other than a closing parenthesis
926 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
927 has been changed to be the same.
928
929 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
930 as not to get warnings when autogen.sh is called. Also changed
931 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
932
933 44. To help people who use pcregrep to scan files containing exceedingly long
934 lines, the following changes have been made:
935
936 (a) The default value of the buffer size parameter has been increased from
937 8K to 20K. (The actual buffer used is three times this size.)
938
939 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
940 PCRE is built.
941
942 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
943 to be set at run time.
944
945 (d) Numerical values in pcregrep options can be followed by K or M, for
946 example --buffer-size=50K.
947
948 (e) If a line being scanned overflows pcregrep's buffer, an error is now
949 given and the return code is set to 2.
950
951 45. Add a pointer to the latest mark to the callout data block.
952
953 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
954 partial match of an empty string instead of no match. This was specific to
955 the use of ".".
956
957 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
958 complete match instead of a partial match. This bug was dependent on both
959 the PCRE_UTF8 and PCRE_DOTALL options being set.
960
961 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
962 starting byte set, because \b was not being ignored.
963
964
965 Version 8.12 15-Jan-2011
966 ------------------------
967
968 1. Fixed some typos in the markup of the man pages, and wrote a script that
969 checks for such things as part of the documentation building process.
970
971 2. On a big-endian 64-bit system, pcregrep did not correctly process the
972 --match-limit and --recursion-limit options (added for 8.11). In
973 particular, this made one of the standard tests fail. (The integer value
974 went into the wrong half of a long int.)
975
976 3. If the --colour option was given to pcregrep with -v (invert match), it
977 did strange things, either producing crazy output, or crashing. It should,
978 of course, ignore a request for colour when reporting lines that do not
979 match.
980
981 4. Another pcregrep bug caused similar problems if --colour was specified with
982 -M (multiline) and the pattern match finished with a line ending.
983
984 5. In pcregrep, when a pattern that ended with a literal newline sequence was
985 matched in multiline mode, the following line was shown as part of the
986 match. This seems wrong, so I have changed it.
987
988 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
989 the check for further matches in the same line (so they could be coloured)
990 to overrun the end of the current line. If another match was found, it was
991 incorrectly shown (and then shown again when found in the next line).
992
993 7. If pcregrep was compiled under Windows, there was a reference to the
994 function pcregrep_exit() before it was defined. I am assuming this was
995 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
996 reported by a user. I've moved the definition above the reference.
997
998
999 Version 8.11 10-Dec-2010
1000 ------------------------
1001
1002 1. (*THEN) was not working properly if there were untried alternatives prior
1003 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1004 backtracked to try for "b" instead of moving to the next alternative branch
1005 at the same level (in this case, to look for "c"). The Perl documentation
1006 is clear that when (*THEN) is backtracked onto, it goes to the "next
1007 alternative in the innermost enclosing group".
1008
1009 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1010 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1011 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1012 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1013 (*THEN).
1014
1015 3. If \s appeared in a character class, it removed the VT character from
1016 the class, even if it had been included by some previous item, for example
1017 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1018 of \s, but is part of the POSIX "space" class.)
1019
1020 4. A partial match never returns an empty string (because you can always
1021 match an empty string at the end of the subject); however the checking for
1022 an empty string was starting at the "start of match" point. This has been
1023 changed to the "earliest inspected character" point, because the returned
1024 data for a partial match starts at this character. This means that, for
1025 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1026 (previously it gave "no match").
1027
1028 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1029 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1030 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1031 has an implication that the given string is incomplete (because a partial
1032 match is preferred over a full match). For this reason, these items now
1033 give a partial match in this situation. [Aside: previously, the one case
1034 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1035 match rather than a full match, which was wrong by the old rules, but is
1036 now correct.]
1037
1038 6. There was a bug in the handling of #-introduced comments, recognized when
1039 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1040 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1041 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1042 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1043 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1044 places in pcre_compile().
1045
1046 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1047 comments when looking ahead for named forward references to subpatterns,
1048 the only newline sequence it recognized was NL. It now handles newlines
1049 according to the set newline convention.
1050
1051 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1052 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1053 cater for a lack of strerror(). These oversights have been fixed.
1054
1055 9. Added --match-limit and --recursion-limit to pcregrep.
1056
1057 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1058
1059 11. When the -o option was used, pcregrep was setting a return code of 1, even
1060 when matches were found, and --line-buffered was not being honoured.
1061
1062 12. Added an optional parentheses number to the -o and --only-matching options
1063 of pcregrep.
1064
1065 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1066 can match an empty string. The code to do it in pcretest and pcredemo
1067 needed fixing:
1068
1069 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1070 only one byte after an empty string match just before CRLF (this case
1071 just got forgotten; "any" and "anycrlf" were OK).
1072
1073 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1074 mode when an empty string match preceded an ASCII character followed by
1075 a non-ASCII character. (The code for advancing by one character rather
1076 than one byte was nonsense.)
1077
1078 (c) The pcredemo.c sample program did not have any code at all to handle
1079 the cases when CRLF is a valid newline sequence.
1080
1081 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1082 as a starting offset was within the subject string. There is now a new
1083 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1084 negative or greater than the length of the string. In order to test this,
1085 pcretest is extended to allow the setting of negative starting offsets.
1086
1087 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1088 starting offset points to the beginning of a UTF-8 character was
1089 unnecessarily clumsy. I tidied it up.
1090
1091 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1092 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1093
1094 17. Nobody had reported that the --include_dir option, which was added in
1095 release 7.7 should have been called --include-dir (hyphen, not underscore)
1096 for compatibility with GNU grep. I have changed it to --include-dir, but
1097 left --include_dir as an undocumented synonym, and the same for
1098 --exclude-dir, though that is not available in GNU grep, at least as of
1099 release 2.5.4.
1100
1101 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1102 characters from a string of bytes) have been redefined so as not to use
1103 loops, in order to improve performance in some environments. At the same
1104 time, I abstracted some of the common code into auxiliary macros to save
1105 repetition (this should not affect the compiled code).
1106
1107 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1108 compile-time error is now given if \c is not followed by an ASCII
1109 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1110 different, and any byte value is allowed.)
1111
1112 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1113 START_OPTIMIZE option, which is now allowed at compile time - but just
1114 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1115 to pcregrep and other applications that have no direct access to PCRE
1116 options. The new /Y option in pcretest sets this option when calling
1117 pcre_compile().
1118
1119 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1120 back references. Groups containing recursive back references were forced to
1121 be atomic by that change, but in the case of named groups, the amount of
1122 memory required was incorrectly computed, leading to "Failed: internal
1123 error: code overflow". This has been fixed.
1124
1125 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1126 pcretest.c, to avoid build problems in some Borland environments.
1127
1128
1129 Version 8.10 25-Jun-2010
1130 ------------------------
1131
1132 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1133 THEN.
1134
1135 2. (*ACCEPT) was not working when inside an atomic group.
1136
1137 3. Inside a character class, \B is treated as a literal by default, but
1138 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1139 causes the error). The code is unchanged, but I tidied the documentation.
1140
1141 4. Inside a character class, PCRE always treated \R and \X as literals,
1142 whereas Perl faults them if its -w option is set. I have changed PCRE so
1143 that it faults them when PCRE_EXTRA is set.
1144
1145 5. Added support for \N, which always matches any character other than
1146 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1147
1148 6. When compiling pcregrep with newer versions of gcc which may have
1149 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1150 declared with attribute warn_unused_result" were given. Just casting the
1151 result to (void) does not stop the warnings; a more elaborate fudge is
1152 needed. I've used a macro to implement this.
1153
1154 7. Minor change to pcretest.c to avoid a compiler warning.
1155
1156 8. Added four artifical Unicode properties to help with an option to make
1157 \s etc use properties (see next item). The new properties are: Xan
1158 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1159
1160 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1161 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1162 this option. Modified pcretest to add /W to test this facility. Added
1163 REG_UCP to make it available via the POSIX interface.
1164
1165 10. Added --line-buffered to pcregrep.
1166
1167 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1168 studied, and the match started with a letter with a code point greater than
1169 127 whose first byte was different to the first byte of the other case of
1170 the letter, the other case of this starting letter was not recognized
1171 (#976).
1172
1173 12. If a pattern that was studied started with a repeated Unicode property
1174 test, for example, \p{Nd}+, there was the theoretical possibility of
1175 setting up an incorrect bitmap of starting bytes, but fortunately it could
1176 not have actually happened in practice until change 8 above was made (it
1177 added property types that matched character-matching opcodes).
1178
1179 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1180 possible starting bytes for non-anchored patterns.
1181
1182 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1183 \R, and also a number of cases that involve Unicode properties, both
1184 explicit and implicit when PCRE_UCP is set.
1185
1186 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1187 input, it could crash or give wrong results if characters with values
1188 greater than 0xc0 were present in the subject string. (Detail: it assumed
1189 UTF-8 input when processing these items.)
1190
1191 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1192 size_t is 64-bit (#991).
1193
1194 17. Added a check for running out of memory when PCRE is compiled with
1195 --disable-stack-for-recursion (#990).
1196
1197 18. If the last data line in a file for pcretest does not have a newline on
1198 the end, a newline was missing in the output.
1199
1200 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1201 less than 128) in its various bitmaps. However, there is a facility for
1202 generating tables according to the current locale when PCRE is compiled. It
1203 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1204 characters, are recognized by isspace() and therefore were getting set in
1205 these tables, and indeed these tables seem to approximate to ISO 8859. This
1206 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1207 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1208 which of course cannot start UTF-8 characters. I have changed the code so
1209 that only real ASCII characters (less than 128) and the correct starting
1210 bytes for UTF-8 encodings are set for characters greater than 127 when in
1211 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1212 altogether.)
1213
1214 20. Added the /T option to pcretest so as to be able to run tests with non-
1215 standard character tables, thus making it possible to include the tests
1216 used for 19 above in the standard set of tests.
1217
1218 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1219 reference to a subpattern the other side of a comment that contains an
1220 opening parenthesis caused either an internal compiling error, or a
1221 reference to the wrong subpattern.
1222
1223
1224 Version 8.02 19-Mar-2010
1225 ------------------------
1226
1227 1. The Unicode data tables have been updated to Unicode 5.2.0.
1228
1229 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1230 configured.
1231
1232 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1233 original author of that file, following a query about its status.
1234
1235 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1236 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1237
1238 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1239 quantifier applied to a forward-referencing subroutine call, could compile
1240 incorrect code or give the error "internal error: previously-checked
1241 referenced subpattern not found".
1242
1243 6. Both MS Visual Studio and Symbian OS have problems with initializing
1244 variables to point to external functions. For these systems, therefore,
1245 pcre_malloc etc. are now initialized to local functions that call the
1246 relevant global functions.
1247
1248 7. There were two entries missing in the vectors called coptable and poptable
1249 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1250 I've fixed the data, and added a kludgy way of testing at compile time that
1251 the lengths are correct (equal to the number of opcodes).
1252
1253 8. Following on from 7, I added a similar kludge to check the length of the
1254 eint vector in pcreposix.c.
1255
1256 9. Error texts for pcre_compile() are held as one long string to avoid too
1257 much relocation at load time. To find a text, the string is searched,
1258 counting zeros. There was no check for running off the end of the string,
1259 which could happen if a new error number was added without updating the
1260 string.
1261
1262 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1263
1264 11. \K was not working if it appeared in an atomic group or in a group that
1265 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1266 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1267 the assertion is positive, but not if it is negative.
1268
1269 12. Change 11 fortuitously reduced the size of the stack frame used in the
1270 "match()" function of pcre_exec.c by one pointer. Forthcoming
1271 implementation of support for (*MARK) will need an extra pointer on the
1272 stack; I have reserved it now, so that the stack frame size does not
1273 decrease.
1274
1275 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1276 item in branch that calls a recursion is a subroutine call - as in the
1277 second branch in the above example - was incorrectly given the compile-
1278 time error "recursive call could loop indefinitely" because pcre_compile()
1279 was not correctly checking the subroutine for matching a non-empty string.
1280
1281 14. The checks for overrunning compiling workspace could trigger after an
1282 overrun had occurred. This is a "should never occur" error, but it can be
1283 triggered by pathological patterns such as hundreds of nested parentheses.
1284 The checks now trigger 100 bytes before the end of the workspace.
1285
1286 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1287
1288
1289 Version 8.01 19-Jan-2010
1290 ------------------------
1291
1292 1. If a pattern contained a conditional subpattern with only one branch (in
1293 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1294 computed the wrong minimum data length (which is of course zero for such
1295 subpatterns). This could cause incorrect "no match" results.
1296
1297 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1298 the pattern is reset in the first branch, pcre_compile() failed with
1299 "internal error: code overflow at offset...". This happened only when
1300 the reset was to the original external option setting. (An optimization
1301 abstracts leading options settings into an external setting, which was the
1302 cause of this.)
1303
1304 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1305 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1306 assertion pattern did not match (meaning that the assertion was true), it
1307 was incorrectly treated as false if the SKIP had been reached during the
1308 matching. This also applied to assertions used as conditions.
1309
1310 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1311 assertion subpattern, including such a pattern used as a condition,
1312 unpredictable results occurred, instead of the error return
1313 PCRE_ERROR_DFA_UITEM.
1314
1315 5. The C++ GlobalReplace function was not working like Perl for the special
1316 situation when an empty string is matched. It now does the fancy magic
1317 stuff that is necessary.
1318
1319 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1320 removed. (These were left over from very, very early versions of PCRE.)
1321
1322 7. Some cosmetic changes to the code to make life easier when compiling it
1323 as part of something else:
1324
1325 (a) Change DEBUG to PCRE_DEBUG.
1326
1327 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1328 called "current" as "current_branch", to prevent a collision with the
1329 Linux macro when compiled as a kernel module.
1330
1331 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1332 prevent a collision with the Linux macro when compiled as a kernel
1333 module.
1334
1335 8. In pcre_compile() there are some checks for integer overflows that used to
1336 cast potentially large values to (double). This has been changed to that
1337 when building, a check for int64_t is made, and if it is found, it is used
1338 instead, thus avoiding the use of floating point arithmetic. (There is no
1339 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1340 double.
1341
1342 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1343 2005 (difference between two addresses compared to an unsigned value).
1344
1345 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1346 custom one, because of the following reported problem in Windows:
1347
1348 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1349 under Win32.
1350 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1351 therefore missing the function definition.
1352 - The compiler thus generates a "C" signature for the test function.
1353 - The linker fails to find the "C" function.
1354 - PCRE fails to configure if asked to do so against libbz2.
1355
1356 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1357 messages were output:
1358
1359 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1360 rerunning libtoolize, to keep the correct libtool macros in-tree.
1361 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1362
1363 I have done both of these things.
1364
1365 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1366 most of the time, it *can* run out if it is given a pattern that contains a
1367 runaway infinite recursion. I updated the discussion in the pcrestack man
1368 page.
1369
1370 13. Now that we have gone to the x.xx style of version numbers, the minor
1371 version may start with zero. Using 08 or 09 is a bad idea because users
1372 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1373 interpreted as invalid octal numbers. I've updated the previous comment in
1374 configure.ac, and also added a check that gives an error if 08 or 09 are
1375 used.
1376
1377 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1378 causing partial matching to fail when the end of the subject matched \W
1379 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1380
1381 15. There were some discrepancies between the declarations in pcre_internal.h
1382 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1383 their definitions. The declarations used "const uschar *" and the
1384 definitions used USPTR. Even though USPTR is normally defined as "const
1385 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1386 reported that: "This difference in casting confuses some C++ compilers, for
1387 example, SunCC recognizes above declarations as different functions and
1388 generates broken code for hbpcre." I have changed the declarations to use
1389 USPTR.
1390
1391 16. GNU libtool is named differently on some systems. The autogen.sh script now
1392 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1393 (FreeBSD).
1394
1395 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1396 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1397 comment: "Figure out how to create a longlong from a string: strtoll and
1398 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1399 instance, but it only takes 2 args instead of 3!"
1400
1401 18. A subtle bug concerned with back references has been fixed by a change of
1402 specification, with a corresponding code fix. A pattern such as
1403 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1404 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1405 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1406 same bug. Such groups have to be quantified to be useful, or contained
1407 inside another quantified group. (If there's no repetition, the reference
1408 can never match.) The problem arises because, having left the group and
1409 moved on to the rest of the pattern, a later failure that backtracks into
1410 the group uses the captured value from the final iteration of the group
1411 rather than the correct earlier one. I have fixed this in PCRE by forcing
1412 any group that contains a reference to itself to be an atomic group; that
1413 is, there cannot be any backtracking into it once it has completed. This is
1414 similar to recursive and subroutine calls.
1415
1416
1417 Version 8.00 19-Oct-09
1418 ----------------------
1419
1420 1. The table for translating pcre_compile() error codes into POSIX error codes
1421 was out-of-date, and there was no check on the pcre_compile() error code
1422 being within the table. This could lead to an OK return being given in
1423 error.
1424
1425 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1426 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1427 in a Windows environment.
1428
1429 3. The pcregrep --count option prints the count for each file even when it is
1430 zero, as does GNU grep. However, pcregrep was also printing all files when
1431 --files-with-matches was added. Now, when both options are given, it prints
1432 counts only for those files that have at least one match. (GNU grep just
1433 prints the file name in this circumstance, but including the count seems
1434 more useful - otherwise, why use --count?) Also ensured that the
1435 combination -clh just lists non-zero counts, with no names.
1436
1437 4. The long form of the pcregrep -F option was incorrectly implemented as
1438 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1439 but it seems right to fix it, and I didn't think it was worth preserving
1440 the old behaviour.
