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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1328 - (show annotations)
Sun May 12 16:28:22 2013 UTC (6 years, 5 months ago) by ph10
File size: 247294 byte(s)
Error occurred while calculating annotation data.
Documentation clarification for 16/32 bit libraries.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.33 28-April-2013
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 28. (*SKIP) is now supported by the JIT compiler.
108
109 29. (*THEN) is now supported by the JIT compiler.
110
111 30. Update RunTest with additional test selector options.
112
113 31. The way PCRE handles backtracking verbs has been changed in two ways.
114
115 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
116 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
117 backtracking. In some cases this makes it more Perl-compatible, but Perl's
118 rather obscure rules do not always do the same thing.
119
120 (2) Previously, backtracking verbs were confined within assertions. This is
121 no longer the case for positive assertions, except for (*ACCEPT). Again,
122 this sometimes improves Perl compatibility, and sometimes does not.
123
124 32. A number of tests that were in test 2 because Perl did things differently
125 have been moved to test 1, because either Perl or PCRE has changed, and
126 these tests are now compatible.
127
128 32. Control verbs are handled in the same way in JIT and interpreter.
129
130 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
131 contained a forward subroutine reference caused a compile error.
132
133 34. Auto-detect and optimize limited repetitions in JIT.
134
135 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
136 blocking (*UTF) etc.
137
138 36. In the interpreter, maximizing pattern repetitions for characters and
139 character types now use tail recursion, which reduces stack usage.
140
141 37. The value of the max lookbehind was not correctly preserved if a compiled
142 and saved regex was reloaded on a host of different endianness.
143
144 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
145 of the compiled pattern block, expand the flags field from 16 to 32 bits
146 because it was almost full.
147
148 39. Try madvise first before posix_madvise.
149
150 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
151 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
152 original change locked out all matches of empty strings. This has been
153 changed so that one match of an empty string per line is recognized.
154 Subsequent searches on the same line (for colouring or for --only-matching,
155 for example) do not recognize empty strings.
156
157 41. Applied a user patch to fix a number of spelling mistakes in comments.
158
159 42. Data lines longer than 65536 caused pcretest to crash.
160
161 44. Clarified the data type for length and startoffset arguments for pcre_exec
162 and pcre_dfa_exec in the function-specific man pages, where they were
163 explicitly stated to be in bytes, never having been updated. I also added
164 some clarification to the pcreapi man page.
165
166
167 Version 8.32 30-November-2012
168 -----------------------------
169
170 1. Improved JIT compiler optimizations for first character search and single
171 character iterators.
172
173 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
174 Patch by Daniel Richard G.
175
176 3. Single character iterator optimizations in the JIT compiler.
177
178 4. Improved JIT compiler optimizations for character ranges.
179
180 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
181 Reported by Giuseppe D'Angelo.
182
183 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
184 of a line, was being set incorrectly in cases where .* appeared inside
185 atomic brackets at the start of a pattern, or where there was a subsequent
186 *PRUNE or *SKIP.
187
188 7. Improved instruction cache flush for POWER/PowerPC.
189 Patch by Daniel Richard G.
190
191 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
192 grep:
193
194 (a) There is now no limit to the number of patterns to be matched.
195
196 (b) An error is given if a pattern is too long.
197
198 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
199 are now supported.
200
201 (d) --exclude-from and --include-from (multiple use) have been added.
202
203 (e) Exclusions and inclusions now apply to all files and directories, not
204 just to those obtained from scanning a directory recursively.
205
206 (f) Multiple uses of -f and --file-list are now supported.
207
208 (g) In a Windows environment, the default for -d has been changed from
209 "read" (the GNU grep default) to "skip", because otherwise the presence
210 of a directory in the file list provokes an error.
211
212 (h) The documentation has been revised and clarified in places.
213
214 9. Improve the matching speed of capturing brackets.
215
216 10. Changed the meaning of \X so that it now matches a Unicode extended
217 grapheme cluster.
218
219 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
220 out POSIX threads when JIT support is configured.
221
222 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
223
224 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
225 structure could go wrong in environments where size_t is not the same size
226 as int.
227
228 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
229 set.
230
231 15. The EBCDIC support had decayed; later updates to the code had included
232 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
233 general tidy up of EBCDIC-related issues, and the documentation was also
234 not quite right. There is now a test that can be run on ASCII systems to
235 check some of the EBCDIC-related things (but is it not a full test).
236
237 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
238 in a small tidy to the code.
239
240 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
241
242 18. If the --only-matching (-o) option in pcregrep is specified multiple
243 times, each one causes appropriate output. For example, -o1 -o2 outputs the
244 substrings matched by the 1st and 2nd capturing parentheses. A separating
245 string can be specified by --om-separator (default empty).
246
247 19. Improving the first n character searches.
248
249 20. Turn case lists for horizontal and vertical white space into macros so that
250 they are defined only once.
251
252 21. This set of changes together give more compatible Unicode case-folding
253 behaviour for characters that have more than one other case when UCP
254 support is available.
255
256 (a) The Unicode property table now has offsets into a new table of sets of
257 three or more characters that are case-equivalent. The MultiStage2.py
258 script that generates these tables (the pcre_ucd.c file) now scans
259 CaseFolding.txt instead of UnicodeData.txt for character case
260 information.
261
262 (b) The code for adding characters or ranges of characters to a character
263 class has been abstracted into a generalized function that also handles
264 case-independence. In UTF-mode with UCP support, this uses the new data
265 to handle characters with more than one other case.
266
267 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
268 whose other case is greater than 256 are now correctly matched
269 caselessly. Previously, the high codepoint matched the low one, but not
270 vice versa.
271
272 (d) The processing of \h, \H, \v, and \ in character classes now makes use
273 of the new class addition function, using character lists defined as
274 macros alongside the case definitions of 20 above.
275
276 (e) Caseless back references now work with characters that have more than
277 one other case.
278
279 (f) General caseless matching of characters with more than one other case
280 is supported.
281
282 22. Unicode character properties were updated from Unicode 6.2.0
283
284 23. Improved CMake support under Windows. Patch by Daniel Richard G.
285
286 24. Add support for 32-bit character strings, and UTF-32
287
288 25. Major JIT compiler update (code refactoring and bugfixing).
289 Experimental Sparc 32 support is added.
290
291 26. Applied a modified version of Daniel Richard G's patch to create
292 pcre.h.generic and config.h.generic by "make" instead of in the
293 PrepareRelease script.
294
295 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
296 pcre_compile.c when checking for a zero character.
297
298 28. Introducing a native interface for JIT. Through this interface, the compiled
299 machine code can be directly executed. The purpose of this interface is to
300 provide fast pattern matching, so several sanity checks are not performed.
301 However, feature tests are still performed. The new interface provides
302 1.4x speedup compared to the old one.
303
304 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
305 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
306 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
307
308 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
309 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
310 the "old" RFC 2279). Instead, it ended up passing a negative length to
311 pcre_exec().
312
313 31. Add support for GCC's visibility feature to hide internal functions.
314
315 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
316 "unknown -C option" after outputting 0 or 1.
317
318 33. There is now support for generating a code coverage report for the test
319 suite in environments where gcc is the compiler and lcov is installed. This
320 is mainly for the benefit of the developers.
321
322 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
323 unaddressable using valgrind annotations, allowing valgrind to detect
324 invalid memory accesses. This is mainly for the benefit of the developers.
325
326 25. (*UTF) can now be used to start a pattern in any of the three libraries.
327
328 26. Give configure error if --enable-cpp but no C++ compiler found.
329
330
331 Version 8.31 06-July-2012
332 -------------------------
333
334 1. Fixing a wrong JIT test case and some compiler warnings.
335
336 2. Removed a bashism from the RunTest script.
337
338 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
339 to unsigned type, result still unsigned" that was given by an MS compiler
340 on encountering the code "-sizeof(xxx)".
341
342 4. Partial matching support is added to the JIT compiler.
343
344 5. Fixed several bugs concerned with partial matching of items that consist
345 of more than one character:
346
347 (a) /^(..)\1/ did not partially match "aba" because checking references was
348 done on an "all or nothing" basis. This also applied to repeated
349 references.
350
351 (b) \R did not give a hard partial match if \r was found at the end of the
352 subject.
353
354 (c) \X did not give a hard partial match after matching one or more
355 characters at the end of the subject.
356
357 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
358 a partial match for the string "\r".
359
360 (e) When newline was set to CRLF, the metacharacter "." did not recognize
361 a partial match for a CR character at the end of the subject string.
362
363 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
364 running pcretest, the text "(JIT)" added to the output whenever JIT is
365 actually used to run the match.
366
367 7. Individual JIT compile options can be set in pcretest by following -s+[+]
368 or /S+[+] with a digit between 1 and 7.
369
370 8. OP_NOT now supports any UTF character not just single-byte ones.
371
372 9. (*MARK) control verb is now supported by the JIT compiler.
373
374 10. The command "./RunTest list" lists the available tests without actually
375 running any of them. (Because I keep forgetting what they all are.)
376
377 11. Add PCRE_INFO_MAXLOOKBEHIND.
378
379 12. Applied a (slightly modified) user-supplied patch that improves performance
380 when the heap is used for recursion (compiled with --disable-stack-for-
381 recursion). Instead of malloc and free for each heap frame each time a
382 logical recursion happens, frames are retained on a chain and re-used where
383 possible. This sometimes gives as much as 30% improvement.
384
385 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
386 call.
387
388 14. As documented, (*COMMIT) is now confined to within a positive assertion.
389
390 15. It is now possible to link pcretest with libedit as an alternative to
391 libreadline.
392
393 16. (*COMMIT) control verb is now supported by the JIT compiler.
394
395 17. The Unicode data tables have been updated to Unicode 6.1.0.
396
397 18. Added --file-list option to pcregrep.
398
399 19. Added binary file support to pcregrep, including the -a, --binary-files,
400 -I, and --text options.
401
402 20. The madvise function is renamed for posix_madvise for QNX compatibility
403 reasons. Fixed by Giuseppe D'Angelo.
404
405 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
406
407 22. Changed the option for creating symbolic links for 16-bit man pages from
408 -s to -sf so that re-installing does not cause issues.
409
410 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
411
412 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
413 matches in certain environments (the workspace was not being correctly
414 retained). Also added to pcre_dfa_exec() a simple plausibility check on
415 some of the workspace data at the beginning of a restart.
416
417 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
418 was not doing so when it should - probably a typo introduced by SVN 528
419 (change 8.10/14).
420
421 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
422 \w+ when the character tables indicated that \x{c4} was a word character.
423 There were several related cases, all because the tests for doing a table
424 lookup were testing for characters less than 127 instead of 255.
425
426 27. If a pattern contains capturing parentheses that are not used in a match,
427 their slots in the ovector are set to -1. For those that are higher than
428 any matched groups, this happens at the end of processing. In the case when
429 there were back references that the ovector was too small to contain
430 (causing temporary malloc'd memory to be used during matching), and the
431 highest capturing number was not used, memory off the end of the ovector
432 was incorrectly being set to -1. (It was using the size of the temporary
433 memory instead of the true size.)
434
435 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
436 ovector size, it uses memory at the end of the block that it has got.
437
438 29. Check for an overlong MARK name and give an error at compile time. The
439 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
440
441 30. JIT compiler update.
442
443 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
444 Rill for the patch.
445
446 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
447
448 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
449
450 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
451 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
452 was enabled.
453
454 35. Improve JIT code generation for greedy plus quantifier.
455
456 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
457 1 to "aa" instead of to an empty string. The bug affected repeated groups
458 that could potentially match an empty string.
459
460 37. Optimizing single character iterators in JIT.
461
462 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
463 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
464 codepoints that are too big for the mode are faulted, and in a UTF mode,
465 disallowed codepoints are also faulted.
466
467 39. If PCRE was compiled with UTF support, in three places in the DFA
468 matcher there was code that should only have been obeyed in UTF mode, but
469 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
470 processing when bytes with values greater than 127 were present. In 16-bit
471 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
472 both cases the values are those that cannot be the first data item in a UTF
473 character. The three items that might have provoked this were recursions,
474 possessively repeated groups, and atomic groups.
475
476 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
477 and pcregrep, because some OS require shared objects to be explicitly
478 passed to ld, causing the link step to fail if they are not.
479
480 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
481 mode, patterns that started with \h* or \R* might be incorrectly matched.
482
483
484 Version 8.30 04-February-2012
485 -----------------------------
486
487 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
488 name is defined in ctype.h.
489
490 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
491 only in quite long subpatterns.
492
493 3. Removed the function pcre_info(), which has been obsolete and deprecated
494 since it was replaced by pcre_fullinfo() in February 2000.
495
496 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
497 match a (*MARK), and the match failed at the start of the subject, a
498 reference to memory before the start of the subject could occur. This bug
499 was introduced by fix 17 of release 8.21.
500
501 5. A reference to an unset group with zero minimum repetition was giving
502 totally wrong answers (in non-JavaScript-compatibility mode). For example,
503 /(another)?(\1?)test/ matched against "hello world test". This bug was
504 introduced in release 8.13.
505
506 6. Add support for 16-bit character strings (a large amount of work involving
507 many changes and refactorings).
508
509 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
510 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
511 from a file.
512
513 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
514 rounding is not applied in this particular case).
515
516 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
517 if they appear, or are escaped, in patterns.
518
519 10. Get rid of a number of -Wunused-but-set-variable warnings.
520
521 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
522 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
523 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
524 also returns the mark "x". This bug applied to capturing parentheses,
525 non-capturing parentheses, and atomic parentheses. It also applied to some
526 assertions.
527
528 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
529 information out of configure.ac instead of relying on pcre.h.generic, which
530 is not stored in the repository.
531
532 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
533 -lreadline.
534
535 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
536
537 15. Applied Graycode's patch to put the top-level frame on the stack rather
538 than the heap when not using the stack for recursion. This gives a
539 performance improvement in many cases when recursion is not deep.
540
541 16. Experimental code added to "pcretest -C" to output the stack frame size.
542
543
544 Version 8.21 12-Dec-2011
545 ------------------------
546
547 1. Updating the JIT compiler.
548
549 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
550 are added as well.
551
552 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
553 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
554 calling _pcre_jit_exec. Some extra comments are added.
555
556 4. (*MARK) settings inside atomic groups that do not contain any capturing
557 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
558 was introduced by change 18 for 8.20.
559
560 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
561 ECMA-262 standard.
562
563 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
564 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
565 This bug was probably introduced by change 9 of 8.13.
566
567 7. While fixing 6 above, I noticed that a number of other items were being
568 incorrectly rejected as "not fixed length". This arose partly because newer
569 opcodes had not been added to the fixed-length checking code. I have (a)
570 corrected the bug and added tests for these items, and (b) arranged for an
571 error to occur if an unknown opcode is encountered while checking for fixed
572 length instead of just assuming "not fixed length". The items that were
573 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
574 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
575 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
576
577 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
578 being incorrectly compiled and would have given unpredicatble results.
579
580 9. A possessively repeated subpattern with minimum repeat count greater than
581 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
582 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
583 the first (A) could occur when it should not.
584
585 10. Add a cast and remove a redundant test from the code.
586
587 11. JIT should use pcre_malloc/pcre_free for allocation.
588
589 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
590 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
591 is anything other than /usr, -L is still shown).
592
593 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
594
595 14. Perl does not support \N without a following name in a [] class; PCRE now
596 also gives an error.
597
598 15. If a forward reference was repeated with an upper limit of around 2000,
599 it caused the error "internal error: overran compiling workspace". The
600 maximum number of forward references (including repeats) was limited by the
601 internal workspace, and dependent on the LINK_SIZE. The code has been
602 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
603 the default depends on LINK_SIZE. There is a new upper limit (for safety)
604 of around 200,000 forward references. While doing this, I also speeded up
605 the filling in of repeated forward references.
606
607 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
608 incorrectly expecting the subject to contain another "a" after the start.
609
610 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
611 in the match, the SKIP should be ignored. This was not happening; instead
612 the SKIP was being treated as NOMATCH. For patterns such as
613 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
614 tested.
615
616 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
617 now much more compatible with Perl, in particular in cases where the result
618 is a non-match for a non-anchored pattern. For example, if
619 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
620 "m", where previously it did not return a name. A side effect of this
621 change is that for partial matches, the last encountered mark name is
622 returned, as for non matches. A number of tests that were previously not
623 Perl-compatible have been moved into the Perl-compatible test files. The
624 refactoring has had the pleasing side effect of removing one argument from
625 the match() function, thus reducing its stack requirements.
626
627 19. If the /S+ option was used in pcretest to study a pattern using JIT,
628 subsequent uses of /S (without +) incorrectly behaved like /S+.
629
630 21. Retrieve executable code size support for the JIT compiler and fixing
631 some warnings.
632
633 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
634 not work when the shorter character appeared right at the end of the
635 subject string.
636
637 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
638 systems.
639
640 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
641 output it when the /M option is used in pcretest.
642
643 25. The CheckMan script was not being included in the distribution. Also, added
644 an explicit "perl" to run Perl scripts from the PrepareRelease script
645 because this is reportedly needed in Windows.
646
647 26. If study data was being save in a file and studying had not found a set of
648 "starts with" bytes for the pattern, the data written to the file (though
649 never used) was taken from uninitialized memory and so caused valgrind to
650 complain.
651
652 27. Updated RunTest.bat as provided by Sheri Pierce.
653
654 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
655
656 29. Computation of memory usage for the table of capturing group names was
657 giving an unnecessarily large value.
658
659
660 Version 8.20 21-Oct-2011
661 ------------------------
662
663 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
664 a POSIX class. After further experiments with Perl, which convinced me that
665 Perl has bugs and confusions, a closing square bracket is no longer allowed
666 in a POSIX name. This bug also affected patterns with classes that started
667 with full stops.
668
669 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
670 captured substring, but while checking the failing first alternative,
671 substring 1 is temporarily captured. If the output vector supplied to
672 pcre_exec() was not big enough for this capture, the yield of the function
673 was still zero ("insufficient space for captured substrings"). This cannot
674 be totally fixed without adding another stack variable, which seems a lot
675 of expense for a edge case. However, I have improved the situation in cases
676 such as /(a)(b)x|abc/ matched against "abc", where the return code
677 indicates that fewer than the maximum number of slots in the ovector have
678 been set.
679
680 3. Related to (2) above: when there are more back references in a pattern than
681 slots in the output vector, pcre_exec() uses temporary memory during
682 matching, and copies in the captures as far as possible afterwards. It was
683 using the entire output vector, but this conflicts with the specification
684 that only 2/3 is used for passing back captured substrings. Now it uses
685 only the first 2/3, for compatibility. This is, of course, another edge
686 case.
687
688 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
689 main code base, and can be used by building with --enable-jit. When this is
690 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
691 runtime --no-jit option is given.
692
693 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
694 ovector, the return from the function was zero, implying that there were
695 other matches that did not fit. The correct "exactly full" value is now
696 returned.
697
698 6. If a subpattern that was called recursively or as a subroutine contained
699 (*PRUNE) or any other control that caused it to give a non-standard return,
700 invalid errors such as "Error -26 (nested recursion at the same subject
701 position)" or even infinite loops could occur.
702
703 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
704 computing the minimum length on reaching *ACCEPT, and so ended up with the
705 wrong value of 1 rather than 0. Further investigation indicates that
706 computing a minimum subject length in the presence of *ACCEPT is difficult
707 (think back references, subroutine calls), and so I have changed the code
708 so that no minimum is registered for a pattern that contains *ACCEPT.
709
710 8. If (*THEN) was present in the first (true) branch of a conditional group,
711 it was not handled as intended. [But see 16 below.]
712
713 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
714 Sheri Pierce.
715
716 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
717 the first byte in a match must be "a".
718
719 11. Change 17 for 8.13 increased the recursion depth for patterns like
720 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
721 pattern contains any instances of (*THEN). If it does not, the old
722 optimizations are restored. It would be nice to do this on a per-group
723 basis, but at the moment that is not feasible.
724
725 12. In some environments, the output of pcretest -C is CRLF terminated. This
726 broke RunTest's code that checks for the link size. A single white space
727 character after the value is now allowed for.
728
729 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
730 For "fr", it uses the Windows-specific input and output files.
731
732 14. If (*THEN) appeared in a group that was called recursively or as a
733 subroutine, it did not work as intended. [But see next item.]
734
735 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
736 pattern fragments (but not containing any | characters). If A and B are
737 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
738 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
739 D. In other words, Perl considers parentheses that do not contain any |
740 characters to be part of a surrounding alternative, whereas PCRE was
741 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
742 differently. PCRE now behaves in the same way as Perl, except in the case
743 of subroutine/recursion calls such as (?1) which have in any case always
744 been different (but PCRE had them first :-).
745
746 16. Related to 15 above: Perl does not treat the | in a conditional group as
747 creating alternatives. Such a group is treated in the same way as an
748 ordinary group without any | characters when processing (*THEN). PCRE has
749 been changed to match Perl's behaviour.
750
751 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
752 RunGrepTest script failed.
753
754 18. Change 22 for version 13 caused atomic groups to use more stack. This is
755 inevitable for groups that contain captures, but it can lead to a lot of
756 stack use in large patterns. The old behaviour has been restored for atomic
757 groups that do not contain any capturing parentheses.