1441
1442 5. The command line items --regex=pattern and --regexp=pattern were not
1443 recognized by pcregrep, which required --regex pattern or --regexp pattern
1444 (with a space rather than an '='). The man page documented the '=' forms,
1445 which are compatible with GNU grep; these now work.
1446
1447 6. No libpcreposix.pc file was created for pkg-config; there was just
1448 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1449
1450 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1451 when UCP support is not needed, by modifying the Python script that
1452 generates it from Unicode data files. This should not matter if the module
1453 is correctly used as a library, but I received one complaint about 50K of
1454 unwanted data. My guess is that the person linked everything into his
1455 program rather than using a library. Anyway, it does no harm.
1456
1457 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1458 was a minimum greater than 1 for a wide character in a possessive
1459 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1460 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1461 character. Chaos in the form of incorrect output or a compiling loop could
1462 result.
1463
1464 9. The restrictions on what a pattern can contain when partial matching is
1465 requested for pcre_exec() have been removed. All patterns can now be
1466 partially matched by this function. In addition, if there are at least two
1467 slots in the offset vector, the offset of the earliest inspected character
1468 for the match and the offset of the end of the subject are set in them when
1469 PCRE_ERROR_PARTIAL is returned.
1470
1471 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1472 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1473 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1474 and may be more useful for multi-segment matching.
1475
1476 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1477 used to be given if ever the end of the subject was reached; now it is
1478 given only if matching could not proceed because another character was
1479 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1480 string "Z", which now yields "no match" instead of "partial match". In the
1481 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1482 final character ended with (*FAIL).
1483
1484 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1485 if the pattern had a "must contain" character that was already found in the
1486 earlier partial match, unless partial matching was again requested. For
1487 example, with the pattern /dog.(body)?/, the "must contain" character is
1488 "g". If the first part-match was for the string "dog", restarting with
1489 "sbody" failed. This bug has been fixed.
1490
1491 13. The string returned by pcre_dfa_exec() after a partial match has been
1492 changed so that it starts at the first inspected character rather than the
1493 first character of the match. This makes a difference only if the pattern
1494 starts with a lookbehind assertion or \b or \B (\K is not supported by
1495 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1496 matching functions compatible, and I think it's the right thing to do.
1497
1498 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1499 so that the demonstration program is easily available in environments where
1500 PCRE has not been installed from source.
1501
1502 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1503 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1504 library.
1505
1506 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1507 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1508 is not the first non-POSIX option to be added. Clearly some people find
1509 these options useful.
1510
1511 17. If a caller to the POSIX matching function regexec() passes a non-zero
1512 value for nmatch with a NULL value for pmatch, the value of
1513 nmatch is forced to zero.
1514
1515 18. RunGrepTest did not have a test for the availability of the -u option of
1516 the diff command, as RunTest does. It now checks in the same way as
1517 RunTest, and also checks for the -b option.
1518
1519 19. If an odd number of negated classes containing just a single character
1520 interposed, within parentheses, between a forward reference to a named
1521 subpattern and the definition of the subpattern, compilation crashed with
1522 an internal error, complaining that it could not find the referenced
1523 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1524 [The bug was that it was starting one character too far in when skipping
1525 over the character class, thus treating the ] as data rather than
1526 terminating the class. This meant it could skip too much.]
1527
1528 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1529 /g option in pcretest when the pattern contains \K, which makes it possible
1530 to have an empty string match not at the start, even when the pattern is
1531 anchored. Updated pcretest and pcredemo to use this option.
1532
1533 21. If the maximum number of capturing subpatterns in a recursion was greater
1534 than the maximum at the outer level, the higher number was returned, but
1535 with unset values at the outer level. The correct (outer level) value is
1536 now given.
1537
1538 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1539 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1540 make it do so. The string so far is captured, making this feature
1541 compatible with Perl.
1542
1543 23. The tests have been re-organized, adding tests 11 and 12, to make it
1544 possible to check the Perl 5.10 features against Perl 5.10.
1545
1546 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1547 pattern matches a fixed length string. PCRE did not allow this; now it
1548 does. Neither allows recursion.
1549
1550 25. I finally figured out how to implement a request to provide the minimum
1551 length of subject string that was needed in order to match a given pattern.
1552 (It was back references and recursion that I had previously got hung up
1553 on.) This code has now been added to pcre_study(); it finds a lower bound
1554 to the length of subject needed. It is not necessarily the greatest lower
1555 bound, but using it to avoid searching strings that are too short does give
1556 some useful speed-ups. The value is available to calling programs via
1557 pcre_fullinfo().
1558
1559 26. While implementing 25, I discovered to my embarrassment that pcretest had
1560 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1561 study optimizations had never been tested with that matching function.
1562 Oops. What is worse, even when it was passed study data, there was a bug in
1563 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1564 were also very few tests of studied patterns with pcre_dfa_exec().
1565
1566 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1567 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1568 on the other side of the coin, they are no longer allowed to have different
1569 names, because these cannot be distinguished in PCRE, and this has caused
1570 confusion. (This is a difference from Perl.)
1571
1572 28. When duplicate subpattern names are present (necessarily with different
1573 numbers, as required by 27 above), and a test is made by name in a
1574 conditional pattern, either for a subpattern having been matched, or for
1575 recursion in such a pattern, all the associated numbered subpatterns are
1576 tested, and the overall condition is true if the condition is true for any
1577 one of them. This is the way Perl works, and is also more like the way
1578 testing by number works.
1579
1580
1581 Version 7.9 11-Apr-09
1582 ---------------------
1583
1584 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1585 (pcretest), all targets were linked against these libraries. This included
1586 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1587 libraries. This caused unwanted dependencies to be created. This problem
1588 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1589 pcretest is linked with readline.
1590
1591 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1592 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1593 moved outside it again, because FALSE and TRUE are already defined in AIX,
1594 but BOOL is not.
1595
1596 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1597 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1598
1599 4. The pcregrep documentation said spaces were inserted as well as colons (or
1600 hyphens) following file names and line numbers when outputting matching
1601 lines. This is not true; no spaces are inserted. I have also clarified the
1602 wording for the --colour (or --color) option.
1603
1604 5. In pcregrep, when --colour was used with -o, the list of matching strings
1605 was not coloured; this is different to GNU grep, so I have changed it to be
1606 the same.
1607
1608 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1609 each matching line was coloured. Now it goes on to look for further matches
1610 of any of the test patterns, which is the same behaviour as GNU grep.
1611
1612 7. A pattern that could match an empty string could cause pcregrep to loop; it
1613 doesn't make sense to accept an empty string match in pcregrep, so I have
1614 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1615 seems to be how GNU grep behaves.
1616
1617 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1618 start or after a newline", because the conditional assertion was not being
1619 correctly handled. The rule now is that both the assertion and what follows
1620 in the first alternative must satisfy the test.
1621
1622 9. If auto-callout was enabled in a pattern with a conditional group whose
1623 condition was an assertion, PCRE could crash during matching, both with
1624 pcre_exec() and pcre_dfa_exec().
1625
1626 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1627 used for matching.
1628
1629 11. Unicode property support in character classes was not working for
1630 characters (bytes) greater than 127 when not in UTF-8 mode.
1631
1632 12. Added the -M command line option to pcretest.
1633
1634 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1635
1636 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1637
1638 16. Added comments and documentation about mis-use of no_arg in the C++
1639 wrapper.
1640
1641 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1642 from Martin Jerabek that uses macro names for all relevant character and
1643 string constants.
1644
1645 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1646 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1647 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1648 these, but not everybody uses configure.
1649
1650 19. A conditional group that had only one branch was not being correctly
1651 recognized as an item that could match an empty string. This meant that an
1652 enclosing group might also not be so recognized, causing infinite looping
1653 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1654 with the subject "ab", where knowledge that the repeated group can match
1655 nothing is needed in order to break the loop.
1656
1657 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1658 exec(), but without supplying a callout function, matching went wrong.
1659
1660 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1661 leak if the size of the offset vector was greater than 30. When the vector
1662 is smaller, the saved offsets during recursion go onto a local stack
1663 vector, but for larger vectors malloc() is used. It was failing to free
1664 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1665 error, in fact).
1666
1667 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1668 heapframe that is used only when UTF-8 support is enabled. This caused no
1669 problem, but was untidy.
1670
1671 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1672 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1673 included within another project.
1674
1675 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1676 slightly modified by me:
1677
1678 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1679 not building pcregrep.
1680
1681 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1682 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1683
1684 25. Forward references, both numeric and by name, in patterns that made use of
1685 duplicate group numbers, could behave incorrectly or give incorrect errors,
1686 because when scanning forward to find the reference group, PCRE was not
1687 taking into account the duplicate group numbers. A pattern such as
1688 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1689
1690 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1691 the feature of a custom pointer more persuasive (as requested by a user).
1692
1693 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1694 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1695 pre-defined.
1696
1697 28. Added support for (*UTF8) at the start of a pattern.
1698
1699 29. Arrange for flags added by the "release type" setting in CMake to be shown
1700 in the configuration summary.
1701
1702
1703 Version 7.8 05-Sep-08
1704 ---------------------
1705
1706 1. Replaced UCP searching code with optimized version as implemented for Ad
1707 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1708 stage table and inline lookup instead of a function, giving speed ups of 2
1709 to 5 times on some simple patterns that I tested. Permission was given to
1710 distribute the MultiStage2.py script that generates the tables (it's not in
1711 the tarball, but is in the Subversion repository).
1712
1713 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1714 scripts.
1715
1716 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1717 a group with a zero qualifier. The result of the study could be incorrect,
1718 or the function might crash, depending on the pattern.
1719
1720 4. Caseless matching was not working for non-ASCII characters in back
1721 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1722 It now works when Unicode Property Support is available.
1723
1724 5. In pcretest, an escape such as \x{de} in the data was always generating
1725 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1726 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1727 truncation.
1728
1729 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1730
1731 7. Added two (int) casts to pcregrep when printing the difference of two
1732 pointers, in case they are 64-bit values.
1733
1734 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1735 test 2 if it fails.
1736
1737 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1738 and a #define of that name to empty if it is not externally set. This is to
1739 allow users of MSVC to set it if necessary.
1740
1741 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1742 the convenience functions in the pcre_get.c source file.
1743
1744 11. An option change at the start of a pattern that had top-level alternatives
1745 could cause overwriting and/or a crash. This command provoked a crash in
1746 some environments:
1747
1748 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1749
1750 This potential security problem was recorded as CVE-2008-2371.
1751
1752 12. For a pattern where the match had to start at the beginning or immediately
1753 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1754 pcre_dfa_exec() could read past the end of the passed subject if there was
1755 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1756 pcretest so that it places the subject at the end of its malloc-ed buffer.
1757
1758 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1759 exec() might read past the end of the data buffer in UTF-8 mode.
1760
1761 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1762 the data contained the byte 0x85 as part of a UTF-8 character within its
1763 first line. This applied both to normal and DFA matching.
1764
1765 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1766 /^[^d]*?$/8 failed to match "abc".
1767
1768 16. Added a missing copyright notice to pcrecpp_internal.h.
1769
1770 17. Make it more clear in the documentation that values returned from
1771 pcre_exec() in ovector are byte offsets, not character counts.
1772
1773 18. Tidied a few places to stop certain compilers from issuing warnings.
1774
1775 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1776 supplied by Stefan Weber. I made a further small update for 7.8 because
1777 there is a change of source arrangements: the pcre_searchfuncs.c module is
1778 replaced by pcre_ucd.c.
1779
1780
1781 Version 7.7 07-May-08
1782 ---------------------
1783
1784 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1785 a string to a long long, pretend we don't even have a long long." This is
1786 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1787
1788 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1789 pre-7.6 versions, which defined a global no_arg variable instead of putting
1790 it in the RE class. (See also #8 below.)
1791
1792 3. Remove a line of dead code, identified by coverity and reported by Nuno
1793 Lopes.
1794
1795 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1796
1797 (1) The include/exclude patterns were being applied to the whole pathnames
1798 of files, instead of just to the final components.
1799
1800 (2) If there was more than one level of directory, the subdirectories were
1801 skipped unless they satisfied the include/exclude conditions. This is
1802 inconsistent with GNU grep (and could even be seen as contrary to the
1803 pcregrep specification - which I improved to make it absolutely clear).
1804 The action now is always to scan all levels of directory, and just
1805 apply the include/exclude patterns to regular files.
1806
1807 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1808 --exclude_dir in the tests to avoid scanning .svn directories.
1809
1810 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1811 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1812 doesn't support NULs in patterns.
1813
1814 7. Added some missing "const"s to declarations of static tables in
1815 pcre_compile.c and pcre_dfa_exec.c.
1816
1817 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1818 caused by fix #2 above. (Subsequently also a second patch to fix the
1819 first patch. And a third patch - this was a messy problem.)
1820
1821 9. Applied Craig's patch to remove the use of push_back().
1822
1823 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1824 matching function regexec().
1825
1826 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1827 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1828 references. PCRE supports relative numbers with this syntax (I don't think
1829 Oniguruma does).
1830
1831 12. Previously, a group with a zero repeat such as (...){0} was completely
1832 omitted from the compiled regex. However, this means that if the group
1833 was called as a subroutine from elsewhere in the pattern, things went wrong
1834 (an internal error was given). Such groups are now left in the compiled
1835 pattern, with a new opcode that causes them to be skipped at execution
1836 time.
1837
1838 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1839 to the way PCRE behaves:
1840
1841 (a) A lone ] character is dis-allowed (Perl treats it as data).
1842
1843 (b) A back reference to an unmatched subpattern matches an empty string
1844 (Perl fails the current match path).
1845
1846 (c) A data ] in a character class must be notated as \] because if the
1847 first data character in a class is ], it defines an empty class. (In
1848 Perl it is not possible to have an empty class.) The empty class []
1849 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1850 The negative empty class [^] matches any one character, independently
1851 of the DOTALL setting.
1852
1853 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1854 non-existent subpattern following a character class starting with ']' and
1855 containing () gave an internal compiling error instead of "reference to
1856 non-existent subpattern". Fortunately, when the pattern did exist, the
1857 compiled code was correct. (When scanning forwards to check for the
1858 existencd of the subpattern, it was treating the data ']' as terminating
1859 the class, so got the count wrong. When actually compiling, the reference
1860 was subsequently set up correctly.)
1861
1862 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1863 it was being rejected as not supported by pcre_dfa_exec(), even though
1864 other assertions are supported. I have made pcre_dfa_exec() support
1865 (*FAIL).
1866
1867 16. The implementation of 13c above involved the invention of a new opcode,
1868 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1869 cannot be changed at match time, I realized I could make a small
1870 improvement to matching performance by compiling OP_ALLANY instead of
1871 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1872 on the OP_ANY path.
1873
1874 17. Compiling pcretest on Windows with readline support failed without the
1875 following two fixes: (1) Make the unistd.h include conditional on
1876 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1877
1878 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1879 ncurses library to be included for pcretest when ReadLine support is
1880 requested, but also to allow for it to be overridden. This patch came from
1881 Daniel Bergström.
1882
1883 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1884 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1885 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1886 spotting this.
1887
1888
1889 Version 7.6 28-Jan-08
1890 ---------------------
1891
1892 1. A character class containing a very large number of characters with
1893 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1894 overflow.
1895
1896 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1897 HAVE_LONG_LONG is not defined.
1898
1899 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1900 bring them up to date and include new features. This patch includes:
1901
1902 - Fixed PH's badly added libz and libbz2 support.
1903 - Fixed a problem with static linking.
1904 - Added pcredemo. [But later removed - see 7 below.]
1905 - Fixed dftables problem and added an option.
1906 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1907 HAVE_LONG_LONG.
1908 - Added readline support for pcretest.
1909 - Added an listing of the option settings after cmake has run.
1910
1911 4. A user submitted a patch to Makefile that makes it easy to create
1912 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1913 Makefile.am that cause it to include this special target, without
1914 affecting anything else. Note that the same mingw target plus all
1915 the other distribution libraries and programs are now supported
1916 when configuring with CMake (see 6 below) instead of with
1917 Configure/Make.
1918
1919 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1920 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1921 exported in the Windows port". It has not yet been confirmed that the patch
1922 solves the problem, but it does no harm.
1923
1924 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1925 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1926 with CMake, and also correct the comment about stack recursion.
1927
1928 7. Remove the automatic building of pcredemo from the ./configure system and
1929 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1930 of a program that users should build themselves after PCRE is installed, so
1931 building it automatically is not really right. What is more, it gave
1932 trouble in some build environments.
1933
1934 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1935
1936
1937 Version 7.5 10-Jan-08
1938 ---------------------
1939
1940 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1941 values in parens when parsing an RE using the C++ wrapper."
1942
1943 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1944 Characters greater than 255 were excluded from the class instead of being
1945 included.
1946
1947 3. The same bug as (2) above applied to negated POSIX classes such as
1948 [:^space:].
1949
1950 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1951 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1952 I have changed it.
1953
1954 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1955 first named subpattern) and a construct such as (?&a) would reference the
1956 first named subpattern whose name started with "a" (in other words, the
1957 length check was missing). Both these problems are fixed. "Subpattern name
1958 expected" is now given for (?&) (a zero-length name), and this patch also
1959 makes it give the same error for \k'' (previously it complained that that
1960 was a reference to a non-existent subpattern).
1961
1962 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1963 this is right because (?- can be followed by option settings as well as by
1964 digits. I have, however, made the messages clearer.
1965
1966 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1967 than the number used in the conditional) now cause a compile-time error.
1968 This is actually not compatible with Perl, which accepts such patterns, but
1969 treats the conditional as always being FALSE (as PCRE used to), but it
1970 seems to me that giving a diagnostic is better.