758
759 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
760 suppress the check for a minimum subject length at run time. (If it was
761 given to pcre_exec() or pcre_dfa_exec() it did work.)
762
763 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
764 fail to work when decoding hex characters in data strings in EBCDIC
765 environments.
766
767 21. It appears that in at least one Mac OS environment, the isxdigit() function
768 is implemented as a macro that evaluates to its argument more than once,
769 contravening the C 90 Standard (I haven't checked a later standard). There
770 was an instance in pcretest which caused it to go wrong when processing
771 \x{...} escapes in subject strings. The has been rewritten to avoid using
772 things like p++ in the argument of isxdigit().
773
774
775 Version 8.13 16-Aug-2011
776 ------------------------
777
778 1. The Unicode data tables have been updated to Unicode 6.0.0.
779
780 2. Two minor typos in pcre_internal.h have been fixed.
781
782 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
783 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
784 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
785
786 4. There were a number of related bugs in the code for matching backrefences
787 caselessly in UTF-8 mode when codes for the characters concerned were
788 different numbers of bytes. For example, U+023A and U+2C65 are an upper
789 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
790 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
791 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
792 2-byte code at the end of the subject (it thought there wasn't enough data
793 left).
794
795 5. Comprehensive information about what went wrong is now returned by
796 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
797 as the output vector has at least 2 elements. The offset of the start of
798 the failing character and a reason code are placed in the vector.
799
800 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
801 now returned is for the first byte of the failing character, instead of the
802 last byte inspected. This is an incompatible change, but I hope it is small
803 enough not to be a problem. It makes the returned offset consistent with
804 pcre_exec() and pcre_dfa_exec().
805
806 7. pcretest now gives a text phrase as well as the error number when
807 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
808 failure, the offset and reason code are output.
809
810 8. When \R was used with a maximizing quantifier it failed to skip backwards
811 over a \r\n pair if the subsequent match failed. Instead, it just skipped
812 back over a single character (\n). This seems wrong (because it treated the
813 two characters as a single entity when going forwards), conflicts with the
814 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
815 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
816 has been changed.
817
818 9. Some internal refactoring has changed the processing so that the handling
819 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
820 time (the PCRE_DOTALL option was changed this way some time ago: version
821 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
822 which was always a bit of a fudge. It also means that there is one less
823 argument for the match() function, which reduces its stack requirements
824 slightly. This change also fixes an incompatibility with Perl: the pattern
825 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
826
827 10. More internal refactoring has drastically reduced the number of recursive
828 calls to match() for possessively repeated groups such as (abc)++ when
829 using pcre_exec().
830
831 11. While implementing 10, a number of bugs in the handling of groups were
832 discovered and fixed:
833
834 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
835 (a|)*(?1) gave a compile-time internal error.
836 ((a|)+)+ did not notice that the outer group could match an empty string.
837 (^a|^)+ was not marked as anchored.
838 (.*a|.*)+ was not marked as matching at start or after a newline.
839
840 12. Yet more internal refactoring has removed another argument from the match()
841 function. Special calls to this function are now indicated by setting a
842 value in a variable in the "match data" data block.
843
844 13. Be more explicit in pcre_study() instead of relying on "default" for
845 opcodes that mean there is no starting character; this means that when new
846 ones are added and accidentally left out of pcre_study(), testing should
847 pick them up.
848
849 14. The -s option of pcretest has been documented for ages as being an old
850 synonym of -m (show memory usage). I have changed it to mean "force study
851 for every regex", that is, assume /S for every regex. This is similar to -i
852 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
853 using it. It makes it easier to run collections of tests with and without
854 study enabled, and thereby test pcre_study() more easily. All the standard
855 tests are now run with and without -s (but some patterns can be marked as
856 "never study" - see 20 below).
857
858 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
859 restoration of the capturing data to the outer values was not happening
860 correctly.
861
862 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
863 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
864 pattern had matched an empty string, and so incorrectly returned a no
865 match.
866
867 17. There was optimizing code for the last branch of non-capturing parentheses,
868 and also for the obeyed branch of a conditional subexpression, which used
869 tail recursion to cut down on stack usage. Unfortunately, now that there is
870 the possibility of (*THEN) occurring in these branches, tail recursion is
871 no longer possible because the return has to be checked for (*THEN). These
872 two optimizations have therefore been removed. [But see 8.20/11 above.]
873
874 18. If a pattern containing \R was studied, it was assumed that \R always
875 matched two bytes, thus causing the minimum subject length to be
876 incorrectly computed because \R can also match just one byte.
877
878 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
879 was incorrectly computed.
880
881 20. If /S is present twice on a test pattern in pcretest input, it now
882 *disables* studying, thereby overriding the use of -s on the command line
883 (see 14 above). This is necessary for one or two tests to keep the output
884 identical in both cases.
885
886 21. When (*ACCEPT) was used in an assertion that matched an empty string and
887 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
888
889 22. When an atomic group that contained a capturing parenthesis was
890 successfully matched, but the branch in which it appeared failed, the
891 capturing was not being forgotten if a higher numbered group was later
892 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
893 group 1 to "a", when in fact it should be unset. This applied to multi-
894 branched capturing and non-capturing groups, repeated or not, and also to
895 positive assertions (capturing in negative assertions does not happen
896 in PCRE) and also to nested atomic groups.
897
898 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
899 subject after a captured substring, to make it easier to tell which of a
900 number of identical substrings has been captured.
901
902 24. The way atomic groups are processed by pcre_exec() has been changed so that
903 if they are repeated, backtracking one repetition now resets captured
904 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
905 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
906 "aaa". Previously, it would have been "a". As part of this code
907 refactoring, the way recursive calls are handled has also been changed.
908
909 25. If an assertion condition captured any substrings, they were not passed
910 back unless some other capturing happened later. For example, if
911 (?(?=(a))a) was matched against "a", no capturing was returned.
912
913 26. When studying a pattern that contained subroutine calls or assertions,
914 the code for finding the minimum length of a possible match was handling
915 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
916 group 1 called group 2 while simultaneously a separate group 2 called group
917 1). A stack overflow occurred in this case. I have fixed this by limiting
918 the recursion depth to 10.
919
920 27. Updated RunTest.bat in the distribution to the version supplied by Tom
921 Fortmann. This supports explicit test numbers on the command line, and has
922 argument validation and error reporting.
923
924 28. An instance of \X with an unlimited repeat could fail if at any point the
925 first character it looked at was a mark character.
926
927 29. Some minor code refactoring concerning Unicode properties and scripts
928 should reduce the stack requirement of match() slightly.
929
930 30. Added the '=' option to pcretest to check the setting of unused capturing
931 slots at the end of the pattern, which are documented as being -1, but are
932 not included in the return count.
933
934 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
935 compiled something random. Now it gives a compile-time error (as does
936 Perl).
937
938 32. A *MARK encountered during the processing of a positive assertion is now
939 recorded and passed back (compatible with Perl).
940
941 33. If --only-matching or --colour was set on a pcregrep call whose pattern
942 had alternative anchored branches, the search for a second match in a line
943 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
944 matched the line "0102" twice. The same bug affected patterns that started
945 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
946 twice.
947
948 34. Previously, PCRE did not allow quantification of assertions. However, Perl
949 does, and because of capturing effects, quantifying parenthesized
950 assertions may at times be useful. Quantifiers are now allowed for
951 parenthesized assertions.
952
953 35. A minor code tidy in pcre_compile() when checking options for \R usage.
954
955 36. \g was being checked for fancy things in a character class, when it should
956 just be a literal "g".
957
958 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
959 appearance of a nested POSIX class supersedes an apparent external class.
960 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
961 unescaped square brackets may also appear as part of class names. For
962 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
963 more like Perl. (But see 8.20/1 above.)
964
965 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
966 was because it thought it was \N{name}, which is not supported).
967
968 39. Add minix to OS list not supporting the -S option in pcretest.
969
970 40. PCRE tries to detect cases of infinite recursion at compile time, but it
971 cannot analyze patterns in sufficient detail to catch mutual recursions
972 such as ((?1))((?2)). There is now a runtime test that gives an error if a
973 subgroup is called recursively as a subpattern for a second time at the
974 same position in the subject string. In previous releases this might have
975 been caught by the recursion limit, or it might have run out of stack.
976
977 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
978 happen only once. PCRE was, however incorrectly giving a compile time error
979 "recursive call could loop indefinitely" because it cannot analyze the
980 pattern in sufficient detail. The compile time test no longer happens when
981 PCRE is compiling a conditional subpattern, but actual runaway loops are
982 now caught at runtime (see 40 above).
983
984 42. It seems that Perl allows any characters other than a closing parenthesis
985 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
986 has been changed to be the same.
987
988 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
989 as not to get warnings when autogen.sh is called. Also changed
990 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
991
992 44. To help people who use pcregrep to scan files containing exceedingly long
993 lines, the following changes have been made:
994
995 (a) The default value of the buffer size parameter has been increased from
996 8K to 20K. (The actual buffer used is three times this size.)
997
998 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
999 PCRE is built.
1000
1001 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1002 to be set at run time.
1003
1004 (d) Numerical values in pcregrep options can be followed by K or M, for
1005 example --buffer-size=50K.
1006
1007 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1008 given and the return code is set to 2.
1009
1010 45. Add a pointer to the latest mark to the callout data block.
1011
1012 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1013 partial match of an empty string instead of no match. This was specific to
1014 the use of ".".
1015
1016 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1017 complete match instead of a partial match. This bug was dependent on both
1018 the PCRE_UTF8 and PCRE_DOTALL options being set.
1019
1020 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1021 starting byte set, because \b was not being ignored.
1022
1023
1024 Version 8.12 15-Jan-2011
1025 ------------------------
1026
1027 1. Fixed some typos in the markup of the man pages, and wrote a script that
1028 checks for such things as part of the documentation building process.
1029
1030 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1031 --match-limit and --recursion-limit options (added for 8.11). In
1032 particular, this made one of the standard tests fail. (The integer value
1033 went into the wrong half of a long int.)
1034
1035 3. If the --colour option was given to pcregrep with -v (invert match), it
1036 did strange things, either producing crazy output, or crashing. It should,
1037 of course, ignore a request for colour when reporting lines that do not
1038 match.
1039
1040 4. Another pcregrep bug caused similar problems if --colour was specified with
1041 -M (multiline) and the pattern match finished with a line ending.
1042
1043 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1044 matched in multiline mode, the following line was shown as part of the
1045 match. This seems wrong, so I have changed it.
1046
1047 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1048 the check for further matches in the same line (so they could be coloured)
1049 to overrun the end of the current line. If another match was found, it was
1050 incorrectly shown (and then shown again when found in the next line).
1051
1052 7. If pcregrep was compiled under Windows, there was a reference to the
1053 function pcregrep_exit() before it was defined. I am assuming this was
1054 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1055 reported by a user. I've moved the definition above the reference.
1056
1057
1058 Version 8.11 10-Dec-2010
1059 ------------------------
1060
1061 1. (*THEN) was not working properly if there were untried alternatives prior
1062 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1063 backtracked to try for "b" instead of moving to the next alternative branch
1064 at the same level (in this case, to look for "c"). The Perl documentation
1065 is clear that when (*THEN) is backtracked onto, it goes to the "next
1066 alternative in the innermost enclosing group".
1067
1068 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1069 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1070 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1071 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1072 (*THEN).
1073
1074 3. If \s appeared in a character class, it removed the VT character from
1075 the class, even if it had been included by some previous item, for example
1076 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1077 of \s, but is part of the POSIX "space" class.)
1078
1079 4. A partial match never returns an empty string (because you can always
1080 match an empty string at the end of the subject); however the checking for
1081 an empty string was starting at the "start of match" point. This has been
1082 changed to the "earliest inspected character" point, because the returned
1083 data for a partial match starts at this character. This means that, for
1084 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1085 (previously it gave "no match").
1086
1087 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1088 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1089 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1090 has an implication that the given string is incomplete (because a partial
1091 match is preferred over a full match). For this reason, these items now
1092 give a partial match in this situation. [Aside: previously, the one case
1093 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1094 match rather than a full match, which was wrong by the old rules, but is
1095 now correct.]
1096
1097 6. There was a bug in the handling of #-introduced comments, recognized when
1098 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1099 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1100 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1101 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1102 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1103 places in pcre_compile().
1104
1105 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1106 comments when looking ahead for named forward references to subpatterns,
1107 the only newline sequence it recognized was NL. It now handles newlines
1108 according to the set newline convention.
1109
1110 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1111 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1112 cater for a lack of strerror(). These oversights have been fixed.
1113
1114 9. Added --match-limit and --recursion-limit to pcregrep.
1115
1116 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1117
1118 11. When the -o option was used, pcregrep was setting a return code of 1, even
1119 when matches were found, and --line-buffered was not being honoured.
1120
1121 12. Added an optional parentheses number to the -o and --only-matching options
1122 of pcregrep.
1123
1124 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1125 can match an empty string. The code to do it in pcretest and pcredemo
1126 needed fixing:
1127
1128 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1129 only one byte after an empty string match just before CRLF (this case
1130 just got forgotten; "any" and "anycrlf" were OK).
1131
1132 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1133 mode when an empty string match preceded an ASCII character followed by
1134 a non-ASCII character. (The code for advancing by one character rather
1135 than one byte was nonsense.)
1136
1137 (c) The pcredemo.c sample program did not have any code at all to handle
1138 the cases when CRLF is a valid newline sequence.
1139
1140 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1141 as a starting offset was within the subject string. There is now a new
1142 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1143 negative or greater than the length of the string. In order to test this,
1144 pcretest is extended to allow the setting of negative starting offsets.
1145
1146 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1147 starting offset points to the beginning of a UTF-8 character was
1148 unnecessarily clumsy. I tidied it up.
1149
1150 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1151 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1152
1153 17. Nobody had reported that the --include_dir option, which was added in
1154 release 7.7 should have been called --include-dir (hyphen, not underscore)
1155 for compatibility with GNU grep. I have changed it to --include-dir, but
1156 left --include_dir as an undocumented synonym, and the same for
1157 --exclude-dir, though that is not available in GNU grep, at least as of
1158 release 2.5.4.
1159
1160 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1161 characters from a string of bytes) have been redefined so as not to use
1162 loops, in order to improve performance in some environments. At the same
1163 time, I abstracted some of the common code into auxiliary macros to save
1164 repetition (this should not affect the compiled code).
1165
1166 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1167 compile-time error is now given if \c is not followed by an ASCII
1168 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1169 different, and any byte value is allowed.)
1170
1171 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1172 START_OPTIMIZE option, which is now allowed at compile time - but just
1173 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1174 to pcregrep and other applications that have no direct access to PCRE
1175 options. The new /Y option in pcretest sets this option when calling
1176 pcre_compile().
1177
1178 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1179 back references. Groups containing recursive back references were forced to
1180 be atomic by that change, but in the case of named groups, the amount of
1181 memory required was incorrectly computed, leading to "Failed: internal
1182 error: code overflow". This has been fixed.
1183
1184 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1185 pcretest.c, to avoid build problems in some Borland environments.
1186
1187
1188 Version 8.10 25-Jun-2010
1189 ------------------------
1190
1191 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1192 THEN.
1193
1194 2. (*ACCEPT) was not working when inside an atomic group.
1195
1196 3. Inside a character class, \B is treated as a literal by default, but
1197 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1198 causes the error). The code is unchanged, but I tidied the documentation.
1199
1200 4. Inside a character class, PCRE always treated \R and \X as literals,
1201 whereas Perl faults them if its -w option is set. I have changed PCRE so
1202 that it faults them when PCRE_EXTRA is set.
1203
1204 5. Added support for \N, which always matches any character other than
1205 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1206
1207 6. When compiling pcregrep with newer versions of gcc which may have
1208 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1209 declared with attribute warn_unused_result" were given. Just casting the
1210 result to (void) does not stop the warnings; a more elaborate fudge is
1211 needed. I've used a macro to implement this.
1212
1213 7. Minor change to pcretest.c to avoid a compiler warning.
1214
1215 8. Added four artifical Unicode properties to help with an option to make
1216 \s etc use properties (see next item). The new properties are: Xan
1217 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1218
1219 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1220 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1221 this option. Modified pcretest to add /W to test this facility. Added
1222 REG_UCP to make it available via the POSIX interface.
1223
1224 10. Added --line-buffered to pcregrep.
1225
1226 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1227 studied, and the match started with a letter with a code point greater than
1228 127 whose first byte was different to the first byte of the other case of
1229 the letter, the other case of this starting letter was not recognized
1230 (#976).
1231
1232 12. If a pattern that was studied started with a repeated Unicode property
1233 test, for example, \p{Nd}+, there was the theoretical possibility of
1234 setting up an incorrect bitmap of starting bytes, but fortunately it could
1235 not have actually happened in practice until change 8 above was made (it
1236 added property types that matched character-matching opcodes).
1237
1238 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1239 possible starting bytes for non-anchored patterns.
1240
1241 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1242 \R, and also a number of cases that involve Unicode properties, both
1243 explicit and implicit when PCRE_UCP is set.
1244
1245 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1246 input, it could crash or give wrong results if characters with values
1247 greater than 0xc0 were present in the subject string. (Detail: it assumed
1248 UTF-8 input when processing these items.)
1249
1250 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1251 size_t is 64-bit (#991).
1252
1253 17. Added a check for running out of memory when PCRE is compiled with
1254 --disable-stack-for-recursion (#990).
1255
1256 18. If the last data line in a file for pcretest does not have a newline on
1257 the end, a newline was missing in the output.
1258
1259 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1260 less than 128) in its various bitmaps. However, there is a facility for
1261 generating tables according to the current locale when PCRE is compiled. It
1262 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1263 characters, are recognized by isspace() and therefore were getting set in
1264 these tables, and indeed these tables seem to approximate to ISO 8859. This
1265 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1266 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1267 which of course cannot start UTF-8 characters. I have changed the code so
1268 that only real ASCII characters (less than 128) and the correct starting
1269 bytes for UTF-8 encodings are set for characters greater than 127 when in
1270 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1271 altogether.)
1272
1273 20. Added the /T option to pcretest so as to be able to run tests with non-
1274 standard character tables, thus making it possible to include the tests
1275 used for 19 above in the standard set of tests.
1276
1277 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1278 reference to a subpattern the other side of a comment that contains an
1279 opening parenthesis caused either an internal compiling error, or a
1280 reference to the wrong subpattern.
1281
1282
1283 Version 8.02 19-Mar-2010
1284 ------------------------
1285
1286 1. The Unicode data tables have been updated to Unicode 5.2.0.
1287
1288 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1289 configured.
1290
1291 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1292 original author of that file, following a query about its status.
1293
1294 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1295 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1296
1297 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1298 quantifier applied to a forward-referencing subroutine call, could compile
1299 incorrect code or give the error "internal error: previously-checked
1300 referenced subpattern not found".
1301
1302 6. Both MS Visual Studio and Symbian OS have problems with initializing
1303 variables to point to external functions. For these systems, therefore,
1304 pcre_malloc etc. are now initialized to local functions that call the
1305 relevant global functions.
1306
1307 7. There were two entries missing in the vectors called coptable and poptable
1308 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1309 I've fixed the data, and added a kludgy way of testing at compile time that
1310 the lengths are correct (equal to the number of opcodes).
1311
1312 8. Following on from 7, I added a similar kludge to check the length of the
1313 eint vector in pcreposix.c.
1314
1315 9. Error texts for pcre_compile() are held as one long string to avoid too
1316 much relocation at load time. To find a text, the string is searched,
1317 counting zeros. There was no check for running off the end of the string,
1318 which could happen if a new error number was added without updating the
1319 string.
1320
1321 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1322
1323 11. \K was not working if it appeared in an atomic group or in a group that
1324 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1325 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1326 the assertion is positive, but not if it is negative.
1327
1328 12. Change 11 fortuitously reduced the size of the stack frame used in the
1329 "match()" function of pcre_exec.c by one pointer. Forthcoming
1330 implementation of support for (*MARK) will need an extra pointer on the
1331 stack; I have reserved it now, so that the stack frame size does not
1332 decrease.
1333
1334 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1335 item in branch that calls a recursion is a subroutine call - as in the
1336 second branch in the above example - was incorrectly given the compile-
1337 time error "recursive call could loop indefinitely" because pcre_compile()
1338 was not correctly checking the subroutine for matching a non-empty string.
1339
1340 14. The checks for overrunning compiling workspace could trigger after an
1341 overrun had occurred. This is a "should never occur" error, but it can be
1342 triggered by pathological patterns such as hundreds of nested parentheses.
1343 The checks now trigger 100 bytes before the end of the workspace.
1344
1345 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1346
1347
1348 Version 8.01 19-Jan-2010
1349 ------------------------
1350
1351 1. If a pattern contained a conditional subpattern with only one branch (in
1352 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1353 computed the wrong minimum data length (which is of course zero for such
1354 subpatterns). This could cause incorrect "no match" results.
1355
1356 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1357 the pattern is reset in the first branch, pcre_compile() failed with
1358 "internal error: code overflow at offset...". This happened only when
1359 the reset was to the original external option setting. (An optimization
1360 abstracts leading options settings into an external setting, which was the
1361 cause of this.)
1362
1363 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1364 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1365 assertion pattern did not match (meaning that the assertion was true), it
1366 was incorrectly treated as false if the SKIP had been reached during the
1367 matching. This also applied to assertions used as conditions.