1971
1972 8. Change "alphameric" to the more common word "alphanumeric" in comments
1973 and messages.
1974
1975 9. Fix two occurrences of "backslash" in comments that should have been
1976 "backspace".
1977
1978 10. Remove two redundant lines of code that can never be obeyed (their function
1979 was moved elsewhere).
1980
1981 11. The program that makes PCRE's Unicode character property table had a bug
1982 which caused it to generate incorrect table entries for sequences of
1983 characters that have the same character type, but are in different scripts.
1984 It amalgamated them into a single range, with the script of the first of
1985 them. In other words, some characters were in the wrong script. There were
1986 thirteen such cases, affecting characters in the following ranges:
1987
1988 U+002b0 - U+002c1
1989 U+0060c - U+0060d
1990 U+0061e - U+00612
1991 U+0064b - U+0065e
1992 U+0074d - U+0076d
1993 U+01800 - U+01805
1994 U+01d00 - U+01d77
1995 U+01d9b - U+01dbf
1996 U+0200b - U+0200f
1997 U+030fc - U+030fe
1998 U+03260 - U+0327f
1999 U+0fb46 - U+0fbb1
2000 U+10450 - U+1049d
2001
2002 12. The -o option (show only the matching part of a line) for pcregrep was not
2003 compatible with GNU grep in that, if there was more than one match in a
2004 line, it showed only the first of them. It now behaves in the same way as
2005 GNU grep.
2006
2007 13. If the -o and -v options were combined for pcregrep, it printed a blank
2008 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2009 does the same. The return code can be used to tell if there were any
2010 non-matching lines.
2011
2012 14. Added --file-offsets and --line-offsets to pcregrep.
2013
2014 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2015 infinitely looping recursion. The bug was that positive lookaheads were not
2016 being skipped when checking for a possible empty match (negative lookaheads
2017 and both kinds of lookbehind were skipped).
2018
2019 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2020 inclusion of <windows.h> to before rather than after the definition of
2021 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2022
2023 17. Specifying a possessive quantifier with a specific limit for a Unicode
2024 character property caused pcre_compile() to compile bad code, which led at
2025 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2026 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2027 caused the error; without that there was no problem.
2028
2029 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2030
2031 19. Added --enable-pcretest-libreadline.
2032
2033 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2034 RE::GlobalReplace(). As a result, the number of replacements returned was
2035 double what it should be. I removed one of the increments, but Craig sent a
2036 later patch that removed the other one (the right fix) and added unit tests
2037 that check the return values (which was not done before).
2038
2039 21. Several CMake things:
2040
2041 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2042 the names libpcre and libpcreposix, not just pcre and pcreposix.
2043
2044 (2) The above change means that pcretest and pcregrep are now correctly
2045 linked with the newly-built libraries, not previously installed ones.
2046
2047 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2048
2049 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2050 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2051 UTF-8 newline character). The key issue is that the pattern starts .*;
2052 this means that the match must be either at the beginning, or after a
2053 newline. The bug was in the code for advancing after a failed match and
2054 checking that the new position followed a newline. It was not taking
2055 account of UTF-8 characters correctly.
2056
2057 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2058 character classes. PCRE was not treating the sequence [:...:] as a
2059 character class unless the ... were all letters. Perl, however, seems to
2060 allow any characters between [: and :], though of course it rejects as
2061 unknown any "names" that contain non-letters, because all the known class
2062 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2063 for example, whereas PCRE did not - it did not recognize a POSIX character
2064 class. This seemed a bit dangerous, so the code has been changed to be
2065 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2066 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2067 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2068 Perl does, and where it didn't before.
2069
2070 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2071 Windows environments %n is disabled by default.
2072
2073
2074 Version 7.4 21-Sep-07
2075 ---------------------
2076
2077 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2078 means that a class such as [\s] counted as "explicit reference to CR or
2079 LF". That isn't really right - the whole point of the change was to try to
2080 help when there was an actual mention of one of the two characters. So now
2081 the change happens only if \r or \n (or a literal CR or LF) character is
2082 encountered.
2083
2084 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2085 of both had grown to the point where there were only 3 bits left.
2086 Fortunately, there was spare space in the data structure, and so I have
2087 moved the internal flags into a new 16-bit field to free up more option
2088 bits.
2089
2090 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2091 but did not set the internal JCHANGED flag - either of these is enough to
2092 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2093 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2094 start sets both bits.
2095
2096 4. Added options (at build time, compile time, exec time) to change \R from
2097 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2098
2099 5. doc/pcresyntax.html was missing from the distribution.
2100
2101 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2102 compatibility, even though it is no longer used.
2103
2104 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2105 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2106 windows.h file is present (where different names are used). [This was
2107 reversed later after testing - see 16 below.]
2108
2109 8. Changed all #include <config.h> to #include "config.h". There were also
2110 some further <pcre.h> cases that I changed to "pcre.h".
2111
2112 9. When pcregrep was used with the --colour option, it missed the line ending
2113 sequence off the lines that it output.
2114
2115 10. It was pointed out to me that arrays of string pointers cause lots of
2116 relocations when a shared library is dynamically loaded. A technique of
2117 using a single long string with a table of offsets can drastically reduce
2118 these. I have refactored PCRE in four places to do this. The result is
2119 dramatic:
2120
2121 Originally: 290
2122 After changing UCP table: 187
2123 After changing error message table: 43
2124 After changing table of "verbs" 36
2125 After changing table of Posix names 22
2126
2127 Thanks to the folks working on Gregex for glib for this insight.
2128
2129 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2130 unicode-properties was also set.
2131
2132 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2133
2134 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2135 checked only for CRLF.
2136
2137 14. Added casts to pcretest.c to avoid compiler warnings.
2138
2139 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2140
2141 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2142 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2143 entirely. This removes changes made in 7 above.
2144
2145 17. The CMake files have been updated, and there is now more information about
2146 building with CMake in the NON-UNIX-USE document.
2147
2148
2149 Version 7.3 28-Aug-07
2150 ---------------------
2151
2152 1. In the rejigging of the build system that eventually resulted in 7.1, the
2153 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2154 brackets there is not right, since it causes compilers to look for an
2155 installed pcre.h, not the version that is in the source that is being
2156 compiled (which of course may be different). I have changed it back to:
2157
2158 #include "pcre.h"
2159
2160 I have a vague recollection that the change was concerned with compiling in
2161 different directories, but in the new build system, that is taken care of
2162 by the VPATH setting the Makefile.
2163
2164 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2165 when the subject happened to end in the byte 0x85 (e.g. if the last
2166 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2167 characters but of course it shouldn't be taken as a newline when it is part
2168 of another character. The bug was that, for an unlimited repeat of . in
2169 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2170 characters when looking for a newline.
2171
2172 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2173
2174 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2175 in debug output.
2176
2177 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2178 long printing in the pcrecpp unittest when running under MinGW.
2179
2180 6. ESC_K was left out of the EBCDIC table.
2181
2182 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2183 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2184 limit also applies to "virtual nesting" when a pattern is recursive, and in
2185 this case 1000 isn't so big. I have been able to remove this limit at the
2186 expense of backing off one optimization in certain circumstances. Normally,
2187 when pcre_exec() would call its internal match() function recursively and
2188 immediately return the result unconditionally, it uses a "tail recursion"
2189 feature to save stack. However, when a subpattern that can match an empty
2190 string has an unlimited repetition quantifier, it no longer makes this
2191 optimization. That gives it a stack frame in which to save the data for
2192 checking that an empty string has been matched. Previously this was taken
2193 from the 1000-entry workspace that had been reserved. So now there is no
2194 explicit limit, but more stack is used.
2195
2196 8. Applied Daniel's patches to solve problems with the import/export magic
2197 syntax that is required for Windows, and which was going wrong for the
2198 pcreposix and pcrecpp parts of the library. These were overlooked when this
2199 problem was solved for the main library.
2200
2201 9. There were some crude static tests to avoid integer overflow when computing
2202 the size of patterns that contain repeated groups with explicit upper
2203 limits. As the maximum quantifier is 65535, the maximum group length was
2204 set at 30,000 so that the product of these two numbers did not overflow a
2205 32-bit integer. However, it turns out that people want to use groups that
2206 are longer than 30,000 bytes (though not repeat them that many times).
2207 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2208 made it possible to implement the integer overflow checks in a much more
2209 dynamic way, which I have now done. The artificial limitation on group
2210 length has been removed - we now have only the limit on the total length of
2211 the compiled pattern, which depends on the LINK_SIZE setting.
2212
2213 10. Fixed a bug in the documentation for get/copy named substring when
2214 duplicate names are permitted. If none of the named substrings are set, the
2215 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2216 empty string.
2217
2218 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2219 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2220 because the ] is interpreted as the first data character and the
2221 terminating ] is not found. PCRE has been made compatible with Perl in this
2222 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2223 cause memory overwriting.
2224
2225 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2226 string has been matched (to stop an infinite loop). It was not recognizing
2227 a conditional subpattern that could match an empty string if that
2228 subpattern was within another subpattern. For example, it looped when
2229 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2230 condition was not nested. This bug has been fixed.
2231
2232 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2233 past the start of the subject in the presence of bytes with the top bit
2234 set, for example "\x8aBCD".
2235
2236 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2237 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2238
2239 14. Optimized (?!) to (*FAIL).
2240
2241 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2242 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2243 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2244 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2245 does: it's just the validity check that is more restrictive.
2246
2247 16. Inserted checks for integer overflows during escape sequence (backslash)
2248 processing, and also fixed erroneous offset values for syntax errors during
2249 backslash processing.
2250
2251 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2252 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2253
2254 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2255 caused an overrun.
2256
2257 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2258 something other than just ASCII characters) inside a group that had an
2259 unlimited repeat caused a loop at compile time (while checking to see
2260 whether the group could match an empty string).
2261
2262 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2263 [\P{Any}] did so. (Error in the code for printing property names.)
2264
2265 21. An orphan \E inside a character class could cause a crash.
2266
2267 22. A repeated capturing bracket such as (A)? could cause a wild memory
2268 reference during compilation.
2269
2270 23. There are several functions in pcre_compile() that scan along a compiled
2271 expression for various reasons (e.g. to see if it's fixed length for look
2272 behind). There were bugs in these functions when a repeated \p or \P was
2273 present in the pattern. These operators have additional parameters compared
2274 with \d, etc, and these were not being taken into account when moving along
2275 the compiled data. Specifically:
2276
2277 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2278 length.
2279
2280 (b) An item such as \pL+ within a repeated group could cause crashes or
2281 loops.
2282
2283 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2284 "reference to non-existent subpattern" error.
2285
2286 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2287
2288 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2289 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2290
2291 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2292
2293 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2294 character were causing crashes (broken optimization).
2295
2296 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2297 \p or \P) caused a compile-time loop.
2298
2299 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2300 break. For example, the unstudied pattern [\r\n]A does not match the string
2301 "\r\nA" because change 7.0/46 below moves the current point on by two
2302 characters after failing to match at the start. However, the pattern \nA
2303 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2304 the same is true. There doesn't seem any very clean way out of this, but
2305 what I have chosen to do makes the common cases work: PCRE now takes note
2306 of whether there can be an explicit match for \r or \n anywhere in the
2307 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2308 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2309 pattern has explicit CR or LF references.
2310
2311 29. Added (*CR) etc for changing newline setting at start of pattern.
2312
2313
2314 Version 7.2 19-Jun-07
2315 ---------------------
2316
2317 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2318 which is apparently normally available under Windows.
2319
2320 2. Re-jig the pcregrep tests with different newline settings in an attempt
2321 to make them independent of the local environment's newline setting.
2322
2323 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2324
2325 4. Some of the "internals" tests were previously cut out when the link size
2326 was not 2, because the output contained actual offsets. The recent new
2327 "Z" feature of pcretest means that these can be cut out, making the tests
2328 usable with all link sizes.
2329
2330 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2331 stack recursion. This gives a massive performance boost under BSD, but just
2332 a small improvement under Linux. However, it saves one field in the frame
2333 in all cases.
2334
2335 6. Added more features from the forthcoming Perl 5.10:
2336
2337 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2338 recursion call. It refers to the nth most recently opened parentheses.
2339
2340 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2341 to be opened parentheses.
2342
2343 (c) Conditions that refer to capturing parentheses can be specified
2344 relatively, for example, (?(-2)... or (?(+3)...
2345
2346 (d) \K resets the start of the current match so that everything before
2347 is not part of it.
2348
2349 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2350
2351 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2352 reference syntax.
2353
2354 (g) (?| introduces a group in which the numbering of parentheses in each
2355 alternative starts with the same number.
2356
2357 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2358
2359 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2360 PCRE_INFO_JCHANGED.
2361
2362 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2363 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2364 for detecting groups that can match an empty string.
2365
2366 9. A pattern with a very large number of alternatives (more than several
2367 hundred) was running out of internal workspace during the pre-compile
2368 phase, where pcre_compile() figures out how much memory will be needed. A
2369 bit of new cunning has reduced the workspace needed for groups with
2370 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2371 workspace instead of running out of the 4096 that are available.
2372
2373 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2374
2375 11. Applied patch from Google to remove an optimization that didn't quite work.
2376 The report of the bug said:
2377
2378 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2379 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2380 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2381
2382 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2383 it matched the wrong number of bytes.
2384
2385
2386 Version 7.1 24-Apr-07
2387 ---------------------
2388
2389 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2390 that is more "standard", making use of automake and other Autotools. There
2391 is some re-arrangement of the files and adjustment of comments consequent
2392 on this.
2393
2394 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2395 for recursive directory scanning broke on some systems because the files
2396 are not scanned in any specific order and on different systems the order
2397 was different. A call to "sort" has been inserted into RunGrepTest for the
2398 approprate test as a short-term fix. In the longer term there may be an
2399 alternative.
2400
2401 3. I had an email from Eric Raymond about problems translating some of PCRE's
2402 man pages to HTML (despite the fact that I distribute HTML pages, some
2403 people do their own conversions for various reasons). The problems
2404 concerned the use of low-level troff macros .br and .in. I have therefore
2405 removed all such uses from the man pages (some were redundant, some could
2406 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2407 HTML has been updated to handle .nf/.fi and to complain if it encounters
2408 .br or .in.
2409
2410 4. Updated comments in configure.ac that get placed in config.h.in and also
2411 arranged for config.h to be included in the distribution, with the name
2412 config.h.generic, for the benefit of those who have to compile without
2413 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2414
2415 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2416 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2417 makevp.bat for the current PCRE, using the additional files
2418 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2419
2420 6. A Windows user reported a minor discrepancy with test 2, which turned out
2421 to be caused by a trailing space on an input line that had got lost in his
2422 copy. The trailing space was an accident, so I've just removed it.
2423
2424 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2425 that is needed.
2426
2427 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2428 as "const" (a) because they are and (b) because it helps the PHP
2429 maintainers who have recently made a script to detect big data structures
2430 in the php code that should be moved to the .rodata section. I remembered
2431 to update Builducptable as well, so it won't revert if ucptable.h is ever
2432 re-created.
2433
2434 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2435 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2436 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2437 support is not required. This saves 1.5-2K of code, which is important in
2438 some applications.
2439
2440 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2441 so as not to refer to the tables, even though these functions will never be
2442 called when UTF-8 support is disabled. Otherwise there are problems with a
2443 shared library.
2444
2445 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2446
2447 (a) It was defining its arguments as char * instead of void *.
2448
2449 (b) It was assuming that all moves were upwards in memory; this was true
2450 a long time ago when I wrote it, but is no longer the case.
2451
2452 The emulated memove() is provided for those environments that have neither
2453 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2454 is clearly not the case, as these two bugs were recently reported.
2455
2456 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2457 and Detrail to create the HTML documentation, the .txt form of the man
2458 pages, and it removes trailing spaces from listed files. It also creates
2459 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2460 case, it wraps all the #defines with #ifndefs. This script should be run
2461 before "make dist".
2462
2463 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2464 with Unicode property support.
2465
2466 (a) For a maximizing quantifier, if the two different cases of the
2467 character were of different lengths in their UTF-8 codings (there are
2468 some cases like this - I found 11), and the matching function had to
2469 back up over a mixture of the two cases, it incorrectly assumed they
2470 were both the same length.
2471
2472 (b) When PCRE was configured to use the heap rather than the stack for
2473 recursion during matching, it was not correctly preserving the data for
2474 the other case of a UTF-8 character when checking ahead for a match
2475 while processing a minimizing repeat. If the check also involved
2476 matching a wide character, but failed, corruption could cause an
2477 erroneous result when trying to check for a repeat of the original
2478 character.
2479
2480 13. Some tidying changes to the testing mechanism:
2481
2482 (a) The RunTest script now detects the internal link size and whether there
2483 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2484 values substituted by "configure". (The RunGrepTest script already did
2485 this for UTF-8.) The configure.ac script no longer substitutes the
2486 relevant variables.
2487
2488 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2489 with length and offset values. This means that the output is different
2490 for different internal link sizes. Test 2 is skipped for link sizes
2491 other than 2 because of this, bypassing the problem. Unfortunately,
2492 there was also a test in test 3 (the locale tests) that used /B and
2493 failed for link sizes other than 2. Rather than cut the whole test out,
2494 I have added a new /Z option to pcretest that replaces the length and
2495 offset values with spaces. This is now used to make test 3 independent
2496 of link size. (Test 2 will be tidied up later.)
2497
2498 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2499 segmentation fault instead of returning the appropriate error message.
2500
2501 15. In multiline mode when the newline sequence was set to "any", the pattern
2502 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2503 This doesn't seem right; it now treats the CRLF combination as the line
2504 ending, and so does not match in that case. It's only a pattern such as ^$
2505 that would hit this one: something like ^ABC$ would have failed after \r
2506 and then tried again after \r\n.