1368
1369 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1370 assertion subpattern, including such a pattern used as a condition,
1371 unpredictable results occurred, instead of the error return
1372 PCRE_ERROR_DFA_UITEM.
1373
1374 5. The C++ GlobalReplace function was not working like Perl for the special
1375 situation when an empty string is matched. It now does the fancy magic
1376 stuff that is necessary.
1377
1378 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1379 removed. (These were left over from very, very early versions of PCRE.)
1380
1381 7. Some cosmetic changes to the code to make life easier when compiling it
1382 as part of something else:
1383
1384 (a) Change DEBUG to PCRE_DEBUG.
1385
1386 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1387 called "current" as "current_branch", to prevent a collision with the
1388 Linux macro when compiled as a kernel module.
1389
1390 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1391 prevent a collision with the Linux macro when compiled as a kernel
1392 module.
1393
1394 8. In pcre_compile() there are some checks for integer overflows that used to
1395 cast potentially large values to (double). This has been changed to that
1396 when building, a check for int64_t is made, and if it is found, it is used
1397 instead, thus avoiding the use of floating point arithmetic. (There is no
1398 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1399 double.
1400
1401 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1402 2005 (difference between two addresses compared to an unsigned value).
1403
1404 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1405 custom one, because of the following reported problem in Windows:
1406
1407 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1408 under Win32.
1409 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1410 therefore missing the function definition.
1411 - The compiler thus generates a "C" signature for the test function.
1412 - The linker fails to find the "C" function.
1413 - PCRE fails to configure if asked to do so against libbz2.
1414
1415 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1416 messages were output:
1417
1418 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1419 rerunning libtoolize, to keep the correct libtool macros in-tree.
1420 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1421
1422 I have done both of these things.
1423
1424 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1425 most of the time, it *can* run out if it is given a pattern that contains a
1426 runaway infinite recursion. I updated the discussion in the pcrestack man
1427 page.
1428
1429 13. Now that we have gone to the x.xx style of version numbers, the minor
1430 version may start with zero. Using 08 or 09 is a bad idea because users
1431 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1432 interpreted as invalid octal numbers. I've updated the previous comment in
1433 configure.ac, and also added a check that gives an error if 08 or 09 are
1434 used.
1435
1436 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1437 causing partial matching to fail when the end of the subject matched \W
1438 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1439
1440 15. There were some discrepancies between the declarations in pcre_internal.h
1441 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1442 their definitions. The declarations used "const uschar *" and the
1443 definitions used USPTR. Even though USPTR is normally defined as "const
1444 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1445 reported that: "This difference in casting confuses some C++ compilers, for
1446 example, SunCC recognizes above declarations as different functions and
1447 generates broken code for hbpcre." I have changed the declarations to use
1448 USPTR.
1449
1450 16. GNU libtool is named differently on some systems. The autogen.sh script now
1451 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1452 (FreeBSD).
1453
1454 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1455 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1456 comment: "Figure out how to create a longlong from a string: strtoll and
1457 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1458 instance, but it only takes 2 args instead of 3!"
1459
1460 18. A subtle bug concerned with back references has been fixed by a change of
1461 specification, with a corresponding code fix. A pattern such as
1462 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1463 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1464 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1465 same bug. Such groups have to be quantified to be useful, or contained
1466 inside another quantified group. (If there's no repetition, the reference
1467 can never match.) The problem arises because, having left the group and
1468 moved on to the rest of the pattern, a later failure that backtracks into
1469 the group uses the captured value from the final iteration of the group
1470 rather than the correct earlier one. I have fixed this in PCRE by forcing
1471 any group that contains a reference to itself to be an atomic group; that
1472 is, there cannot be any backtracking into it once it has completed. This is
1473 similar to recursive and subroutine calls.
1474
1475
1476 Version 8.00 19-Oct-09
1477 ----------------------
1478
1479 1. The table for translating pcre_compile() error codes into POSIX error codes
1480 was out-of-date, and there was no check on the pcre_compile() error code
1481 being within the table. This could lead to an OK return being given in
1482 error.
1483
1484 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1485 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1486 in a Windows environment.
1487
1488 3. The pcregrep --count option prints the count for each file even when it is
1489 zero, as does GNU grep. However, pcregrep was also printing all files when
1490 --files-with-matches was added. Now, when both options are given, it prints
1491 counts only for those files that have at least one match. (GNU grep just
1492 prints the file name in this circumstance, but including the count seems
1493 more useful - otherwise, why use --count?) Also ensured that the
1494 combination -clh just lists non-zero counts, with no names.
1495
1496 4. The long form of the pcregrep -F option was incorrectly implemented as
1497 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1498 but it seems right to fix it, and I didn't think it was worth preserving
1499 the old behaviour.
1500
1501 5. The command line items --regex=pattern and --regexp=pattern were not
1502 recognized by pcregrep, which required --regex pattern or --regexp pattern
1503 (with a space rather than an '='). The man page documented the '=' forms,
1504 which are compatible with GNU grep; these now work.
1505
1506 6. No libpcreposix.pc file was created for pkg-config; there was just
1507 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1508
1509 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1510 when UCP support is not needed, by modifying the Python script that
1511 generates it from Unicode data files. This should not matter if the module
1512 is correctly used as a library, but I received one complaint about 50K of
1513 unwanted data. My guess is that the person linked everything into his
1514 program rather than using a library. Anyway, it does no harm.
1515
1516 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1517 was a minimum greater than 1 for a wide character in a possessive
1518 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1519 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1520 character. Chaos in the form of incorrect output or a compiling loop could
1521 result.
1522
1523 9. The restrictions on what a pattern can contain when partial matching is
1524 requested for pcre_exec() have been removed. All patterns can now be
1525 partially matched by this function. In addition, if there are at least two
1526 slots in the offset vector, the offset of the earliest inspected character
1527 for the match and the offset of the end of the subject are set in them when
1528 PCRE_ERROR_PARTIAL is returned.
1529
1530 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1531 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1532 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1533 and may be more useful for multi-segment matching.
1534
1535 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1536 used to be given if ever the end of the subject was reached; now it is
1537 given only if matching could not proceed because another character was
1538 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1539 string "Z", which now yields "no match" instead of "partial match". In the
1540 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1541 final character ended with (*FAIL).
1542
1543 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1544 if the pattern had a "must contain" character that was already found in the
1545 earlier partial match, unless partial matching was again requested. For
1546 example, with the pattern /dog.(body)?/, the "must contain" character is
1547 "g". If the first part-match was for the string "dog", restarting with
1548 "sbody" failed. This bug has been fixed.
1549
1550 13. The string returned by pcre_dfa_exec() after a partial match has been
1551 changed so that it starts at the first inspected character rather than the
1552 first character of the match. This makes a difference only if the pattern
1553 starts with a lookbehind assertion or \b or \B (\K is not supported by
1554 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1555 matching functions compatible, and I think it's the right thing to do.
1556
1557 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1558 so that the demonstration program is easily available in environments where
1559 PCRE has not been installed from source.
1560
1561 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1562 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1563 library.
1564
1565 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1566 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1567 is not the first non-POSIX option to be added. Clearly some people find
1568 these options useful.
1569
1570 17. If a caller to the POSIX matching function regexec() passes a non-zero
1571 value for nmatch with a NULL value for pmatch, the value of
1572 nmatch is forced to zero.
1573
1574 18. RunGrepTest did not have a test for the availability of the -u option of
1575 the diff command, as RunTest does. It now checks in the same way as
1576 RunTest, and also checks for the -b option.
1577
1578 19. If an odd number of negated classes containing just a single character
1579 interposed, within parentheses, between a forward reference to a named
1580 subpattern and the definition of the subpattern, compilation crashed with
1581 an internal error, complaining that it could not find the referenced
1582 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1583 [The bug was that it was starting one character too far in when skipping
1584 over the character class, thus treating the ] as data rather than
1585 terminating the class. This meant it could skip too much.]
1586
1587 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1588 /g option in pcretest when the pattern contains \K, which makes it possible
1589 to have an empty string match not at the start, even when the pattern is
1590 anchored. Updated pcretest and pcredemo to use this option.
1591
1592 21. If the maximum number of capturing subpatterns in a recursion was greater
1593 than the maximum at the outer level, the higher number was returned, but
1594 with unset values at the outer level. The correct (outer level) value is
1595 now given.
1596
1597 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1598 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1599 make it do so. The string so far is captured, making this feature
1600 compatible with Perl.
1601
1602 23. The tests have been re-organized, adding tests 11 and 12, to make it
1603 possible to check the Perl 5.10 features against Perl 5.10.
1604
1605 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1606 pattern matches a fixed length string. PCRE did not allow this; now it
1607 does. Neither allows recursion.
1608
1609 25. I finally figured out how to implement a request to provide the minimum
1610 length of subject string that was needed in order to match a given pattern.
1611 (It was back references and recursion that I had previously got hung up
1612 on.) This code has now been added to pcre_study(); it finds a lower bound
1613 to the length of subject needed. It is not necessarily the greatest lower
1614 bound, but using it to avoid searching strings that are too short does give
1615 some useful speed-ups. The value is available to calling programs via
1616 pcre_fullinfo().
1617
1618 26. While implementing 25, I discovered to my embarrassment that pcretest had
1619 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1620 study optimizations had never been tested with that matching function.
1621 Oops. What is worse, even when it was passed study data, there was a bug in
1622 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1623 were also very few tests of studied patterns with pcre_dfa_exec().
1624
1625 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1626 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1627 on the other side of the coin, they are no longer allowed to have different
1628 names, because these cannot be distinguished in PCRE, and this has caused
1629 confusion. (This is a difference from Perl.)
1630
1631 28. When duplicate subpattern names are present (necessarily with different
1632 numbers, as required by 27 above), and a test is made by name in a
1633 conditional pattern, either for a subpattern having been matched, or for
1634 recursion in such a pattern, all the associated numbered subpatterns are
1635 tested, and the overall condition is true if the condition is true for any
1636 one of them. This is the way Perl works, and is also more like the way
1637 testing by number works.
1638
1639
1640 Version 7.9 11-Apr-09
1641 ---------------------
1642
1643 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1644 (pcretest), all targets were linked against these libraries. This included
1645 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1646 libraries. This caused unwanted dependencies to be created. This problem
1647 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1648 pcretest is linked with readline.
1649
1650 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1651 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1652 moved outside it again, because FALSE and TRUE are already defined in AIX,
1653 but BOOL is not.
1654
1655 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1656 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1657
1658 4. The pcregrep documentation said spaces were inserted as well as colons (or
1659 hyphens) following file names and line numbers when outputting matching
1660 lines. This is not true; no spaces are inserted. I have also clarified the
1661 wording for the --colour (or --color) option.