2507
2508 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2509 in an attempt to make files that differ only in their line terminators
2510 compare equal. This works on Linux.
2511
2512 17. Under certain error circumstances pcregrep might try to free random memory
2513 as it exited. This is now fixed, thanks to valgrind.
2514
2515 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2516 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2517 was because its rules for how to advance for /g after matching an empty
2518 string at the end of a line did not allow for this case. They now check for
2519 it specially.
2520
2521 20. pcretest is supposed to handle patterns and data of any length, by
2522 extending its buffers when necessary. It was getting this wrong when the
2523 buffer for a data line had to be extended.
2524
2525 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2526 CRLF as a newline sequence.
2527
2528 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2529 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2530 I have nevertheless tidied it up.
2531
2532 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2533
2534 24. Added a man page for pcre-config.
2535
2536
2537 Version 7.0 19-Dec-06
2538 ---------------------
2539
2540 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2541 moving to gcc 4.1.1.
2542
2543 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2544 sys/time.h, which is documented as needed for this function. It doesn't
2545 seem to matter on Linux, but it showed up on some releases of OS X.
2546
2547 3. It seems that there are systems where bytes whose values are greater than
2548 127 match isprint() in the "C" locale. The "C" locale should be the
2549 default when a C program starts up. In most systems, only ASCII printing
2550 characters match isprint(). This difference caused the output from pcretest
2551 to vary, making some of the tests fail. I have changed pcretest so that:
2552
2553 (a) When it is outputting text in the compiled version of a pattern, bytes
2554 other than 32-126 are always shown as hex escapes.
2555
2556 (b) When it is outputting text that is a matched part of a subject string,
2557 it does the same, unless a different locale has been set for the match
2558 (using the /L modifier). In this case, it uses isprint() to decide.
2559
2560 4. Fixed a major bug that caused incorrect computation of the amount of memory
2561 required for a compiled pattern when options that changed within the
2562 pattern affected the logic of the preliminary scan that determines the
2563 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2564 that the computed length was too small. The symptoms of this bug were
2565 either the PCRE error "internal error: code overflow" from pcre_compile(),
2566 or a glibc crash with a message such as "pcretest: free(): invalid next
2567 size (fast)". Examples of patterns that provoked this bug (shown in
2568 pcretest format) are:
2569
2570 /(?-x: )/x
2571 /(?x)(?-x: \s*#\s*)/
2572 /((?i)[\x{c0}])/8
2573 /(?i:[\x{c0}])/8
2574
2575 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2576 is now done differently.
2577
2578 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2579 wrapper classes; (b) implement a new function in the C++ scanner that is
2580 more efficient than the old way of doing things because it avoids levels of
2581 recursion in the regex matching; (c) add a paragraph to the documentation
2582 for the FullMatch() function.
2583
2584 6. The escape sequence \n was being treated as whatever was defined as
2585 "newline". Not only was this contrary to the documentation, which states
2586 that \n is character 10 (hex 0A), but it also went horribly wrong when
2587 "newline" was defined as CRLF. This has been fixed.
2588
2589 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2590 was being set to -1 for the "end of line" case (supposedly a value that no
2591 character can have). Though this value is never used (the check for end of
2592 line is "zero bytes in current character"), it caused compiler complaints.
2593 I've changed it to 0xffffffff.
2594
2595 8. In pcre_version.c, the version string was being built by a sequence of
2596 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2597 string (as it is for production releases) called a macro with an empty
2598 argument. The C standard says the result of this is undefined. The gcc
2599 compiler treats it as an empty string (which was what was wanted) but it is
2600 reported that Visual C gives an error. The source has been hacked around to
2601 avoid this problem.
2602
2603 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2604 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2605 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2606 of them did).
2607
2608 10. Originally, pcretest opened its input and output without "b"; then I was
2609 told that "b" was needed in some environments, so it was added for release
2610 5.0 to both the input and output. (It makes no difference on Unix-like
2611 systems.) Later I was told that it is wrong for the input on Windows. I've
2612 now abstracted the modes into two macros, to make it easier to fiddle with
2613 them, and removed "b" from the input mode under Windows.
2614
2615 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2616
2617 12. Added -help and --help to pcretest as an official way of being reminded
2618 of the options.
2619
2620 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2621 and pcrecpp.cc because they annoy compilers at high warning levels.
2622
2623 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2624
2625 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2626 scripts are not C programs :-) and which was not noticed because it works
2627 on Linux.
2628
2629 16. pcretest is supposed to handle any length of pattern and data line (as one
2630 line or as a continued sequence of lines) by extending its input buffer if
2631 necessary. This feature was broken for very long pattern lines, leading to
2632 a string of junk being passed to pcre_compile() if the pattern was longer
2633 than about 50K.
2634
2635 17. I have done a major re-factoring of the way pcre_compile() computes the
2636 amount of memory needed for a compiled pattern. Previously, there was code
2637 that made a preliminary scan of the pattern in order to do this. That was
2638 OK when PCRE was new, but as the facilities have expanded, it has become
2639 harder and harder to keep it in step with the real compile phase, and there
2640 have been a number of bugs (see for example, 4 above). I have now found a
2641 cunning way of running the real compile function in a "fake" mode that
2642 enables it to compute how much memory it would need, while actually only
2643 ever using a few hundred bytes of working memory and without too many
2644 tests of the mode. This should make future maintenance and development
2645 easier. A side effect of this work is that the limit of 200 on the nesting
2646 depth of parentheses has been removed (though this was never a serious
2647 limitation, I suspect). However, there is a downside: pcre_compile() now
2648 runs more slowly than before (30% or more, depending on the pattern). I
2649 hope this isn't a big issue. There is no effect on runtime performance.
2650
2651 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2652 newline (only possible for the last line of a file) and it was a
2653 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2654
2655 19. Added additional timing features to pcretest. (1) The -tm option now times
2656 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2657 separate command line item, by a number that specifies the number of
2658 repeats to use when timing. The default is 50000; this gives better
2659 precision, but takes uncomfortably long for very large patterns.
2660
2661 20. Extended pcre_study() to be more clever in cases where a branch of a
2662 subpattern has no definite first character. For example, (a*|b*)[cd] would
2663 previously give no result from pcre_study(). Now it recognizes that the
2664 first character must be a, b, c, or d.
2665
2666 21. There was an incorrect error "recursive call could loop indefinitely" if
2667 a subpattern (or the entire pattern) that was being tested for matching an
2668 empty string contained only one non-empty item after a nested subpattern.
2669 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2670 incorrectly, because the \d was being skipped in the check.
2671
2672 22. The pcretest program now has a new pattern option /B and a command line
2673 option -b, which is equivalent to adding /B to every pattern. This causes
2674 it to show the compiled bytecode, without the additional information that
2675 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2676 is the same as /B/I).
2677
2678 23. A new optimization is now able automatically to treat some sequences such
2679 as a*b as a*+b. More specifically, if something simple (such as a character
2680 or a simple class like \d) has an unlimited quantifier, and is followed by
2681 something that cannot possibly match the quantified thing, the quantifier
2682 is automatically "possessified".
2683
2684 24. A recursive reference to a subpattern whose number was greater than 39
2685 went wrong under certain circumstances in UTF-8 mode. This bug could also
2686 have affected the operation of pcre_study().
2687
2688 25. Realized that a little bit of performance could be had by replacing
2689 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2690
2691 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2692
2693 27. Possessive quantifiers such as a++ were previously implemented by turning
2694 them into atomic groups such as ($>a+). Now they have their own opcodes,
2695 which improves performance. This includes the automatically created ones
2696 from 23 above.
2697
2698 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2699 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2700 the first matched character to be a colon. This applied both to named and
2701 numbered groups.
2702
2703 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2704
2705 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2706 building PCRE on Windows easier, so I have included it in the distribution.
2707
2708 31. There is now a check in pcretest against a ridiculously large number being
2709 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2710 loop, the loop is abandoned.
2711
2712 32. Forward references to subpatterns in conditions such as (?(2)...) where
2713 subpattern 2 is defined later cause pcre_compile() to search forwards in
2714 the pattern for the relevant set of parentheses. This search went wrong
2715 when there were unescaped parentheses in a character class, parentheses
2716 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2717
2718 33. "Subroutine" calls and backreferences were previously restricted to
2719 referencing subpatterns earlier in the regex. This restriction has now
2720 been removed.
2721
2722 34. Added a number of extra features that are going to be in Perl 5.10. On the
2723 whole, these are just syntactic alternatives for features that PCRE had
2724 previously implemented using the Python syntax or my own invention. The
2725 other formats are all retained for compatibility.
2726
2727 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2728 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2729 also .NET compatible.
2730
2731 (b) A recursion or subroutine call to a named group can now be defined as
2732 (?&name) as well as (?P>name).
2733
2734 (c) A backreference to a named group can now be defined as \k<name> or
2735 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2736 5.10, are also .NET compatible.
2737
2738 (d) A conditional reference to a named group can now use the syntax
2739 (?(<name>) or (?('name') as well as (?(name).
2740
2741 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2742 groups (named and numbered) that are never evaluated inline, but can be
2743 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2744 is always false. There may be only one alternative in such a group.
2745
2746 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2747 as the simple (?(R). The condition is true only if the most recent
2748 recursion is that of the given number or name. It does not search out
2749 through the entire recursion stack.
2750
2751 (g) The escape \gN or \g{N} has been added, where N is a positive or
2752 negative number, specifying an absolute or relative reference.
2753
2754 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2755 some "unreachable code" warnings.
2756
2757 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2758 things, this adds five new scripts.
2759
2760 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2761 There were also incompatibilities regarding the handling of \Q..\E inside
2762 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2763 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2764
2765 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2766 matches an empty string, and forcibly breaks the loop. There were bugs in
2767 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2768 against aaaa the result was just "a" rather than "aaaa", for example. Two
2769 separate and independent bugs (that affected different cases) have been
2770 fixed.
2771
2772 39. Refactored the code to abolish the use of different opcodes for small
2773 capturing bracket numbers. This is a tidy that I avoided doing when I
2774 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2775 The new approach is not only tidier, it makes it possible to reduce the
2776 memory needed to fix the previous bug (38).
2777
2778 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2779 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2780 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2781 mode.
2782
2783 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2784 report.
2785
2786 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2787 copy construction and assignment in the C++ wrapper.
2788
2789 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2790 couple of bugs that could have given wrong results in the "--newline=crlf"
2791 case.
2792
2793 44. Added a number of casts and did some reorganization of signed/unsigned int
2794 variables following suggestions from Dair Grant. Also renamed the variable
2795 "this" as "item" because it is a C++ keyword.
2796
2797 45. Arranged for dftables to add
2798
2799 #include "pcre_internal.h"
2800
2801 to pcre_chartables.c because without it, gcc 4.x may remove the array
2802 definition from the final binary if PCRE is built into a static library and
2803 dead code stripping is activated.
2804
2805 46. For an unanchored pattern, if a match attempt fails at the start of a
2806 newline sequence, and the newline setting is CRLF or ANY, and the next two
2807 characters are CRLF, advance by two characters instead of one.
2808
2809
2810 Version 6.7 04-Jul-06
2811 ---------------------
2812
2813 1. In order to handle tests when input lines are enormously long, pcretest has
2814 been re-factored so that it automatically extends its buffers when
2815 necessary. The code is crude, but this _is_ just a test program. The
2816 default size has been increased from 32K to 50K.
2817
2818 2. The code in pcre_study() was using the value of the re argument before
2819 testing it for NULL. (Of course, in any sensible call of the function, it
2820 won't be NULL.)
2821
2822 3. The memmove() emulation function in pcre_internal.h, which is used on
2823 systems that lack both memmove() and bcopy() - that is, hardly ever -
2824 was missing a "static" storage class specifier.
2825
2826 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2827 containing an extended class (one that cannot be represented by a bitmap
2828 because it contains high-valued characters or Unicode property items, e.g.
2829 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2830 pattern, but PCRE should not loop if you do not (it no longer does).
2831 [Detail: two cases were found: (a) a repeated subpattern containing an
2832 extended class; (b) a recursive reference to a subpattern that followed a
2833 previous extended class. It wasn't skipping over the extended class
2834 correctly when UTF-8 mode was not set.]
2835
2836 5. A negated single-character class was not being recognized as fixed-length
2837 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2838 compile error "lookbehind assertion is not fixed length".
2839
2840 6. The RunPerlTest auxiliary script was showing an unexpected difference
2841 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2842 write a Perl script that can interpret lines of an input file either as
2843 byte characters or as UTF-8, which is what "perltest" was being required to
2844 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2845 can't do is switch easily at run time between having the "use utf8;" pragma
2846 or not. In the end, I fudged it by using the RunPerlTest script to insert
2847 "use utf8;" explicitly for the UTF-8 tests.
2848
2849 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2850 the end of the subject string, contrary to the documentation and to what
2851 Perl does. This was true of both matching functions. Now it matches only at
2852 the start of the subject and immediately after *internal* newlines.
2853
2854 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2855 a pointer to an int instead of a pointer to an unsigned long int. This
2856 caused problems on 64-bit systems.
2857
2858 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2859 instance of the 'standard' template library not being so standard".
2860
2861 10. There was no check on the number of named subpatterns nor the maximum
2862 length of a subpattern name. The product of these values is used to compute
2863 the size of the memory block for a compiled pattern. By supplying a very
2864 long subpattern name and a large number of named subpatterns, the size
2865 computation could be caused to overflow. This is now prevented by limiting
2866 the length of names to 32 characters, and the number of named subpatterns
2867 to 10,000.
2868
2869 11. Subpatterns that are repeated with specific counts have to be replicated in
2870 the compiled pattern. The size of memory for this was computed from the
2871 length of the subpattern and the repeat count. The latter is limited to
2872 65535, but there was no limit on the former, meaning that integer overflow
2873 could in principle occur. The compiled length of a repeated subpattern is
2874 now limited to 30,000 bytes in order to prevent this.
2875
2876 12. Added the optional facility to have named substrings with the same name.
2877
2878 13. Added the ability to use a named substring as a condition, using the
2879 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2880 are numbers (not recommended). Forward references are permitted.
2881
2882 14. Added forward references in named backreferences (if you see what I mean).
2883
2884 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2885 pattern could run off the end of the subject. For example, the pattern
2886 "(?s)(.{1,5})"8 did this with the subject "ab".
2887
2888 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2889 PCRE_CASELESS was set when matching characters that were quantified with ?
2890 or *.
2891
2892 17. A character class other than a single negated character that had a minimum
2893 but no maximum quantifier - for example [ab]{6,} - was not handled
2894 correctly by pce_dfa_exec(). It would match only one character.
2895
2896 18. A valid (though odd) pattern that looked like a POSIX character
2897 class but used an invalid character after [ (for example [[,abc,]]) caused
2898 pcre_compile() to give the error "Failed: internal error: code overflow" or
2899 in some cases to crash with a glibc free() error. This could even happen if
2900 the pattern terminated after [[ but there just happened to be a sequence of
2901 letters, a binary zero, and a closing ] in the memory that followed.
2902
2903 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2904 over the years. Originally (before any Unicode support), just the bottom 8
2905 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2906 output from "man perlunicode" includes this:
2907
2908 The regular expression compiler produces polymorphic opcodes. That
2909 is, the pattern adapts to the data and automatically switches to
2910 the Unicode character scheme when presented with Unicode data--or
2911 instead uses a traditional byte scheme when presented with byte
2912 data.
2913
2914 Sadly, a wide octal escape does not cause a switch, and in a string with
2915 no other multibyte characters, these octal escapes are treated as before.
2916 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2917 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2918 Unicode string.
2919
2920 I have not perpetrated such confusion in PCRE. Up till now, it took just
2921 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2922 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2923 translate to the appropriate multibyte character.
2924
2925 29. Applied some refactoring to reduce the number of warnings from Microsoft
2926 and Borland compilers. This has included removing the fudge introduced
2927 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2928 a warning about an unused variable.
2929
2930 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2931 characters since release 4.0, because Perl (from release 5.004) does not.
2932 [Or at least, is documented not to: some releases seem to be in conflict
2933 with the documentation.] However, when a pattern was studied with
2934 pcre_study() and all its branches started with \s, PCRE still included VT
2935 as a possible starting character. Of course, this did no harm; it just
2936 caused an unnecessary match attempt.
2937
2938 22. Removed a now-redundant internal flag bit that recorded the fact that case
2939 dependency changed within the pattern. This was once needed for "required
2940 byte" processing, but is no longer used. This recovers a now-scarce options
2941 bit. Also moved the least significant internal flag bit to the most-
2942 significant bit of the word, which was not previously used (hangover from
2943 the days when it was an int rather than a uint) to free up another bit for
2944 the future.
2945
2946 23. Added support for CRLF line endings as well as CR and LF. As well as the
2947 default being selectable at build time, it can now be changed at runtime
2948 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2949 specify that it is scanning data with non-default line endings.
2950
2951 24. Changed the definition of CXXLINK to make it agree with the definition of
2952 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2953
2954 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2955 recursions. This makes a big different to stack usage for some patterns.
2956
2957 26. If a subpattern containing a named recursion or subroutine reference such
2958 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2959 the space required for the compiled pattern went wrong and gave too small a
2960 value. Depending on the environment, this could lead to "Failed: internal
2961 error: code overflow at offset 49" or "glibc detected double free or
2962 corruption" errors.
2963
2964 27. Applied patches from Google (a) to support the new newline modes and (b) to
2965 advance over multibyte UTF-8 characters in GlobalReplace.
2966
2967 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2968 difference for some implementation of PCRE in some Windows version.
2969
2970 29. Added some extra testing facilities to pcretest:
2971
2972 \q<number> in a data line sets the "match limit" value
2973 \Q<number> in a data line sets the "match recursion limt" value
2974 -S <number> sets the stack size, where <number> is in megabytes
2975
2976 The -S option isn't available for Windows.
2977
2978
2979 Version 6.6 06-Feb-06
2980 ---------------------
2981
2982 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2983 in pcreposix.h. I have copied the definition from pcre.h.
2984
2985 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2986 because pcre.h is no longer a built file.
2987
2988 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2989 not normally included in the compiled code.
2990
2991
2992 Version 6.5 01-Feb-06
2993 ---------------------
2994
2995 1. When using the partial match feature with pcre_dfa_exec(), it was not
2996 anchoring the second and subsequent partial matches at the new starting
2997 point. This could lead to incorrect results. For example, with the pattern
2998 /1234/, partially matching against "123" and then "a4" gave a match.
2999
3000 2. Changes to pcregrep:
3001
3002 (a) All non-match returns from pcre_exec() were being treated as failures
3003 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3004 error message is output. Some extra information is given for the
3005 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3006 probably the only errors that are likely to be caused by users (by
3007 specifying a regex that has nested indefinite repeats, for instance).
3008 If there are more than 20 of these errors, pcregrep is abandoned.
3009
3010 (b) A binary zero was treated as data while matching, but terminated the
3011 output line if it was written out. This has been fixed: binary zeroes
3012 are now no different to any other data bytes.
3013
3014 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3015 used to set a locale for matching. The --locale=xxxx long option has
3016 been added (no short equivalent) to specify a locale explicitly on the
3017 pcregrep command, overriding the environment variables.
3018
3019 (d) When -B was used with -n, some line numbers in the output were one less
3020 than they should have been.
3021
3022 (e) Added the -o (--only-matching) option.
3023
3024 (f) If -A or -C was used with -c (count only), some lines of context were
3025 accidentally printed for the final match.
3026
3027 (g) Added the -H (--with-filename) option.
3028
3029 (h) The combination of options -rh failed to suppress file names for files
3030 that were found from directory arguments.
3031
3032 (i) Added the -D (--devices) and -d (--directories) options.
3033
3034 (j) Added the -F (--fixed-strings) option.
3035
3036 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3037
3038 (l) Added the --colo(u)r option.
3039
3040 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3041 is not present by default.
3042
3043 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3044 items such as (?R) or (?1), when the recursion could match a number of
3045 alternatives. If it matched one of the alternatives, but subsequently,
3046 outside the recursion, there was a failure, the code tried to back up into
3047 the recursion. However, because of the way PCRE is implemented, this is not
3048 possible, and the result was an incorrect result from the match.
3049
3050 In order to prevent this happening, the specification of recursion has
3051 been changed so that all such subpatterns are automatically treated as
3052 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3053
3054 4. I had overlooked the fact that, in some locales, there are characters for
3055 which isalpha() is true but neither isupper() nor islower() are true. In
3056 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3057 and ordfeminine) are like this. This affected the treatment of \w and \W
3058 when they appeared in character classes, but not when they appeared outside
3059 a character class. The bit map for "word" characters is now created
3060 separately from the results of isalnum() instead of just taking it from the
3061 upper, lower, and digit maps. (Plus the underscore character, of course.)
3062
3063 5. The above bug also affected the handling of POSIX character classes such as
3064 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3065 permanent tables. Instead, the bit maps for such a class were previously
3066 created as the appropriate unions of the upper, lower, and digit bitmaps.
3067 Now they are created by subtraction from the [[:word:]] class, which has
3068 its own bitmap.
3069
3070 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3071 It is created by subtracting the vertical space characters (\x09, \x0a,
3072 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3073 subtraction was done in the overall bitmap for a character class, meaning
3074 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3075 be recognized. This bug has been fixed.
3076
3077 7. Patches from the folks at Google:
3078
3079 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3080 real life, but is still worth protecting against".
3081
3082 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3083 regular expressions".
3084
3085 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3086 have it.
3087
3088 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3089 "configure" and the latter not, in order to fix a problem somebody had
3090 with compiling the Arg class on HP-UX.
3091
3092 (e) Improve the error-handling of the C++ wrapper a little bit.
3093
3094 (f) New tests for checking recursion limiting.
3095
3096 8. The pcre_memmove() function, which is used only if the environment does not
3097 have a standard memmove() function (and is therefore rarely compiled),
3098 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3099 returning a result (though PCRE never actually uses the result).
3100
3101 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3102 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3103 returned instead of calling malloc() with an overflowing number that would
3104 most likely cause subsequent chaos.
3105
3106 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3107
3108 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3109 with this option is matched, the nmatch and pmatch options of regexec() are
3110 ignored.
3111
3112 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3113 provided in case anyone wants to the the POSIX interface with UTF-8
3114 strings.
3115
3116 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3117 C++ linking (needed for some HP-UX environments).
3118
3119 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3120 (unused parameter) and in the pcre_printint() function (omitted "default"
3121 switch label when the default is to do nothing).
3122
3123 15. Added some code to make it possible, when PCRE is compiled as a C++
3124 library, to replace subject pointers for pcre_exec() with a smart pointer
3125 class, thus making it possible to process discontinuous strings.
3126
3127 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3128 much the same function. They were added by different people who were trying
3129 to make PCRE easy to compile on non-Unix systems. It has been suggested
3130 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3131 for compiling on Windows systems. I have therefore replaced it with
3132 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3133 defaults to "extern" for C or "extern C" for C++, which works fine on
3134 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3135 SCOPE with something explicit in config.h. In addition:
3136
3137 (a) pcreposix.h still had just "extern" instead of either of these macros;
3138 I have replaced it with PCRE_DATA_SCOPE.
3139
3140 (b) Functions such as _pcre_xclass(), which are internal to the library,
3141 but external in the C sense, all had PCRE_EXPORT in their definitions.
3142 This is apparently wrong for the Windows case, so I have removed it.
3143 (It makes no difference on Unix-like systems.)
3144
3145 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3146 of recursive calls to match(). This is different to MATCH_LIMIT because
3147 that limits the total number of calls to match(), not all of which increase
3148 the depth of recursion. Limiting the recursion depth limits the amount of
3149 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3150 when PCRE is compiled, and changed at run time. A patch from Google adds
3151 this functionality to the C++ interface.
3152
3153 18. Changes to the handling of Unicode character properties:
3154
3155 (a) Updated the table to Unicode 4.1.0.
3156
3157 (b) Recognize characters that are not in the table as "Cn" (undefined).
3158
3159 (c) I revised the way the table is implemented to a much improved format
3160 which includes recognition of ranges. It now supports the ranges that
3161 are defined in UnicodeData.txt, and it also amalgamates other
3162 characters into ranges. This has reduced the number of entries in the
3163 table from around 16,000 to around 3,000, thus reducing its size
3164 considerably. I realized I did not need to use a tree structure after
3165 all - a binary chop search is just as efficient. Having reduced the
3166 number of entries, I extended their size from 6 bytes to 8 bytes to
3167 allow for more data.
3168
3169 (d) Added support for Unicode script names via properties such as \p{Han}.
3170
3171 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3172 matching that character.
3173
3174 20. When matching a repeated Unicode property with a minimum greater than zero,
3175 (for example \pL{2,}), PCRE could look past the end of the subject if it
3176 reached it while seeking the minimum number of characters. This could
3177 happen only if some of the characters were more than one byte long, because
3178 there is a check for at least the minimum number of bytes.
3179
3180 21. Refactored the implementation of \p and \P so as to be more general, to
3181 allow for more different types of property in future. This has changed the
3182 compiled form incompatibly. Anybody with saved compiled patterns that use
3183 \p or \P will have to recompile them.
3184
3185 22. Added "Any" and "L&" to the supported property types.
3186
3187 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3188 but give a compile time error if the value is greater than 0xff.
3189
3190 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3191 accidentally not being installed or uninstalled.
3192
3193 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3194 made were to insert the current release number. This seemed silly, because
3195 it made things harder for people building PCRE on systems that don't run
3196 "configure". I have turned pcre.h into a distributed file, no longer built
3197 by "configure", with the version identification directly included. There is
3198 no longer a pcre.h.in file.
3199
3200 However, this change necessitated a change to the pcre-config script as
3201 well. It is built from pcre-config.in, and one of the substitutions was the
3202 release number. I have updated configure.ac so that ./configure now finds
3203 the release number by grepping pcre.h.
3204
3205 26. Added the ability to run the tests under valgrind.
3206
3207
3208 Version 6.4 05-Sep-05
3209 ---------------------
3210
3211 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3212 "--" to be printed when multiple files were scanned, even when none of the
3213 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3214 consider it to be a bug, and have restored the previous behaviour.
3215
3216 2. A couple of code tidies to get rid of compiler warnings.
3217
3218 3. The pcretest program used to cheat by referring to symbols in the library
3219 whose names begin with _pcre_. These are internal symbols that are not
3220 really supposed to be visible externally, and in some environments it is
3221 possible to suppress them. The cheating is now confined to including
3222 certain files from the library's source, which is a bit cleaner.
3223
3224 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3225 file's purpose clearer.
3226
3227 5. Reorganized pcre_ucp_findchar().
3228
3229
3230 Version 6.3 15-Aug-05
3231 ---------------------
3232
3233 1. The file libpcre.pc.in did not have general read permission in the tarball.
3234
3235 2. There were some problems when building without C++ support:
3236
3237 (a) If C++ support was not built, "make install" and "make test" still
3238 tried to test it.
3239
3240 (b) There were problems when the value of CXX was explicitly set. Some
3241 changes have been made to try to fix these, and ...
3242
3243 (c) --disable-cpp can now be used to explicitly disable C++ support.
3244
3245 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3246 backslash in a target when C++ was disabled. This confuses some
3247 versions of "make", apparently. Using an intermediate variable solves
3248 this. (Same for CPP_LOBJ.)
3249
3250 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3251 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3252 necessary on certain architectures.
3253
3254 4. Added a setting of -export-symbols-regex to the link command to remove
3255 those symbols that are exported in the C sense, but actually are local
3256 within the library, and not documented. Their names all begin with
3257 "_pcre_". This is not a perfect job, because (a) we have to except some
3258 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3259 available (and never for static libraries). I have made a note to try to
3260 find a way round (a) in the future.
3261
3262
3263 Version 6.2 01-Aug-05
3264 ---------------------
3265
3266 1. There was no test for integer overflow of quantifier values. A construction
3267 such as {1111111111111111} would give undefined results. What is worse, if
3268 a minimum quantifier for a parenthesized subpattern overflowed and became
3269 negative, the calculation of the memory size went wrong. This could have
3270 led to memory overwriting.
3271
3272 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3273
3274 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3275 operating environments where this matters.
3276
3277 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3278 PCRE options from within the C++ wrapper.
3279
3280 5. Named capturing subpatterns were not being correctly counted when a pattern
3281 was compiled. This caused two problems: (a) If there were more than 100
3282 such subpatterns, the calculation of the memory needed for the whole
3283 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3284 back references of the form \12, where the number was greater than 9, were
3285 not recognized as back references, even though there were sufficient
3286 previous subpatterns.
3287
3288 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3289 versions of gcc, e.g. 2.95.4.
3290
3291
3292 Version 6.1 21-Jun-05
3293 ---------------------
3294
3295 1. There was one reference to the variable "posix" in pcretest.c that was not
3296 surrounded by "#if !defined NOPOSIX".
3297
3298 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3299 the cross-check on the old pcre_info() function, for the benefit of the
3300 cut-down version of PCRE that is currently imported into Exim.
3301
3302 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3303 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3304 patterns that start (?i) but I don't think that matters. The use of (?i) is
3305 just an example; this all applies to the other options as well.
3306
3307 4. Since libtool seems to echo the compile commands it is issuing, the output
3308 from "make" can be reduced a bit by putting "@" in front of each libtool
3309 compile command.
3310
3311 5. Patch from the folks at Google for configure.in to be a bit more thorough
3312 in checking for a suitable C++ installation before trying to compile the
3313 C++ stuff. This should fix a reported problem when a compiler was present,
3314 but no suitable headers.
3315
3316 6. The man pages all had just "PCRE" as their title. I have changed them to
3317 be the relevant file name. I have also arranged that these names are
3318 retained in the file doc/pcre.txt, which is a concatenation in text format
3319 of all the man pages except the little individual ones for each function.
3320
3321 7. The NON-UNIX-USE file had not been updated for the different set of source
3322 files that come with release 6. I also added a few comments about the C++
3323 wrapper.
3324
3325
3326 Version 6.0 07-Jun-05
3327 ---------------------
3328
3329 1. Some minor internal re-organization to help with my DFA experiments.
3330
3331 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3332 didn't matter for the library itself when fully configured, but did matter
3333 when compiling without UCP support, or within Exim, where the ucp files are
3334 not imported.
3335
3336 3. Refactoring of the library code to split up the various functions into
3337 different source modules. The addition of the new DFA matching code (see
3338 below) to a single monolithic source would have made it really too
3339 unwieldy, quite apart from causing all the code to be include in a
3340 statically linked application, when only some functions are used. This is
3341 relevant even without the DFA addition now that patterns can be compiled in
3342 one application and matched in another.
3343
3344 The downside of splitting up is that there have to be some external
3345 functions and data tables that are used internally in different modules of
3346 the library but which are not part of the API. These have all had their
3347 names changed to start with "_pcre_" so that they are unlikely to clash
3348 with other external names.
3349
3350 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3351 a different (DFA) algorithm. Although it is slower than the original
3352 function, it does have some advantages for certain types of matching
3353 problem.
3354
3355 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3356 including restarting after a partial match.
3357
3358 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3359 defined when compiling for Windows was sent to me. I have put it into the
3360 code, though I have no means of testing or verifying it.
3361
3362 7. Added the pcre_refcount() auxiliary function.
3363
3364 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3365 match before or at the first newline in the subject string. In pcretest,
3366 the /f option on a pattern can be used to set this.
3367
3368 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3369 would behave wrongly. This has been present in PCRE since release 4.0.
3370
3371 10. A number of changes to the pcregrep command:
3372
3373 (a) Refactored how -x works; insert ^(...)$ instead of setting
3374 PCRE_ANCHORED and checking the length, in preparation for adding
3375 something similar for -w.
3376
3377 (b) Added the -w (match as a word) option.
3378
3379 (c) Refactored the way lines are read and buffered so as to have more
3380 than one at a time available.
3381
3382 (d) Implemented a pcregrep test script.
3383
3384 (e) Added the -M (multiline match) option. This allows patterns to match
3385 over several lines of the subject. The buffering ensures that at least
3386 8K, or the rest of the document (whichever is the shorter) is available
3387 for matching (and similarly the previous 8K for lookbehind assertions).
3388
3389 (f) Changed the --help output so that it now says
3390
3391 -w, --word-regex(p)
3392
3393 instead of two lines, one with "regex" and the other with "regexp"
3394 because that confused at least one person since the short forms are the
3395 same. (This required a bit of code, as the output is generated
3396 automatically from a table. It wasn't just a text change.)
3397
3398 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3399 option but starts with a hyphen. Could be a pattern or a path name
3400 starting with a hyphen, for instance.
3401
3402 (h) "-" can be given as a file name to represent stdin.
3403
3404 (i) When file names are being printed, "(standard input)" is used for
3405 the standard input, for compatibility with GNU grep. Previously
3406 "<stdin>" was used.
3407
3408 (j) The option --label=xxx can be used to supply a name to be used for
3409 stdin when file names are being printed. There is no short form.
3410
3411 (k) Re-factored the options decoding logic because we are going to add
3412 two more options that take data. Such options can now be given in four
3413 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3414
3415 (l) Added the -A, -B, and -C options for requesting that lines of context
3416 around matches be printed.
3417
3418 (m) Added the -L option to print the names of files that do not contain
3419 any matching lines, that is, the complement of -l.
3420
3421 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3422 continue to scan other files.
3423
3424 (o) The -s option was incorrectly implemented. For compatibility with other
3425 greps, it now suppresses the error message for a non-existent or non-
3426 accessible file (but not the return code). There is a new option called
3427 -q that suppresses the output of matching lines, which was what -s was
3428 previously doing.
3429
3430 (p) Added --include and --exclude options to specify files for inclusion
3431 and exclusion when recursing.
3432
3433 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3434 Hopefully, it now does.
3435
3436 12. Missing cast in pcre_study().
3437
3438 13. Added an "uninstall" target to the makefile.
3439
3440 14. Replaced "extern" in the function prototypes in Makefile.in with
3441 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3442 world, but is set differently for Windows.
3443
3444 15. Added a second compiling function called pcre_compile2(). The only
3445 difference is that it has an extra argument, which is a pointer to an
3446 integer error code. When there is a compile-time failure, this is set
3447 non-zero, in addition to the error test pointer being set to point to an
3448 error message. The new argument may be NULL if no error number is required
3449 (but then you may as well call pcre_compile(), which is now just a
3450 wrapper). This facility is provided because some applications need a
3451 numeric error indication, but it has also enabled me to tidy up the way
3452 compile-time errors are handled in the POSIX wrapper.
3453
3454 16. Added VPATH=.libs to the makefile; this should help when building with one
3455 prefix path and installing with another. (Or so I'm told by someone who
3456 knows more about this stuff than I do.)
3457
3458 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3459 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3460 match everything, including newlines. This is not POSIX-compatible, but
3461 somebody wanted the feature. From pcretest it can be activated by using
3462 both the P and the s flags.
3463
3464 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3465
3466 19. libpcre.pc was being incorrectly installed as executable.