1662
1663 5. In pcregrep, when --colour was used with -o, the list of matching strings
1664 was not coloured; this is different to GNU grep, so I have changed it to be
1665 the same.
1666
1667 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1668 each matching line was coloured. Now it goes on to look for further matches
1669 of any of the test patterns, which is the same behaviour as GNU grep.
1670
1671 7. A pattern that could match an empty string could cause pcregrep to loop; it
1672 doesn't make sense to accept an empty string match in pcregrep, so I have
1673 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1674 seems to be how GNU grep behaves. [But see later change 40 for release
1675 8.33.]
1676
1677 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1678 start or after a newline", because the conditional assertion was not being
1679 correctly handled. The rule now is that both the assertion and what follows
1680 in the first alternative must satisfy the test.
1681
1682 9. If auto-callout was enabled in a pattern with a conditional group whose
1683 condition was an assertion, PCRE could crash during matching, both with
1684 pcre_exec() and pcre_dfa_exec().
1685
1686 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1687 used for matching.
1688
1689 11. Unicode property support in character classes was not working for
1690 characters (bytes) greater than 127 when not in UTF-8 mode.
1691
1692 12. Added the -M command line option to pcretest.
1693
1694 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1695
1696 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1697
1698 16. Added comments and documentation about mis-use of no_arg in the C++
1699 wrapper.
1700
1701 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1702 from Martin Jerabek that uses macro names for all relevant character and
1703 string constants.
1704
1705 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1706 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1707 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1708 these, but not everybody uses configure.
1709
1710 19. A conditional group that had only one branch was not being correctly
1711 recognized as an item that could match an empty string. This meant that an
1712 enclosing group might also not be so recognized, causing infinite looping
1713 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1714 with the subject "ab", where knowledge that the repeated group can match
1715 nothing is needed in order to break the loop.
1716
1717 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1718 exec(), but without supplying a callout function, matching went wrong.
1719
1720 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1721 leak if the size of the offset vector was greater than 30. When the vector
1722 is smaller, the saved offsets during recursion go onto a local stack
1723 vector, but for larger vectors malloc() is used. It was failing to free
1724 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1725 error, in fact).
1726
1727 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1728 heapframe that is used only when UTF-8 support is enabled. This caused no
1729 problem, but was untidy.
1730
1731 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1732 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1733 included within another project.
1734
1735 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1736 slightly modified by me:
1737
1738 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1739 not building pcregrep.
1740
1741 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1742 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1743
1744 25. Forward references, both numeric and by name, in patterns that made use of
1745 duplicate group numbers, could behave incorrectly or give incorrect errors,
1746 because when scanning forward to find the reference group, PCRE was not
1747 taking into account the duplicate group numbers. A pattern such as
1748 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1749
1750 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1751 the feature of a custom pointer more persuasive (as requested by a user).
1752
1753 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1754 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1755 pre-defined.
1756
1757 28. Added support for (*UTF8) at the start of a pattern.
1758
1759 29. Arrange for flags added by the "release type" setting in CMake to be shown
1760 in the configuration summary.
1761
1762
1763 Version 7.8 05-Sep-08
1764 ---------------------
1765
1766 1. Replaced UCP searching code with optimized version as implemented for Ad
1767 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1768 stage table and inline lookup instead of a function, giving speed ups of 2
1769 to 5 times on some simple patterns that I tested. Permission was given to
1770 distribute the MultiStage2.py script that generates the tables (it's not in
1771 the tarball, but is in the Subversion repository).
1772
1773 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1774 scripts.
1775
1776 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1777 a group with a zero qualifier. The result of the study could be incorrect,
1778 or the function might crash, depending on the pattern.
1779
1780 4. Caseless matching was not working for non-ASCII characters in back
1781 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1782 It now works when Unicode Property Support is available.
1783
1784 5. In pcretest, an escape such as \x{de} in the data was always generating
1785 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1786 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1787 truncation.
1788
1789 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1790
1791 7. Added two (int) casts to pcregrep when printing the difference of two
1792 pointers, in case they are 64-bit values.
1793
1794 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1795 test 2 if it fails.
1796
1797 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1798 and a #define of that name to empty if it is not externally set. This is to
1799 allow users of MSVC to set it if necessary.
1800
1801 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1802 the convenience functions in the pcre_get.c source file.
1803
1804 11. An option change at the start of a pattern that had top-level alternatives
1805 could cause overwriting and/or a crash. This command provoked a crash in
1806 some environments:
1807
1808 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1809
1810 This potential security problem was recorded as CVE-2008-2371.
1811
1812 12. For a pattern where the match had to start at the beginning or immediately
1813 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1814 pcre_dfa_exec() could read past the end of the passed subject if there was
1815 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1816 pcretest so that it places the subject at the end of its malloc-ed buffer.
1817
1818 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1819 exec() might read past the end of the data buffer in UTF-8 mode.
1820
1821 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1822 the data contained the byte 0x85 as part of a UTF-8 character within its
1823 first line. This applied both to normal and DFA matching.
1824
1825 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1826 /^[^d]*?$/8 failed to match "abc".
1827
1828 16. Added a missing copyright notice to pcrecpp_internal.h.
1829
1830 17. Make it more clear in the documentation that values returned from
1831 pcre_exec() in ovector are byte offsets, not character counts.
1832
1833 18. Tidied a few places to stop certain compilers from issuing warnings.
1834
1835 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1836 supplied by Stefan Weber. I made a further small update for 7.8 because
1837 there is a change of source arrangements: the pcre_searchfuncs.c module is
1838 replaced by pcre_ucd.c.
1839
1840
1841 Version 7.7 07-May-08
1842 ---------------------
1843
1844 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1845 a string to a long long, pretend we don't even have a long long." This is
1846 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1847
1848 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1849 pre-7.6 versions, which defined a global no_arg variable instead of putting
1850 it in the RE class. (See also #8 below.)
1851
1852 3. Remove a line of dead code, identified by coverity and reported by Nuno
1853 Lopes.
1854
1855 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1856
1857 (1) The include/exclude patterns were being applied to the whole pathnames
1858 of files, instead of just to the final components.
1859
1860 (2) If there was more than one level of directory, the subdirectories were
1861 skipped unless they satisfied the include/exclude conditions. This is
1862 inconsistent with GNU grep (and could even be seen as contrary to the
1863 pcregrep specification - which I improved to make it absolutely clear).
1864 The action now is always to scan all levels of directory, and just
1865 apply the include/exclude patterns to regular files.
1866
1867 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1868 --exclude_dir in the tests to avoid scanning .svn directories.
1869
1870 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1871 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1872 doesn't support NULs in patterns.
1873
1874 7. Added some missing "const"s to declarations of static tables in
1875 pcre_compile.c and pcre_dfa_exec.c.
1876
1877 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1878 caused by fix #2 above. (Subsequently also a second patch to fix the
1879 first patch. And a third patch - this was a messy problem.)
1880
1881 9. Applied Craig's patch to remove the use of push_back().
1882
1883 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1884 matching function regexec().
1885
1886 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1887 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1888 references. PCRE supports relative numbers with this syntax (I don't think
1889 Oniguruma does).
1890
1891 12. Previously, a group with a zero repeat such as (...){0} was completely
1892 omitted from the compiled regex. However, this means that if the group
1893 was called as a subroutine from elsewhere in the pattern, things went wrong
1894 (an internal error was given). Such groups are now left in the compiled
1895 pattern, with a new opcode that causes them to be skipped at execution
1896 time.
1897
1898 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1899 to the way PCRE behaves:
1900
1901 (a) A lone ] character is dis-allowed (Perl treats it as data).
1902
1903 (b) A back reference to an unmatched subpattern matches an empty string
1904 (Perl fails the current match path).
1905
1906 (c) A data ] in a character class must be notated as \] because if the
1907 first data character in a class is ], it defines an empty class. (In
1908 Perl it is not possible to have an empty class.) The empty class []
1909 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1910 The negative empty class [^] matches any one character, independently
1911 of the DOTALL setting.
1912
1913 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1914 non-existent subpattern following a character class starting with ']' and
1915 containing () gave an internal compiling error instead of "reference to
1916 non-existent subpattern". Fortunately, when the pattern did exist, the
1917 compiled code was correct. (When scanning forwards to check for the
1918 existence of the subpattern, it was treating the data ']' as terminating
1919 the class, so got the count wrong. When actually compiling, the reference
1920 was subsequently set up correctly.)
1921
1922 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1923 it was being rejected as not supported by pcre_dfa_exec(), even though
1924 other assertions are supported. I have made pcre_dfa_exec() support
1925 (*FAIL).
1926
1927 16. The implementation of 13c above involved the invention of a new opcode,
1928 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1929 cannot be changed at match time, I realized I could make a small
1930 improvement to matching performance by compiling OP_ALLANY instead of
1931 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1932 on the OP_ANY path.
1933
1934 17. Compiling pcretest on Windows with readline support failed without the
1935 following two fixes: (1) Make the unistd.h include conditional on
1936 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1937
1938 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1939 ncurses library to be included for pcretest when ReadLine support is
1940 requested, but also to allow for it to be overridden. This patch came from
1941 Daniel Bergström.
1942
1943 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1944 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1945 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1946 spotting this.
1947
1948
1949 Version 7.6 28-Jan-08
1950 ---------------------
1951
1952 1. A character class containing a very large number of characters with
1953 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1954 overflow.
1955
1956 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1957 HAVE_LONG_LONG is not defined.
1958
1959 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1960 bring them up to date and include new features. This patch includes:
1961
1962 - Fixed PH's badly added libz and libbz2 support.
1963 - Fixed a problem with static linking.
1964 - Added pcredemo. [But later removed - see 7 below.]
1965 - Fixed dftables problem and added an option.
1966 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1967 HAVE_LONG_LONG.
1968 - Added readline support for pcretest.
1969 - Added an listing of the option settings after cmake has run.
1970
1971 4. A user submitted a patch to Makefile that makes it easy to create
1972 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1973 Makefile.am that cause it to include this special target, without
1974 affecting anything else. Note that the same mingw target plus all
1975 the other distribution libraries and programs are now supported
1976 when configuring with CMake (see 6 below) instead of with
1977 Configure/Make.
1978
1979 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1980 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1981 exported in the Windows port". It has not yet been confirmed that the patch
1982 solves the problem, but it does no harm.
1983
1984 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1985 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1986 with CMake, and also correct the comment about stack recursion.
1987
1988 7. Remove the automatic building of pcredemo from the ./configure system and
1989 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1990 of a program that users should build themselves after PCRE is installed, so
1991 building it automatically is not really right. What is more, it gave
1992 trouble in some build environments.
1993
1994 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1995
1996
1997 Version 7.5 10-Jan-08
1998 ---------------------
1999
2000 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2001 values in parens when parsing an RE using the C++ wrapper."
2002
2003 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2004 Characters greater than 255 were excluded from the class instead of being
2005 included.
2006
2007 3. The same bug as (2) above applied to negated POSIX classes such as
2008 [:^space:].
2009
2010 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2011 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2012 I have changed it.
2013
2014 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2015 first named subpattern) and a construct such as (?&a) would reference the
2016 first named subpattern whose name started with "a" (in other words, the
2017 length check was missing). Both these problems are fixed. "Subpattern name
2018 expected" is now given for (?&) (a zero-length name), and this patch also
2019 makes it give the same error for \k'' (previously it complained that that
2020 was a reference to a non-existent subpattern).
2021
2022 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2023 this is right because (?- can be followed by option settings as well as by
2024 digits. I have, however, made the messages clearer.
2025
2026 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2027 than the number used in the conditional) now cause a compile-time error.
2028 This is actually not compatible with Perl, which accepts such patterns, but
2029 treats the conditional as always being FALSE (as PCRE used to), but it
2030 seems to me that giving a diagnostic is better.
2031
2032 8. Change "alphameric" to the more common word "alphanumeric" in comments
2033 and messages.
2034
2035 9. Fix two occurrences of "backslash" in comments that should have been
2036 "backspace".
2037
2038 10. Remove two redundant lines of code that can never be obeyed (their function
2039 was moved elsewhere).
2040
2041 11. The program that makes PCRE's Unicode character property table had a bug
2042 which caused it to generate incorrect table entries for sequences of
2043 characters that have the same character type, but are in different scripts.
2044 It amalgamated them into a single range, with the script of the first of
2045 them. In other words, some characters were in the wrong script. There were
2046 thirteen such cases, affecting characters in the following ranges:
2047
2048 U+002b0 - U+002c1
2049 U+0060c - U+0060d
2050 U+0061e - U+00612
2051 U+0064b - U+0065e
2052 U+0074d - U+0076d
2053 U+01800 - U+01805
2054 U+01d00 - U+01d77
2055 U+01d9b - U+01dbf
2056 U+0200b - U+0200f
2057 U+030fc - U+030fe
2058 U+03260 - U+0327f
2059 U+0fb46 - U+0fbb1
2060 U+10450 - U+1049d
2061
2062 12. The -o option (show only the matching part of a line) for pcregrep was not
2063 compatible with GNU grep in that, if there was more than one match in a
2064 line, it showed only the first of them. It now behaves in the same way as
2065 GNU grep.
2066
2067 13. If the -o and -v options were combined for pcregrep, it printed a blank
2068 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2069 does the same. The return code can be used to tell if there were any
2070 non-matching lines.
2071
2072 14. Added --file-offsets and --line-offsets to pcregrep.
2073
2074 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2075 infinitely looping recursion. The bug was that positive lookaheads were not
2076 being skipped when checking for a possible empty match (negative lookaheads
2077 and both kinds of lookbehind were skipped).
2078
2079 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2080 inclusion of <windows.h> to before rather than after the definition of
2081 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2082
2083 17. Specifying a possessive quantifier with a specific limit for a Unicode
2084 character property caused pcre_compile() to compile bad code, which led at
2085 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2086 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2087 caused the error; without that there was no problem.
2088
2089 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2090
2091 19. Added --enable-pcretest-libreadline.
2092
2093 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2094 RE::GlobalReplace(). As a result, the number of replacements returned was
2095 double what it should be. I removed one of the increments, but Craig sent a
2096 later patch that removed the other one (the right fix) and added unit tests
2097 that check the return values (which was not done before).
2098
2099 21. Several CMake things:
2100
2101 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2102 the names libpcre and libpcreposix, not just pcre and pcreposix.
2103
2104 (2) The above change means that pcretest and pcregrep are now correctly
2105 linked with the newly-built libraries, not previously installed ones.
2106
2107 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2108
2109 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2110 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2111 UTF-8 newline character). The key issue is that the pattern starts .*;
2112 this means that the match must be either at the beginning, or after a
2113 newline. The bug was in the code for advancing after a failed match and
2114 checking that the new position followed a newline. It was not taking
2115 account of UTF-8 characters correctly.
2116
2117 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2118 character classes. PCRE was not treating the sequence [:...:] as a
2119 character class unless the ... were all letters. Perl, however, seems to
2120 allow any characters between [: and :], though of course it rejects as
2121 unknown any "names" that contain non-letters, because all the known class
2122 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2123 for example, whereas PCRE did not - it did not recognize a POSIX character
2124 class. This seemed a bit dangerous, so the code has been changed to be
2125 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2126 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2127 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2128 Perl does, and where it didn't before.
2129
2130 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2131 Windows environments %n is disabled by default.
2132
2133
2134 Version 7.4 21-Sep-07
2135 ---------------------
2136
2137 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2138 means that a class such as [\s] counted as "explicit reference to CR or
2139 LF". That isn't really right - the whole point of the change was to try to
2140 help when there was an actual mention of one of the two characters. So now
2141 the change happens only if \r or \n (or a literal CR or LF) character is
2142 encountered.
2143
2144 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2145 of both had grown to the point where there were only 3 bits left.
2146 Fortunately, there was spare space in the data structure, and so I have
2147 moved the internal flags into a new 16-bit field to free up more option
2148 bits.
2149
2150 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2151 but did not set the internal JCHANGED flag - either of these is enough to
2152 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2153 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2154 start sets both bits.
2155
2156 4. Added options (at build time, compile time, exec time) to change \R from
2157 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2158
2159 5. doc/pcresyntax.html was missing from the distribution.
2160
2161 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2162 compatibility, even though it is no longer used.
2163
2164 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2165 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2166 windows.h file is present (where different names are used). [This was
2167 reversed later after testing - see 16 below.]
2168
2169 8. Changed all #include <config.h> to #include "config.h". There were also
2170 some further <pcre.h> cases that I changed to "pcre.h".
2171
2172 9. When pcregrep was used with the --colour option, it missed the line ending
2173 sequence off the lines that it output.
2174
2175 10. It was pointed out to me that arrays of string pointers cause lots of
2176 relocations when a shared library is dynamically loaded. A technique of
2177 using a single long string with a table of offsets can drastically reduce
2178 these. I have refactored PCRE in four places to do this. The result is
2179 dramatic:
2180
2181 Originally: 290
2182 After changing UCP table: 187
2183 After changing error message table: 43
2184 After changing table of "verbs" 36
2185 After changing table of Posix names 22
2186
2187 Thanks to the folks working on Gregex for glib for this insight.
2188
2189 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2190 unicode-properties was also set.
2191
2192 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2193
2194 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2195 checked only for CRLF.
2196
2197 14. Added casts to pcretest.c to avoid compiler warnings.
2198
2199 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2200
2201 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2202 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2203 entirely. This removes changes made in 7 above.
2204
2205 17. The CMake files have been updated, and there is now more information about
2206 building with CMake in the NON-UNIX-USE document.
2207
2208
2209 Version 7.3 28-Aug-07
2210 ---------------------
2211
2212 1. In the rejigging of the build system that eventually resulted in 7.1, the
2213 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2214 brackets there is not right, since it causes compilers to look for an
2215 installed pcre.h, not the version that is in the source that is being
2216 compiled (which of course may be different). I have changed it back to:
2217
2218 #include "pcre.h"
2219
2220 I have a vague recollection that the change was concerned with compiling in
2221 different directories, but in the new build system, that is taken care of
2222 by the VPATH setting the Makefile.
2223
2224 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2225 when the subject happened to end in the byte 0x85 (e.g. if the last
2226 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2227 characters but of course it shouldn't be taken as a newline when it is part
2228 of another character. The bug was that, for an unlimited repeat of . in
2229 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2230 characters when looking for a newline.
2231
2232 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2233
2234 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2235 in debug output.
2236
2237 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2238 long printing in the pcrecpp unittest when running under MinGW.
2239
2240 6. ESC_K was left out of the EBCDIC table.
2241
2242 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2243 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2244 limit also applies to "virtual nesting" when a pattern is recursive, and in
2245 this case 1000 isn't so big. I have been able to remove this limit at the
2246 expense of backing off one optimization in certain circumstances. Normally,
2247 when pcre_exec() would call its internal match() function recursively and
2248 immediately return the result unconditionally, it uses a "tail recursion"
2249 feature to save stack. However, when a subpattern that can match an empty
2250 string has an unlimited repetition quantifier, it no longer makes this
2251 optimization. That gives it a stack frame in which to save the data for
2252 checking that an empty string has been matched. Previously this was taken
2253 from the 1000-entry workspace that had been reserved. So now there is no
2254 explicit limit, but more stack is used.
2255
2256 8. Applied Daniel's patches to solve problems with the import/export magic
2257 syntax that is required for Windows, and which was going wrong for the
2258 pcreposix and pcrecpp parts of the library. These were overlooked when this
2259 problem was solved for the main library.
2260
2261 9. There were some crude static tests to avoid integer overflow when computing
2262 the size of patterns that contain repeated groups with explicit upper
2263 limits. As the maximum quantifier is 65535, the maximum group length was
2264 set at 30,000 so that the product of these two numbers did not overflow a
2265 32-bit integer. However, it turns out that people want to use groups that
2266 are longer than 30,000 bytes (though not repeat them that many times).
2267 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2268 made it possible to implement the integer overflow checks in a much more
2269 dynamic way, which I have now done. The artificial limitation on group
2270 length has been removed - we now have only the limit on the total length of
2271 the compiled pattern, which depends on the LINK_SIZE setting.
2272
2273 10. Fixed a bug in the documentation for get/copy named substring when
2274 duplicate names are permitted. If none of the named substrings are set, the
2275 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2276 empty string.
2277
2278 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2279 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2280 because the ] is interpreted as the first data character and the
2281 terminating ] is not found. PCRE has been made compatible with Perl in this
2282 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2283 cause memory overwriting.
2284
2285 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2286 string has been matched (to stop an infinite loop). It was not recognizing
2287 a conditional subpattern that could match an empty string if that
2288 subpattern was within another subpattern. For example, it looped when
2289 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2290 condition was not nested. This bug has been fixed.
2291
2292 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2293 past the start of the subject in the presence of bytes with the top bit
2294 set, for example "\x8aBCD".
2295
2296 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2297 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2298
2299 14. Optimized (?!) to (*FAIL).
2300
2301 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2302 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2303 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2304 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2305 does: it's just the validity check that is more restrictive.
2306
2307 16. Inserted checks for integer overflows during escape sequence (backslash)
2308 processing, and also fixed erroneous offset values for syntax errors during
2309 backslash processing.
2310
2311 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2312 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2313
2314 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2315 caused an overrun.