3467
3468 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3469 it now also looks for '\r' so that it will work unmodified on Windows.
3470
3471 21. Added Google's contributed C++ wrapper to the distribution.
3472
3473 22. Added some untidy missing memory free() calls in pcretest, to keep
3474 Electric Fence happy when testing.
3475
3476
3477
3478 Version 5.0 13-Sep-04
3479 ---------------------
3480
3481 1. Internal change: literal characters are no longer packed up into items
3482 containing multiple characters in a single byte-string. Each character
3483 is now matched using a separate opcode. However, there may be more than one
3484 byte in the character in UTF-8 mode.
3485
3486 2. The pcre_callout_block structure has two new fields: pattern_position and
3487 next_item_length. These contain the offset in the pattern to the next match
3488 item, and its length, respectively.
3489
3490 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3491 insertion of callouts before each pattern item. Added the /C option to
3492 pcretest to make use of this.
3493
3494 4. On the advice of a Windows user, the lines
3495
3496 #if defined(_WIN32) || defined(WIN32)
3497 _setmode( _fileno( stdout ), 0x8000 );
3498 #endif /* defined(_WIN32) || defined(WIN32) */
3499
3500 have been added to the source of pcretest. This apparently does useful
3501 magic in relation to line terminators.
3502
3503 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3504 for the benefit of those environments where the "b" makes a difference.
3505
3506 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3507 to know about it. I have put a hack into configure.in that adds in code
3508 to set GCC=yes if CC=icc. This seems to end up at a point in the
3509 generated configure script that is early enough to affect the setting of
3510 compiler options, which is what is needed, but I have no means of testing
3511 whether it really works. (The user who reported this had patched the
3512 generated configure script, which of course I cannot do.)
3513
3514 LATER: After change 22 below (new libtool files), the configure script
3515 seems to know about icc (and also ecc). Therefore, I have commented out
3516 this hack in configure.in.
3517
3518 7. Added support for pkg-config (2 patches were sent in).
3519
3520 8. Negated POSIX character classes that used a combination of internal tables
3521 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3522 [[:^ascii]]. Typically, they would match almost any characters. The other
3523 POSIX classes were not broken in this way.
3524
3525 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3526 to find the match, as PCRE was deluded into thinking that the match had to
3527 start at the start point or following a newline. The same bug applied to
3528 patterns with negative forward assertions or any backward assertions
3529 preceding ".*" at the start, unless the pattern required a fixed first
3530 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3531
3532 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3533 starting at the last subject character, bytes beyond the end of the subject
3534 string were read.
3535
3536 11. Renamed the variable "class" as "classbits" to make life easier for C++
3537 users. (Previously there was a macro definition, but it apparently wasn't
3538 enough.)
3539
3540 12. Added the new field "tables" to the extra data so that tables can be passed
3541 in at exec time, or the internal tables can be re-selected. This allows
3542 a compiled regex to be saved and re-used at a later time by a different
3543 program that might have everything at different addresses.
3544
3545 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3546 -R library as well as a -L library.
3547
3548 14. The debugging options of pcretest (-d on the command line or D on a
3549 pattern) showed incorrect output for anything following an extended class
3550 that contained multibyte characters and which was followed by a quantifier.
3551
3552 15. Added optional support for general category Unicode character properties
3553 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3554 support. It adds about 90K to the size of the library. The meanings of the
3555 inbuilt class escapes such as \d and \s have NOT been changed.
3556
3557 16. Updated pcredemo.c to include calls to free() to release the memory for the
3558 compiled pattern.
3559
3560 17. The generated file chartables.c was being created in the source directory
3561 instead of in the building directory. This caused the build to fail if the
3562 source directory was different from the building directory, and was
3563 read-only.
3564
3565 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3566 file. No doubt somebody will tell me if they don't make sense... Also added
3567 Dan Mooney's comments about building on OpenVMS.
3568
3569 19. Added support for partial matching via the PCRE_PARTIAL option for
3570 pcre_exec() and the \P data escape in pcretest.
3571
3572 20. Extended pcretest with 3 new pattern features:
3573
3574 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3575 write the compiled pattern to the file whose name is "rest-of-line".
3576 This is a straight binary dump of the data, with the saved pointer to
3577 the character tables forced to be NULL. The study data, if any, is
3578 written too. After writing, pcretest reads a new pattern.
3579
3580 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3581 compiled pattern from the given file. There must not be any
3582 occurrences of "<" in the file name (pretty unlikely); if there are,
3583 pcretest will instead treat the initial "<" as a pattern delimiter.
3584 After reading in the pattern, pcretest goes on to read data lines as
3585 usual.
3586
3587 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3588 and 16-bit fields in a compiled pattern, to simulate a pattern that
3589 was compiled on a host of opposite endianness.
3590
3591 21. The pcre-exec() function can now cope with patterns that were compiled on
3592 hosts of opposite endianness, with this restriction:
3593
3594 As for any compiled expression that is saved and used later, the tables
3595 pointer field cannot be preserved; the extra_data field in the arguments
3596 to pcre_exec() should be used to pass in a tables address if a value
3597 other than the default internal tables were used at compile time.
3598
3599 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3600 now diagnosed as an error. Previously, most of the time, a negative number
3601 would have been treated as zero, but if in addition "ovector" was passed as
3602 NULL, a crash could occur.
3603
3604 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3605 new versions from the libtool 1.5 distribution (the last one is a copy of
3606 a file called libtool.m4). This seems to have fixed the need to patch
3607 "configure" to support Darwin 1.3 (which I used to do). However, I still
3608 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3609 workstation).
3610
3611 24. Changed the PCRE licence to be the more standard "BSD" licence.
3612
3613
3614 Version 4.5 01-Dec-03
3615 ---------------------
3616
3617 1. There has been some re-arrangement of the code for the match() function so
3618 that it can be compiled in a version that does not call itself recursively.
3619 Instead, it keeps those local variables that need separate instances for
3620 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3621 needs to "recurse". Keeping track of where control must go is done by means
3622 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3623 hide most of the details from the main code, and operates only if
3624 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3625 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3626 operating.
3627
3628 To make it easier for callers to provide specially tailored get/free
3629 functions for this usage, two new functions, pcre_stack_malloc, and
3630 pcre_stack_free, are used. They are always called in strict stacking order,
3631 and the size of block requested is always the same.
3632
3633 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3634 PCRE has been compiled to use the stack or the heap for recursion. The
3635 -C option of pcretest uses this to show which version is compiled.
3636
3637 A new data escape \S, is added to pcretest; it causes the amounts of store
3638 obtained and freed by both kinds of malloc/free at match time to be added
3639 to the output.
3640
3641 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3642 what's available on my current Linux desktop machine.
3643
3644 3. When matching a UTF-8 string, the test for a valid string at the start has
3645 been extended. If start_offset is not zero, PCRE now checks that it points
3646 to a byte that is the start of a UTF-8 character. If not, it returns
3647 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3648 this is necessary because there may be backward assertions in the pattern.
3649 When matching the same subject several times, it may save resources to use
3650 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3651
3652 4. The code for checking the validity of UTF-8 strings has been tightened so
3653 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3654 containing "overlong sequences".
3655
3656 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3657 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3658 should have been "&", but it just so happened that all the cases this let
3659 through by mistake were picked up later in the function.
3660
3661 6. I had used a variable called "isblank" - this is a C99 function, causing
3662 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3663
3664 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3665 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3666 so the version is shown; (c) stop "make test" echoing "./RunTest".
3667
3668 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3669
3670 9. The prototype for memmove() for systems that don't have it was using
3671 size_t, but the inclusion of the header that defines size_t was later. I've
3672 moved the #includes for the C headers earlier to avoid this.
3673
3674 10. Added some adjustments to the code to make it easier to compiler on certain
3675 special systems:
3676
3677 (a) Some "const" qualifiers were missing.
3678 (b) Added the macro EXPORT before all exported functions; by default this
3679 is defined to be empty.
3680 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3681 that it reads its output file name as an argument instead of writing
3682 to the standard output and assuming this can be redirected.
3683
3684 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3685 class containing characters with values greater than 255, PCRE compilation
3686 went into a loop.
3687
3688 12. A recursive reference to a subpattern that was within another subpattern
3689 that had a minimum quantifier of zero caused PCRE to crash. For example,
3690 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3691 recursion. If the recursively-called subpattern itself had a zero repeat,
3692 that was OK.
3693
3694 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3695 buffer into which it was copied (for escape processing) was still set at
3696 1024, so long lines caused crashes.
3697
3698 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3699 "internal error: code overflow...". This applied to any character class
3700 that was followed by a possessive quantifier.
3701
3702 15. Modified the Makefile to add libpcre.la as a prerequisite for
3703 libpcreposix.la because I was told this is needed for a parallel build to
3704 work.
3705
3706 16. If a pattern that contained .* following optional items at the start was
3707 studied, the wrong optimizing data was generated, leading to matching
3708 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3709 matching string must start with a or b or c. The correct conclusion for
3710 this pattern is that a match can start with any character.
3711
3712
3713 Version 4.4 13-Aug-03
3714 ---------------------
3715
3716 1. In UTF-8 mode, a character class containing characters with values between
3717 127 and 255 was not handled correctly if the compiled pattern was studied.
3718 In fixing this, I have also improved the studying algorithm for such
3719 classes (slightly).
3720
3721 2. Three internal functions had redundant arguments passed to them. Removal
3722 might give a very teeny performance improvement.
3723
3724 3. Documentation bug: the value of the capture_top field in a callout is *one
3725 more than* the number of the hightest numbered captured substring.
3726
3727 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3728 in incorrectly linking with a previously installed version. They now link
3729 explicitly with libpcre.la.
3730
3731 5. configure.in no longer needs to recognize Cygwin specially.
3732
3733 6. A problem in pcre.in for Windows platforms is fixed.
3734
3735 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3736 pcretest, it used to include the size of the study block as part of its
3737 output. Unfortunately, the structure contains a field that has a different
3738 size on different hardware architectures. This meant that the tests that
3739 showed this size failed. As the block is currently always of a fixed size,
3740 this information isn't actually particularly useful in pcretest output, so
3741 I have just removed it.
3742
3743 8. Three pre-processor statements accidentally did not start in column 1.
3744 Sadly, there are *still* compilers around that complain, even though
3745 standard C has not required this for well over a decade. Sigh.
3746
3747 9. In pcretest, the code for checking callouts passed small integers in the
3748 callout_data field, which is a void * field. However, some picky compilers
3749 complained about the casts involved for this on 64-bit systems. Now
3750 pcretest passes the address of the small integer instead, which should get
3751 rid of the warnings.
3752
3753 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3754 both compile and run time, and gives an error if an invalid UTF-8 sequence
3755 is found. There is a option for disabling this check in cases where the
3756 string is known to be correct and/or the maximum performance is wanted.
3757
3758 11. In response to a bug report, I changed one line in Makefile.in from
3759
3760 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3761 to
3762 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3763
3764 to look similar to other lines, but I have no way of telling whether this
3765 is the right thing to do, as I do not use Windows. No doubt I'll get told
3766 if it's wrong...
3767
3768
3769 Version 4.3 21-May-03
3770 ---------------------
3771
3772 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3773 Makefile.
3774
3775 2. Some refactoring to improve the quality of the code:
3776
3777 (i) The utf8_table... variables are now declared "const".
3778
3779 (ii) The code for \cx, which used the "case flipping" table to upper case
3780 lower case letters, now just substracts 32. This is ASCII-specific,
3781 but the whole concept of \cx is ASCII-specific, so it seems
3782 reasonable.
3783
3784 (iii) PCRE was using its character types table to recognize decimal and
3785 hexadecimal digits in the pattern. This is silly, because it handles
3786 only 0-9, a-f, and A-F, but the character types table is locale-
3787 specific, which means strange things might happen. A private
3788 table is now used for this - though it costs 256 bytes, a table is
3789 much faster than multiple explicit tests. Of course, the standard
3790 character types table is still used for matching digits in subject
3791 strings against \d.
3792
3793 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3794 ending in _t are). So I've renamed it as ESC_tee.
3795
3796 3. The first argument for regexec() in the POSIX wrapper should have been
3797 defined as "const".
3798
3799 4. Changed pcretest to use malloc() for its buffers so that they can be
3800 Electric Fenced for debugging.
3801
3802 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3803 to read one or more bytes before the start of the subject string. Often this
3804 had no effect on PCRE's behaviour, but in some circumstances it could
3805 provoke a segmentation fault.
3806
3807 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3808 to try to read one or more bytes before the start of the subject string.
3809
3810 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3811 UTF-8 support could misbehave in various ways if the subject string
3812 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3813 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3814 back over UTF-8 characters.)
3815
3816
3817 Version 4.2 14-Apr-03
3818 ---------------------
3819
3820 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3821
3822 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3823 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3824 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3825 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3826 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3827 and BUILD_EXEEXT
3828 Note: automatic setting of the BUILD variables is not yet working
3829 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3830 compile-time but not at link-time
3831 [LINK]: use for linking executables only
3832 make different versions for Windows and non-Windows
3833 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3834 libraries
3835 [LINK_FOR_BUILD]: new variable
3836 [OBJEXT]: use throughout
3837 [EXEEXT]: use throughout
3838 <winshared>: new target
3839 <wininstall>: new target
3840 <dftables.o>: use native compiler
3841 <dftables>: use native linker
3842 <install>: handle Windows platform correctly
3843 <clean>: ditto
3844 <check>: ditto
3845 copy DLL to top builddir before testing
3846
3847 As part of these changes, -no-undefined was removed again. This was reported
3848 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3849 in any case.
3850
3851 3. Some tidies to get rid of compiler warnings:
3852
3853 . In the match_data structure, match_limit was an unsigned long int, whereas
3854 match_call_count was an int. I've made them both unsigned long ints.
3855
3856 . In pcretest the fact that a const uschar * doesn't automatically cast to
3857 a void * provoked a warning.
3858
3859 . Turning on some more compiler warnings threw up some "shadow" variables
3860 and a few more missing casts.
3861
3862 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3863 option, a class that contained a single character with a value between 128
3864 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3865
3866 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3867 option, a class that contained several characters, but with at least one
3868 whose value was between 128 and 255 caused PCRE to crash.
3869
3870
3871 Version 4.1 12-Mar-03
3872 ---------------------
3873
3874 1. Compiling with gcc -pedantic found a couple of places where casts were
3875 needed, and a string in dftables.c that was longer than standard compilers are
3876 required to support.
3877
3878 2. Compiling with Sun's compiler found a few more places where the code could
3879 be tidied up in order to avoid warnings.
3880
3881 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3882 first of these names is deprecated in the latest Autoconf in favour of the name
3883 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3884 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3885 analogy I have changed it to CFLAGS_FOR_BUILD.
3886
3887 4. Added -no-undefined to the linking command in the Makefile, because this is
3888 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3889 linking step for the pcreposix library.
3890
3891 5. PCRE was failing to diagnose the case of two named groups with the same
3892 name.
3893
3894 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3895 literal character that is needed in the subject for a match, and scans along to
3896 ensure that it is present before embarking on the full matching process. This
3897 saves time in cases of nested unlimited repeats that are never going to match.
3898 Problem: the scan can take a lot of time if the subject is very long (e.g.
3899 megabytes), thus penalizing straightforward matches. It is now done only if the
3900 amount of subject to be scanned is less than 1000 bytes.
3901
3902 7. A lesser problem with the same optimization is that it was recording the
3903 first character of an anchored pattern as "needed", thus provoking a search
3904 right along the subject, even when the first match of the pattern was going to
3905 fail. The "needed" character is now not set for anchored patterns, unless it
3906 follows something in the pattern that is of non-fixed length. Thus, it still
3907 fulfils its original purpose of finding quick non-matches in cases of nested
3908 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3909
3910
3911 Version 4.0 17-Feb-03
3912 ---------------------
3913
3914 1. If a comment in an extended regex that started immediately after a meta-item
3915 extended to the end of string, PCRE compiled incorrect data. This could lead to
3916 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3917
3918 2. Moved to autoconf 2.53 and libtool 1.4.2.
3919
3920 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3921 the special perltest8 script is no longer needed - all the tests can be run
3922 from a single perltest script.
3923
3924 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3925 by \s. It has now been removed in PCRE. This means it isn't recognized as
3926 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3927 class [:space:] *does* include VT, thereby creating a mess.
3928
3929 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3930 space and tab.
3931
3932 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3933 its new features into the main test script, reducing the number of scripts.
3934
3935 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3936 were backward compatible, and made the (?i) apply to the whole pattern, as if
3937 /i were given. Now it behaves more logically, and applies the option setting
3938 only to what follows. PCRE has been changed to follow suit. However, if it
3939 finds options settings right at the start of the pattern, it extracts them into
3940 the global options, as before. Thus, they show up in the info data.
3941
3942 8. Added support for the \Q...\E escape sequence. Characters in between are
3943 treated as literals. This is slightly different from Perl in that $ and @ are
3944 also handled as literals inside the quotes. In Perl, they will cause variable
3945 interpolation. Note the following examples:
3946
3947 Pattern PCRE matches Perl matches
3948
3949 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3950 \Qabc\$xyz\E abc\$xyz abc\$xyz
3951 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3952
3953 For compatibility with Perl, \Q...\E sequences are recognized inside character
3954 classes as well as outside them.
3955
3956 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3957 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3958 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3959 signed/unsigned warnings.
3960
3961 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3962 option for pcretest, so I've replaced it by a simple function that does just
3963 that job.
3964
3965 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3966 "pcregrep -".
3967
3968 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3969 Java package. This provides some syntactic sugar for simple cases of what my
3970 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3971 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3972 item, you can use this simplified notation. Note that only makes sense with
3973 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3974 greediness, whatever the setting of the PCRE_UNGREEDY option.
3975
3976 13. A change of greediness default within a pattern was not taking effect at
3977 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3978 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3979 was abstracted outside.
3980
3981 14. PCRE now supports the \G assertion. It is true when the current matching
3982 position is at the start point of the match. This differs from \A when the
3983 starting offset is non-zero. Used with the /g option of pcretest (or similar
3984 code), it works in the same way as it does for Perl's /g option. If all
3985 alternatives of a regex begin with \G, the expression is anchored to the start
3986 match position, and the "anchored" flag is set in the compiled expression.
3987
3988 15. Some bugs concerning the handling of certain option changes within patterns
3989 have been fixed. These applied to options other than (?ims). For example,
3990 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3991 been the other way round. Some of this was related to change 7 above.
3992
3993 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3994 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3995 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3996 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3997
3998 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3999 mode. Unlike ".", it always matches newline, whatever the setting of
4000 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4001 assertions. Perl allows it, but it doesn't (in general) work because it can't
4002 calculate the length of the lookbehind. At least, that's the case for Perl
4003 5.8.0 - I've been told they are going to document that it doesn't work in
4004 future.
4005
4006 18. Added an error diagnosis for escapes that PCRE does not support: these are
4007 \L, \l, \N, \P, \p, \U, \u, and \X.
4008
4009 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4010 reading past the end of the pattern in cases such as /[abcd/.
4011
4012 20. PCRE was getting more memory than necessary for patterns with classes that
4013 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4014
4015 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4016 compiling PCRE for use with Virtual Pascal.
4017
4018 22. Small fix to the Makefile to make it work properly if the build is done
4019 outside the source tree.
4020
4021 23. Added a new extension: a condition to go with recursion. If a conditional
4022 subpattern starts with (?(R) the "true" branch is used if recursion has
4023 happened, whereas the "false" branch is used only at the top level.
4024
4025 24. When there was a very long string of literal characters (over 255 bytes
4026 without UTF support, over 250 bytes with UTF support), the computation of how
4027 much memory was required could be incorrect, leading to segfaults or other
4028 strange effects.
4029
4030 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4031 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4032 there was a subsequent back reference to those brackets. This meant that, for
4033 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4034 possible to check for precisely this case. All we can do is abandon the
4035 optimization if .* occurs inside capturing brackets when there are any back
4036 references whatsoever. (See below for a better fix that came later.)
4037
4038 26. The handling of the optimization for finding the first character of a
4039 non-anchored pattern, and for finding a character that is required later in the
4040 match were failing in some cases. This didn't break the matching; it just
4041 failed to optimize when it could. The way this is done has been re-implemented.
4042
4043 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4044
4045 28. Added a new feature that provides some of the functionality that Perl
4046 provides with (?{...}). The facility is termed a "callout". The way it is done
4047 in PCRE is for the caller to provide an optional function, by setting
4048 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4049 global variable. By default it is unset, which disables all calling out. To get
4050 the function called, the regex must include (?C) at appropriate points. This
4051 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4052 This provides a means of identifying different callout points. When PCRE
4053 reaches such a point in the regex, if pcre_callout has been set, the external
4054 function is called. It is provided with data in a structure called
4055 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4056 matching continues; if it returns a non-zero value, the match at the current
4057 point fails. However, backtracking will occur if possible. [This was changed
4058 later and other features added - see item 49 below.]
4059
4060 29. pcretest is upgraded to test the callout functionality. It provides a
4061 callout function that displays information. By default, it shows the start of
4062 the match and the current position in the text. There are some new data escapes
4063 to vary what happens:
4064
4065 \C+ in addition, show current contents of captured substrings
4066 \C- do not supply a callout function
4067 \C!n return 1 when callout number n is reached
4068 \C!n!m return 1 when callout number n is reached for the mth time
4069
4070 30. If pcregrep was called with the -l option and just a single file name, it
4071 output "<stdin>" if a match was found, instead of the file name.
4072
4073 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4074 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4075 pcre_exec(). This saves a malloc/free per call. The default value of
4076 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4077 when configuring.
4078
4079 32. The default maximum size of a compiled pattern is 64K. There have been a
4080 few cases of people hitting this limit. The code now uses macros to handle the
4081 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4082 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4083 configuring. Tests 2 and 5 work only with 2-byte links because they output
4084 debugging information about compiled patterns.
4085
4086 33. Internal code re-arrangements:
4087
4088 (a) Moved the debugging function for printing out a compiled regex into
4089 its own source file (printint.c) and used #include to pull it into
4090 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4091 separate copies.
4092
4093 (b) Defined the list of op-code names for debugging as a macro in
4094 internal.h so that it is next to the definition of the opcodes.
4095
4096 (c) Defined a table of op-code lengths for simpler skipping along compiled
4097 code. This is again a macro in internal.h so that it is next to the
4098 definition of the opcodes.
4099
4100 34. Added support for recursive calls to individual subpatterns, along the
4101 lines of Robin Houston's patch (but implemented somewhat differently).
4102
4103 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4104 allow it to read and process whole directories in Win32. This code was
4105 contributed by Lionel Fourquaux; it has not been tested by me.
4106
4107 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4108 used to name a group. Names consist of alphanumerics and underscores, and must
4109 be unique. Back references use the syntax (?P=name) and recursive calls use
4110 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4111 numbers. The function pcre_fullinfo() can be used after compilation to extract
4112 a name/number map. There are three relevant calls:
4113
4114 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4115 PCRE_INFO_NAMECOUNT yields the number of entries
4116 PCRE_INFO_NAMETABLE yields a pointer to the map.
4117
4118 The map is a vector of fixed-size entries. The size of each entry depends on
4119 the length of the longest name used. The first two bytes of each entry are the
4120 group number, most significant byte first. There follows the corresponding
4121 name, zero terminated. The names are in alphabetical order.
4122
4123 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4124 case instead of 255. Making it the same both with and without UTF-8 support
4125 means that the same test output works with both.
4126
4127 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4128 calling malloc() with a zero argument.
4129
4130 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4131 optimization. I've improved things by keeping a bitmap of backreferences with
4132 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4133 fact referenced, the optimization can be applied. It is unlikely that a
4134 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4135 the match to follow \n) will appear inside brackets with a number greater than
4136 31, but if it does, any back reference > 31 suppresses the optimization.
4137
4138 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4139 of disabling numbered capturing parentheses. Any opening parenthesis that is
4140 not followed by ? behaves as if it were followed by ?: but named parentheses
4141 can still be used for capturing (and they will acquire numbers in the usual
4142 way).
4143
4144 41. Redesigned the return codes from the match() function into yes/no/error so
4145 that errors can be passed back from deep inside the nested calls. A malloc
4146 failure while inside a recursive subpattern call now causes the
4147 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4148
4149 42. It is now possible to set a limit on the number of times the match()
4150 function is called in a call to pcre_exec(). This facility makes it possible to
4151 limit the amount of recursion and backtracking, though not in a directly
4152 obvious way, because the match() function is used in a number of different
4153 circumstances. The count starts from zero for each position in the subject
4154 string (for non-anchored patterns). The default limit is, for compatibility, a
4155 large number, namely 10 000 000. You can change this in two ways:
4156
4157 (a) When configuring PCRE before making, you can use --with-match-limit=n
4158 to set a default value for the compiled library.
4159
4160 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4161 a different value is set. See 45 below.
4162
4163 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4164
4165 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4166 of things that can be changed at compile time. The first argument specifies
4167 what is wanted and the second points to where the information is to be placed.
4168 The current list of available information is:
4169
4170 PCRE_CONFIG_UTF8
4171
4172 The output is an integer that is set to one if UTF-8 support is available;
4173 otherwise it is set to zero.
4174
4175 PCRE_CONFIG_NEWLINE
4176
4177 The output is an integer that it set to the value of the code that is used for
4178 newline. It is either LF (10) or CR (13).
4179
4180 PCRE_CONFIG_LINK_SIZE
4181
4182 The output is an integer that contains the number of bytes used for internal
4183 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4184
4185 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4186
4187 The output is an integer that contains the threshold above which the POSIX
4188 interface uses malloc() for output vectors. See item 31 above.
4189
4190 PCRE_CONFIG_MATCH_LIMIT
4191
4192 The output is an unsigned integer that contains the default limit of the number
4193 of match() calls in a pcre_exec() execution. See 42 above.
4194
4195 44. pcretest has been upgraded by the addition of the -C option. This causes it
4196 to extract all the available output from the new pcre_config() function, and to
4197 output it. The program then exits immediately.
4198
4199 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4200 order to support additional features. One way would have been to define
4201 pcre_exec2() (for example) with extra arguments, but this would not have been
4202 extensible, and would also have required all calls to the original function to
4203 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4204 is used for passing in "extra" data from pcre_study().
4205
4206 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4207 contains the following fields:
4208
4209 flags a bitmap indicating which of the following fields are set
4210 study_data opaque data from pcre_study()
4211 match_limit a way of specifying a limit on match() calls for a specific
4212 call to pcre_exec()
4213 callout_data data for callouts (see 49 below)
4214
4215 The flag bits are also defined in pcre.h, and are
4216
4217 PCRE_EXTRA_STUDY_DATA
4218 PCRE_EXTRA_MATCH_LIMIT
4219 PCRE_EXTRA_CALLOUT_DATA
4220
4221 The pcre_study() function now returns one of these new pcre_extra blocks, with
4222 the actual study data pointed to by the study_data field, and the
4223 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4224 before. That is, this change is entirely upwards-compatible and requires no
4225 change to existing code.
4226
4227 If you want to pass in additional data to pcre_exec(), you can either place it
4228 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4229 block.
4230
4231 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4232 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4233 times with different match limits, until it finds the minimum value needed for
4234 pcre_exec() to complete. The value is then output. This can be instructive; for
4235 most simple matches the number is quite small, but for pathological cases it
4236 gets very large very quickly.
4237
4238 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4239 returns the size of the data block pointed to by the study_data field in a
4240 pcre_extra block, that is, the value that was passed as the argument to
4241 pcre_malloc() when PCRE was getting memory in which to place the information
4242 created by pcre_study(). The fourth argument should point to a size_t variable.
4243 pcretest has been extended so that this information is shown after a successful
4244 pcre_study() call when information about the compiled regex is being displayed.
4245
4246 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4247 because what follows is always an absolute path. (Later: it turns out that this
4248 is more than cosmetic for MinGW, because it doesn't like empty path
4249 components.)
4250
4251 49. Some changes have been made to the callout feature (see 28 above):
4252
4253 (i) A callout function now has three choices for what it returns:
4254
4255 0 => success, carry on matching
4256 > 0 => failure at this point, but backtrack if possible
4257 < 0 => serious error, return this value from pcre_exec()
4258
4259 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4260 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4261 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4262 use by callout functions. It will never be used by PCRE itself.
4263
4264 (ii) The pcre_extra structure (see 45 above) has a void * field called
4265 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4266 pcre_callout_block structure has a field of the same name. The contents of
4267 the field passed in the pcre_extra structure are passed to the callout
4268 function in the corresponding field in the callout block. This makes it
4269 easier to use the same callout-containing regex from multiple threads. For
4270 testing, the pcretest program has a new data escape
4271
4272 \C*n pass the number n (may be negative) as callout_data
4273
4274 If the callout function in pcretest receives a non-zero value as
4275 callout_data, it returns that value.
4276
4277 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4278 there were some redundant $(CFLAGS) in commands that are now specified as
4279 $(LINK), which already includes $(CFLAGS).
4280
4281 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4282 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4283 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4284 one-byte characters throughout. Note that case-insensitive matching applies
4285 only to characters whose values are less than 256. PCRE doesn't support the
4286 notion of cases for higher-valued characters.
4287
4288 (i) A character class whose characters are all within 0-255 is handled as
4289 a bit map, and the map is inverted for negative classes. Previously, a
4290 character > 255 always failed to match such a class; however it should
4291 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4292
4293 (ii) A negated character class with a single character < 255 is coded as
4294 "not this character" (OP_NOT). This wasn't working properly when the test
4295 character was multibyte, either singly or repeated.
4296
4297 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4298 mode, for example: \x{100}{2,3}.
4299
4300 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4301 singly or repeated) now correctly test multibyte characters. However,
4302 PCRE doesn't recognize any characters with values greater than 255 as
4303 digits, spaces, or word characters. Such characters always match \D, \S,
4304 and \W, and never match \d, \s, or \w.
4305
4306 (v) Classes may now contain characters and character ranges with values
4307 greater than 255. For example: [ab\x{100}-\x{400}].
4308
4309 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4310 PCRE in UTF-8 mode.
4311
4312 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4313 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4314 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4315 value.)
4316
4317 53. The single man page has become too large. I have therefore split it up into
4318 a number of separate man pages. These also give rise to individual HTML pages;
4319 these are now put in a separate directory, and there is an index.html page that
4320 lists them all. Some hyperlinking between the pages has been installed.
4321
4322 54. Added convenience functions for handling named capturing parentheses.
4323
4324 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4325 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4326 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4327 are faulted.
4328
4329 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4330 calling configure. These values are used when compiling the dftables.c program
4331 which is run to generate the source of the default character tables. They
4332 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4333 you will need to set these values.
4334
4335 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4336
4337
4338 Version 3.9 02-Jan-02
4339 ---------------------
4340
4341 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4342
4343 2. If --disable-static was given, the building process failed when trying to
4344 build pcretest and pcregrep. (For some reason it was using libtool to compile
4345 them, which is not right, as they aren't part of the library.)
4346
4347
4348 Version 3.8 18-Dec-01
4349 ---------------------
4350
4351 1. The experimental UTF-8 code was completely screwed up. It was packing the
4352 bytes in the wrong order. How dumb can you get?
4353
4354
4355 Version 3.7 29-Oct-01
4356 ---------------------
4357
4358 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4359 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4360 this didn't happen under Solaris 8, where I normally test things.
4361
4362 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4363 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4364 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4365 this fix an hour or so after the initial 3.7 release.)
4366
4367
4368 Version 3.6 23-Oct-01
4369 ---------------------
4370
4371 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4372 offsets passed as NULL with zero offset count.
4373
4374 2. The config.guess and config.sub files had not been updated when I moved to
4375 the latest autoconf.
4376
4377
4378 Version 3.5 15-Aug-01
4379 ---------------------
4380
4381 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4382 had been forgotten.
4383
4384 2. By using declared but undefined structures, we can avoid using "void"
4385 definitions in pcre.h while keeping the internal definitions of the structures
4386 private.
4387
4388 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4389 user point of view, this means that both static and shared libraries are built
4390 by default, but this can be individually controlled. More of the work of
4391 handling this static/shared cases is now inside libtool instead of PCRE's make
4392 file.
4393
4394 4. The pcretest utility is now installed along with pcregrep because it is
4395 useful for users (to test regexs) and by doing this, it automatically gets
4396 relinked by libtool. The documentation has been turned into a man page, so
4397 there are now .1, .txt, and .html versions in /doc.
4398
4399 5. Upgrades to pcregrep:
4400 (i) Added long-form option names like gnu grep.
4401 (ii) Added --help to list all options with an explanatory phrase.
4402 (iii) Added -r, --recursive to recurse into sub-directories.
4403 (iv) Added -f, --file to read patterns from a file.
4404
4405 6. pcre_exec() was referring to its "code" argument before testing that
4406 argument for NULL (and giving an error if it was NULL).
4407
4408 7. Upgraded Makefile.in to allow for compiling in a different directory from
4409 the source directory.
4410
4411 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4412 options bits, the pointer it was passed was to an int instead of to an unsigned
4413 long int. This mattered only on 64-bit systems.
4414
4415 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4416 generated) instead of pcre.in, which it its source. Also made the same change
4417 in several of the .c files.
4418
4419 10. A new release of gcc defines printf() as a macro, which broke pcretest
4420 because it had an ifdef in the middle of a string argument for printf(). Fixed
4421 by using separate calls to printf().
4422
4423 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4424 script, to force use of CR or LF instead of \n in the source. On non-Unix
4425 systems, the value can be set in config.h.
4426
4427 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4428 absolute limit. Changed the text of the error message to make this clear, and
4429 likewise updated the man page.
4430
4431 13. The limit of 99 on the number of capturing subpatterns has been removed.
4432 The new limit is 65535, which I hope will not be a "real" limit.
4433
4434
4435 Version 3.4 22-Aug-00
4436 ---------------------
4437
4438 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4439
4440 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4441
4442
4443 Version 3.3 01-Aug-00
4444 ---------------------
4445
4446 1. If an octal character was given, but the value was greater than \377, it
4447 was not getting masked to the least significant bits, as documented. This could
4448 lead to crashes in some systems.
4449
4450 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4451 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4452
4453 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4454 These just pass their arguments on to (pcre_free)(), but they are provided
4455 because some uses of PCRE bind it to non-C systems that can call its functions,
4456 but cannot call free() or pcre_free() directly.
4457
4458 4. Add "make test" as a synonym for "make check". Corrected some comments in
4459 the Makefile.
4460
4461 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4462 Makefile.
4463
4464 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4465 command called pgrep for grepping around the active processes.
4466
4467 7. Added the beginnings of support for UTF-8 character strings.
4468
4469 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4470 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4471 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4472 out for the ar command.)
4473
4474
4475 Version 3.2 12-May-00
4476 ---------------------
4477
4478 This is purely a bug fixing release.
4479
4480 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4481 of ZA. This was just one example of several cases that could provoke this bug,
4482 which was introduced by change 9 of version 2.00. The code for breaking
4483 infinite loops after an iteration that matches an empty string was't working
4484 correctly.
4485
4486 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4487 when matched against abbab (for example). After matching an empty string, it
4488 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4489 caused it to match further down the string than it should.
4490
4491 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4492 was there because it doesn't seem to be needed, and it causes trouble on some
4493 systems, as it is not a Standard C header. It has been removed.
4494
4495 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4496 were reported on the Macintosh. The changes were from
4497
4498 while ((c = *(++ptr)) != 0 && c != '\n');
4499 to
4500 while ((c = *(++ptr)) != 0 && c != '\n') ;
4501
4502 Totally extraordinary, but if that's what it takes...
4503
4504 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4505 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4506 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4507 assumes the way PCRE uses memmove() (always moving upwards).
4508
4509 6. PCRE is being used in one environment where strchr() is not available. There
4510 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4511 faster code anyway.
4512
4513
4514 Version 3.1 09-Feb-00
4515 ---------------------
4516
4517 The only change in this release is the fixing of some bugs in Makefile.in for
4518 the "install" target:
4519
4520 (1) It was failing to install pcreposix.h.
4521
4522 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4523
4524
4525 Version 3.0 01-Feb-00
4526 ---------------------
4527
4528 1. Add support for the /+ modifier to perltest (to output $` like it does in
4529 pcretest).
4530
4531 2. Add support for the /g modifier to perltest.
4532
4533 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4534 matches null strings.
4535
4536 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4537 pattern. Perl treats empty patterns specially - it reuses the most recent
4538 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4539 effect.
4540
4541 5. The POSIX interface was broken in that it was just handing over the POSIX
4542 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4543 required a bigger vector, with some working space on the end. This means that
4544 the POSIX wrapper now has to get and free some memory, and copy the results.
4545
4546 6. Added some simple autoconf support, placing the test data and the
4547 documentation in separate directories, re-organizing some of the
4548 information files, and making it build pcre-config (a GNU standard). Also added
4549 libtool support for building PCRE as a shared library, which is now the
4550 default.
4551
4552 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4553 09 are not valid octal constants. Single digits will be used for minor values
4554 less than 10.
4555
4556 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4557 existing programs that set these in the POSIX interface can use PCRE without
4558 modification.
4559
4560 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4561 return all that pcre_info() returns, plus additional data. The pcre_info()
4562 function is retained for compatibility, but is considered to be obsolete.
4563
4564 10. Added experimental recursion feature (?R) to handle one common case that
4565 Perl 5.6 will be able to do with (?p{...}).
4566
4567 11. Added support for POSIX character classes like [:alpha:], which Perl is
4568 adopting.
4569
4570
4571 Version 2.08 31-Aug-99
4572 ----------------------
4573
4574 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4575 trying to match at the startoffset position, but instead was moving forward to
4576 the next newline as if a previous match had failed.
4577
4578 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4579 and could get into a loop if a null string was matched other than at the start
4580 of the subject.
4581
4582 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4583 be distinguished at compile time, and for completeness also added PCRE_DATE.
4584
4585 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4586 in GnuWin32 environments.
4587
4588
4589 Version 2.07 29-Jul-99
4590 ----------------------
4591
4592 1. The documentation is now supplied in plain text form and HTML as well as in
4593 the form of man page sources.
4594
4595 2. C++ compilers don't like assigning (void *) values to other pointer types.
4596 In particular this affects malloc(). Although there is no problem in Standard
4597 C, I've put in casts to keep C++ compilers happy.
4598
4599 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4600 should be (const char *).
4601
4602 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4603 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4604 However, I haven't made this a standard facility. The documentation doesn't
4605 mention it, and the Makefile doesn't support it.
4606
4607 5. The Makefile now contains an "install" target, with editable destinations at
4608 the top of the file. The pcretest program is not installed.
4609
4610 6. pgrep -V now gives the PCRE version number and date.
4611
4612 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4613 causing the entire string to be ignored, instead of just the last character.
4614
4615 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4616 non-matching string, it can take a very, very long time, even for strings of
4617 quite modest length, because of the nested recursion. PCRE now does better in
4618 some of these cases. It does this by remembering the last required literal
4619 character in the pattern, and pre-searching the subject to ensure it is present
4620 before running the real match. In other words, it applies a heuristic to detect
4621 some types of certain failure quickly, and in the above example, if presented
4622 with a string that has no trailing " it gives "no match" very quickly.
4623
4624 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4625 other alternatives are tried instead.
4626
4627
4628 Version 2.06 09-Jun-99
4629 ----------------------
4630
4631 1. Change pcretest's output for amount of store used to show just the code
4632 space, because the remainder (the data block) varies in size between 32-bit and
4633 64-bit systems.
4634
4635 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4636 start matching at. This allows lookbehinds to work when searching for multiple
4637 occurrences in a string.
4638
4639 3. Added additional options to pcretest for testing multiple occurrences:
4640
4641 /+ outputs the rest of the string that follows a match
4642 /g loops for multiple occurrences, using the new startoffset argument
4643 /G loops for multiple occurrences by passing an incremented pointer
4644
4645 4. PCRE wasn't doing the "first character" optimization for patterns starting
4646 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4647 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4648 the letter 'x'. On long subject strings, this gives a significant speed-up.
4649
4650
4651 Version 2.05 21-Apr-99
4652 ----------------------
4653
4654 1. Changed the type of magic_number from int to long int so that it works
4655 properly on 16-bit systems.
4656
4657 2. Fixed a bug which caused patterns starting with .* not to work correctly
4658 when the subject string contained newline characters. PCRE was assuming
4659 anchoring for such patterns in all cases, which is not correct because .* will
4660 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4661 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4662 must be retried after every newline in the subject.
4663
4664
4665 Version 2.04 18-Feb-99
4666 ----------------------
4667
4668 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4669 computation of the store needed to hold the pattern was incorrect (too large).
4670 If such patterns were nested a few deep, this could multiply and become a real
4671 problem.
4672
4673 2. Added /M option to pcretest to show the memory requirement of a specific
4674 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4675
4676 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4677 compiled in such a way that the backtracking after subsequent failure was
4678 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4679 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4680
4681
4682 Version 2.03 02-Feb-99
4683 ----------------------
4684
4685 1. Fixed typo and small mistake in man page.
4686
4687 2. Added 4th condition (GPL supersedes if conflict) and created separate
4688 LICENCE file containing the conditions.
4689
4690 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4691 Perl, that is the internal \ allows the delimiter to be included in the
4692 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4693 the final delimiter, add \ to the end of the pattern (to test the error).
4694
4695 4. Added the convenience functions for extracting substrings after a successful
4696 match. Updated pcretest to make it able to test these functions.
4697
4698
4699 Version 2.02 14-Jan-99
4700 ----------------------
4701
4702 1. Initialized the working variables associated with each extraction so that
4703 their saving and restoring doesn't refer to uninitialized store.
4704
4705 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4706 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4707 fix the problem.
4708
4709 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4710 calls, and wasn't printing the correct value for compiling calls. Increased the
4711 default value of LOOPREPEAT, and the number of significant figures in the
4712 times.
4713
4714 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4715
4716 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4717 a building problem on Windows NT with a FAT file system.
4718
4719
4720 Version 2.01 21-Oct-98
4721 ----------------------
4722
4723 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4724 to character tables built by pcre_maketables() in the current locale. If NULL
4725 is passed, the default tables are used.
4726
4727
4728 Version 2.00 24-Sep-98
4729 ----------------------
4730
4731 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4732 it any more.
4733
4734 2. Allow quantification of (?>) groups, and make it work correctly.
4735
4736 3. The first character computation wasn't working for (?>) groups.
4737
4738 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4739 end of the subject) and add 5.005's \z, which really does match only at the
4740 very end of the subject.
4741
4742 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4743
4744 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4745 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4746 localized options. All options to pcre_study() were also removed.
4747
4748 7. Add other new features from 5.005:
4749
4750 $(?<= positive lookbehind
4751 $(?<! negative lookbehind
4752 (?imsx-imsx) added the unsetting capability
4753 such a setting is global if at outer level; local otherwise
4754 (?imsx-imsx:) non-capturing groups with option setting
4755 (?(cond)re|re) conditional pattern matching
4756
4757 A backreference to itself in a repeated group matches the previous
4758 captured string.
4759
4760 8. General tidying up of studying (both automatic and via "study")
4761 consequential on the addition of new assertions.
4762
4763 9. As in 5.005, unlimited repeated groups that could match an empty substring
4764 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4765 runtime if any iteration does actually match an empty substring.
4766
4767 10. Include the RunTest script in the distribution.
4768
4769 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4770 discrepancies, some of which were old and were also with respect to 5.004. They
4771 have now been fixed.
4772
4773
4774 Version 1.09 28-Apr-98
4775 ----------------------
4776
4777 1. A negated single character class followed by a quantifier with a minimum
4778 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4779 program crashes, or just wrong answers. This did not apply to negated classes
4780 containing more than one character, or to minima other than one.
4781
4782
4783 Version 1.08 27-Mar-98
4784 ----------------------
4785
4786 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4787
4788 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4789 latter must appear before anything that relies on it in the pattern.
4790
4791
4792 Version 1.07 16-Feb-98
4793 ----------------------
4794
4795 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4796 repeat of a potentially empty string).
4797
4798
4799 Version 1.06 23-Jan-98
4800 ----------------------
4801
4802 1. Added Markus Oberhumer's little patches for C++.
4803
4804 2. Literal strings longer than 255 characters were broken.
4805
4806
4807 Version 1.05 23-Dec-97
4808 ----------------------
4809
4810 1. Negated character classes containing more than one character were failing if
4811 PCRE_CASELESS was set at run time.
4812
4813
4814 Version 1.04 19-Dec-97
4815 ----------------------
4816
4817 1. Corrected the man page, where some "const" qualifiers had been omitted.
4818
4819 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4820 input syntax.
4821
4822 3. Fixed memory leak which occurred when a regex with back references was
4823 matched with an offsets vector that wasn't big enough. The temporary memory
4824 that is used in this case wasn't being freed if the match failed.
4825
4826 4. Tidied pcretest to ensure it frees memory that it gets.
4827
4828 5. Temporary memory was being obtained in the case where the passed offsets
4829 vector was exactly big enough.
4830
4831 6. Corrected definition of offsetof() from change 5 below.
4832
4833 7. I had screwed up change 6 below and broken the rules for the use of
4834 setjmp(). Now fixed.
4835
4836
4837 Version 1.03 18-Dec-97
4838 ----------------------
4839
4840 1. A erroneous regex with a missing opening parenthesis was correctly
4841 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4842 on some systems.
4843
4844 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4845 it was reported that one broken compiler failed on the former because "code" is
4846 also an independent variable.
4847
4848 3. The erroneous regex a[]b caused an array overrun reference.
4849
4850 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4851 fail on data ending with that character. (It was going on too far, and checking
4852 the next character, typically a binary zero.) This was specific to the
4853 optimized code for single-character negative classes.
4854
4855 5. Added a contributed patch from the TIN world which does the following:
4856
4857 + Add an undef for memmove, in case the the system defines a macro for it.
4858
4859 + Add a definition of offsetof(), in case there isn't one. (I don't know
4860 the reason behind this - offsetof() is part of the ANSI standard - but
4861 it does no harm).
4862
4863 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4864 most of the places where whitespace preceded '#'. I have given up and
4865 allowed the remaining 2 cases to be at the margin.
4866
4867 + Rename some variables in pcre to eliminate shadowing. This seems very
4868 pedantic, but does no harm, of course.
4869
4870 6. Moved the call to setjmp() into its own function, to get rid of warnings
4871 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4872
4873 7. Constructs such as \d{8,} were compiling into the equivalent of
4874 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4875 outcome, but in this particular case used more store than had been allocated,
4876 which caused the bug to be discovered because it threw up an internal error.
4877
4878 8. The debugging code in both pcre and pcretest for outputting the compiled
4879 form of a regex was going wrong in the case of back references followed by
4880 curly-bracketed repeats.
4881
4882
4883 Version 1.02 12-Dec-97
4884 ----------------------
4885
4886 1. Typos in pcre.3 and comments in the source fixed.
4887
4888 2. Applied a contributed patch to get rid of places where it used to remove
4889 'const' from variables, and fixed some signed/unsigned and uninitialized
4890 variable warnings.
4891
4892 3. Added the "runtest" target to Makefile.
4893
4894 4. Set default compiler flag to -O2 rather than just -O.
4895
4896
4897 Version 1.01 19-Nov-97
4898 ----------------------
4899
4900 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4901 like /([ab]*)*/, that is, for classes with more than one character in them.
4902
4903 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4904 as /((?>a*))*/ (a PCRE_EXTRA facility).
4905
4906
4907 Version 1.00 18-Nov-97
4908 ----------------------
4909
4910 1. Added compile-time macros to support systems such as SunOS4 which don't have
4911 memmove() or strerror() but have other things that can be used instead.
4912
4913 2. Arranged that "make clean" removes the executables.
4914
4915
4916 Version 0.99 27-Oct-97
4917 ----------------------
4918
4919 1. Fixed bug in code for optimizing classes with only one character. It was
4920 initializing a 32-byte map regardless, which could cause it to run off the end
4921 of the memory it had got.
4922
4923 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4924
4925
4926 Version 0.98 22-Oct-97
4927 ----------------------
4928
4929 1. Fixed bug in code for handling temporary memory usage when there are more
4930 back references than supplied space in the ovector. This could cause segfaults.
4931
4932
4933 Version 0.97 21-Oct-97
4934 ----------------------
4935
4936 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4937
4938 2. Optimized negated single characters not to use a bit map.
4939
4940 3. Brought error texts together as macro definitions; clarified some of them;
4941 fixed one that was wrong - it said "range out of order" when it meant "invalid
4942 escape sequence".
4943
4944 4. Changed some char * arguments to const char *.
4945
4946 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4947
4948 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4949 pcretest.
4950
4951
4952 Version 0.96 16-Oct-97
4953 ----------------------
4954
4955 1. Added a simple "pgrep" utility to the distribution.
4956
4957 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4958 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4959 where "ddd" means "one or more decimal digits".
4960
4961 3. Fixed serious bug. If a pattern had a back reference, but the call to
4962 pcre_exec() didn't supply a large enough ovector to record the related
4963 identifying subpattern, the match always failed. PCRE now remembers the number
4964 of the largest back reference, and gets some temporary memory in which to save
4965 the offsets during matching if necessary, in order to ensure that
4966 backreferences always work.
4967
4968 4. Increased the compatibility with Perl in a number of ways:
4969
4970 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4971 to request this handling. The option can be set at compile or exec time.
4972
4973 (b) $ matches before a terminating newline by default; an option
4974 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4975 mode). The option can be set at compile or exec time.
4976
4977 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4978 the same as Perl's. If the decimal number it represents is less than 10
4979 or there aren't that many previous left capturing parentheses, an octal
4980 escape is read. Inside a character class, it's always an octal escape,
4981 even if it is a single digit.
4982
4983 (d) An escaped but undefined alphabetic character is taken as a literal,
4984 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4985 escapes.
4986
4987 (e) {0} is now permitted. (The previous item is removed from the compiled
4988 pattern).
4989
4990 5. Changed all the names of code files so that the basic parts are no longer
4991 than 10 characters, and abolished the teeny "globals.c" file.
4992
4993 6. Changed the handling of character classes; they are now done with a 32-byte
4994 bit map always.
4995
4996 7. Added the -d and /D options to pcretest to make it possible to look at the
4997 internals of compilation without having to recompile pcre.
4998
4999
5000 Version 0.95 23-Sep-97
5001 ----------------------
5002
5003 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5004 \x20 at the start of a run of normal characters. These were being treated as
5005 real characters, instead of the source characters being re-checked.
5006
5007
5008 Version 0.94 18-Sep-97
5009 ----------------------
5010
5011 1. The functions are now thread-safe, with the caveat that the global variables
5012 containing pointers to malloc() and free() or alternative functions are the
5013 same for all threads.
5014
5015 2. Get pcre_study() to generate a bitmap of initial characters for non-
5016 anchored patterns when this is possible, and use it if passed to pcre_exec().
5017
5018
5019 Version 0.93 15-Sep-97
5020 ----------------------
5021
5022 1. /(b)|(:+)/ was computing an incorrect first character.
5023
5024 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5025 but not actually doing anything yet.
5026
5027 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5028 as Perl does (e.g. [-az] or [az-]).
5029
5030 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5031 all possible positions.
5032
5033 5. Split up into different modules to avoid including unneeded functions in a
5034 compiled binary. However, compile and exec are still in one module. The "study"
5035 function is split off.
5036
5037 6. The character tables are now in a separate module whose source is generated
5038 by an auxiliary program - but can then be edited by hand if required. There are
5039 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5040 toupper() in the code.
5041
5042 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5043 make them global. Abolish the function for setting them, as the caller can now
5044 set them directly.
5045
5046
5047 Version 0.92 11-Sep-97
5048 ----------------------
5049
5050 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5051 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5052
5053 2. Caseless matching was not working in character classes if the characters in
5054 the pattern were in upper case.
5055
5056 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5057
5058 4. Make PCRE_ANCHORED public and accept as a compile option.
5059
5060 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5061 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5062 pass them.
5063
5064 6. Give an error if bad option bits passed at compile or run time.
5065
5066 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5067 pcretest to cause it to pass that flag.
5068
5069 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5070 options, and the first character, if set.
5071
5072 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5073
5074
5075 Version 0.91 10-Sep-97
5076 ----------------------
5077
5078 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5079 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5080
5081 2. PCRE was looping on encountering an indefinitely repeated back reference to
5082 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5083 Perl does - treats the match as successful.
5084
5085 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5