2316
2317 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2318 something other than just ASCII characters) inside a group that had an
2319 unlimited repeat caused a loop at compile time (while checking to see
2320 whether the group could match an empty string).
2321
2322 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2323 [\P{Any}] did so. (Error in the code for printing property names.)
2324
2325 21. An orphan \E inside a character class could cause a crash.
2326
2327 22. A repeated capturing bracket such as (A)? could cause a wild memory
2328 reference during compilation.
2329
2330 23. There are several functions in pcre_compile() that scan along a compiled
2331 expression for various reasons (e.g. to see if it's fixed length for look
2332 behind). There were bugs in these functions when a repeated \p or \P was
2333 present in the pattern. These operators have additional parameters compared
2334 with \d, etc, and these were not being taken into account when moving along
2335 the compiled data. Specifically:
2336
2337 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2338 length.
2339
2340 (b) An item such as \pL+ within a repeated group could cause crashes or
2341 loops.
2342
2343 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2344 "reference to non-existent subpattern" error.
2345
2346 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2347
2348 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2349 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2350
2351 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2352
2353 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2354 character were causing crashes (broken optimization).
2355
2356 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2357 \p or \P) caused a compile-time loop.
2358
2359 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2360 break. For example, the unstudied pattern [\r\n]A does not match the string
2361 "\r\nA" because change 7.0/46 below moves the current point on by two
2362 characters after failing to match at the start. However, the pattern \nA
2363 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2364 the same is true. There doesn't seem any very clean way out of this, but
2365 what I have chosen to do makes the common cases work: PCRE now takes note
2366 of whether there can be an explicit match for \r or \n anywhere in the
2367 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2368 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2369 pattern has explicit CR or LF references.
2370
2371 29. Added (*CR) etc for changing newline setting at start of pattern.
2372
2373
2374 Version 7.2 19-Jun-07
2375 ---------------------
2376
2377 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2378 which is apparently normally available under Windows.
2379
2380 2. Re-jig the pcregrep tests with different newline settings in an attempt
2381 to make them independent of the local environment's newline setting.
2382
2383 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2384
2385 4. Some of the "internals" tests were previously cut out when the link size
2386 was not 2, because the output contained actual offsets. The recent new
2387 "Z" feature of pcretest means that these can be cut out, making the tests
2388 usable with all link sizes.
2389
2390 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2391 stack recursion. This gives a massive performance boost under BSD, but just
2392 a small improvement under Linux. However, it saves one field in the frame
2393 in all cases.
2394
2395 6. Added more features from the forthcoming Perl 5.10:
2396
2397 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2398 recursion call. It refers to the nth most recently opened parentheses.
2399
2400 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2401 to be opened parentheses.
2402
2403 (c) Conditions that refer to capturing parentheses can be specified
2404 relatively, for example, (?(-2)... or (?(+3)...
2405
2406 (d) \K resets the start of the current match so that everything before
2407 is not part of it.
2408
2409 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2410
2411 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2412 reference syntax.
2413
2414 (g) (?| introduces a group in which the numbering of parentheses in each
2415 alternative starts with the same number.
2416
2417 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2418
2419 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2420 PCRE_INFO_JCHANGED.
2421
2422 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2423 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2424 for detecting groups that can match an empty string.
2425
2426 9. A pattern with a very large number of alternatives (more than several
2427 hundred) was running out of internal workspace during the pre-compile
2428 phase, where pcre_compile() figures out how much memory will be needed. A
2429 bit of new cunning has reduced the workspace needed for groups with
2430 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2431 workspace instead of running out of the 4096 that are available.
2432
2433 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2434
2435 11. Applied patch from Google to remove an optimization that didn't quite work.
2436 The report of the bug said:
2437
2438 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2439 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2440 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2441
2442 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2443 it matched the wrong number of bytes.
2444
2445
2446 Version 7.1 24-Apr-07
2447 ---------------------
2448
2449 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2450 that is more "standard", making use of automake and other Autotools. There
2451 is some re-arrangement of the files and adjustment of comments consequent
2452 on this.
2453
2454 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2455 for recursive directory scanning broke on some systems because the files
2456 are not scanned in any specific order and on different systems the order
2457 was different. A call to "sort" has been inserted into RunGrepTest for the
2458 approprate test as a short-term fix. In the longer term there may be an
2459 alternative.
2460
2461 3. I had an email from Eric Raymond about problems translating some of PCRE's
2462 man pages to HTML (despite the fact that I distribute HTML pages, some
2463 people do their own conversions for various reasons). The problems
2464 concerned the use of low-level troff macros .br and .in. I have therefore
2465 removed all such uses from the man pages (some were redundant, some could
2466 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2467 HTML has been updated to handle .nf/.fi and to complain if it encounters
2468 .br or .in.
2469
2470 4. Updated comments in configure.ac that get placed in config.h.in and also
2471 arranged for config.h to be included in the distribution, with the name
2472 config.h.generic, for the benefit of those who have to compile without
2473 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2474
2475 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2476 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2477 makevp.bat for the current PCRE, using the additional files
2478 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2479
2480 6. A Windows user reported a minor discrepancy with test 2, which turned out
2481 to be caused by a trailing space on an input line that had got lost in his
2482 copy. The trailing space was an accident, so I've just removed it.
2483
2484 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2485 that is needed.
2486
2487 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2488 as "const" (a) because they are and (b) because it helps the PHP
2489 maintainers who have recently made a script to detect big data structures
2490 in the php code that should be moved to the .rodata section. I remembered
2491 to update Builducptable as well, so it won't revert if ucptable.h is ever
2492 re-created.
2493
2494 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2495 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2496 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2497 support is not required. This saves 1.5-2K of code, which is important in
2498 some applications.
2499
2500 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2501 so as not to refer to the tables, even though these functions will never be
2502 called when UTF-8 support is disabled. Otherwise there are problems with a
2503 shared library.
2504
2505 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2506
2507 (a) It was defining its arguments as char * instead of void *.
2508
2509 (b) It was assuming that all moves were upwards in memory; this was true
2510 a long time ago when I wrote it, but is no longer the case.
2511
2512 The emulated memove() is provided for those environments that have neither
2513 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2514 is clearly not the case, as these two bugs were recently reported.
2515
2516 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2517 and Detrail to create the HTML documentation, the .txt form of the man
2518 pages, and it removes trailing spaces from listed files. It also creates
2519 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2520 case, it wraps all the #defines with #ifndefs. This script should be run
2521 before "make dist".
2522
2523 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2524 with Unicode property support.
2525
2526 (a) For a maximizing quantifier, if the two different cases of the
2527 character were of different lengths in their UTF-8 codings (there are
2528 some cases like this - I found 11), and the matching function had to
2529 back up over a mixture of the two cases, it incorrectly assumed they
2530 were both the same length.
2531
2532 (b) When PCRE was configured to use the heap rather than the stack for
2533 recursion during matching, it was not correctly preserving the data for
2534 the other case of a UTF-8 character when checking ahead for a match
2535 while processing a minimizing repeat. If the check also involved
2536 matching a wide character, but failed, corruption could cause an
2537 erroneous result when trying to check for a repeat of the original
2538 character.
2539
2540 13. Some tidying changes to the testing mechanism:
2541
2542 (a) The RunTest script now detects the internal link size and whether there
2543 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2544 values substituted by "configure". (The RunGrepTest script already did
2545 this for UTF-8.) The configure.ac script no longer substitutes the
2546 relevant variables.
2547
2548 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2549 with length and offset values. This means that the output is different
2550 for different internal link sizes. Test 2 is skipped for link sizes
2551 other than 2 because of this, bypassing the problem. Unfortunately,
2552 there was also a test in test 3 (the locale tests) that used /B and
2553 failed for link sizes other than 2. Rather than cut the whole test out,
2554 I have added a new /Z option to pcretest that replaces the length and
2555 offset values with spaces. This is now used to make test 3 independent
2556 of link size. (Test 2 will be tidied up later.)
2557
2558 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2559 segmentation fault instead of returning the appropriate error message.
2560
2561 15. In multiline mode when the newline sequence was set to "any", the pattern
2562 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2563 This doesn't seem right; it now treats the CRLF combination as the line
2564 ending, and so does not match in that case. It's only a pattern such as ^$
2565 that would hit this one: something like ^ABC$ would have failed after \r
2566 and then tried again after \r\n.
2567
2568 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2569 in an attempt to make files that differ only in their line terminators
2570 compare equal. This works on Linux.
2571
2572 17. Under certain error circumstances pcregrep might try to free random memory
2573 as it exited. This is now fixed, thanks to valgrind.
2574
2575 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2576 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2577 was because its rules for how to advance for /g after matching an empty
2578 string at the end of a line did not allow for this case. They now check for
2579 it specially.
2580
2581 20. pcretest is supposed to handle patterns and data of any length, by
2582 extending its buffers when necessary. It was getting this wrong when the
2583 buffer for a data line had to be extended.
2584
2585 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2586 CRLF as a newline sequence.
2587
2588 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2589 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2590 I have nevertheless tidied it up.
2591
2592 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2593
2594 24. Added a man page for pcre-config.
2595
2596
2597 Version 7.0 19-Dec-06
2598 ---------------------
2599
2600 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2601 moving to gcc 4.1.1.
2602
2603 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2604 sys/time.h, which is documented as needed for this function. It doesn't
2605 seem to matter on Linux, but it showed up on some releases of OS X.
2606
2607 3. It seems that there are systems where bytes whose values are greater than
2608 127 match isprint() in the "C" locale. The "C" locale should be the
2609 default when a C program starts up. In most systems, only ASCII printing
2610 characters match isprint(). This difference caused the output from pcretest
2611 to vary, making some of the tests fail. I have changed pcretest so that:
2612
2613 (a) When it is outputting text in the compiled version of a pattern, bytes
2614 other than 32-126 are always shown as hex escapes.
2615
2616 (b) When it is outputting text that is a matched part of a subject string,
2617 it does the same, unless a different locale has been set for the match
2618 (using the /L modifier). In this case, it uses isprint() to decide.
2619
2620 4. Fixed a major bug that caused incorrect computation of the amount of memory
2621 required for a compiled pattern when options that changed within the
2622 pattern affected the logic of the preliminary scan that determines the
2623 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2624 that the computed length was too small. The symptoms of this bug were
2625 either the PCRE error "internal error: code overflow" from pcre_compile(),
2626 or a glibc crash with a message such as "pcretest: free(): invalid next
2627 size (fast)". Examples of patterns that provoked this bug (shown in
2628 pcretest format) are:
2629
2630 /(?-x: )/x
2631 /(?x)(?-x: \s*#\s*)/
2632 /((?i)[\x{c0}])/8
2633 /(?i:[\x{c0}])/8
2634
2635 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2636 is now done differently.
2637
2638 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2639 wrapper classes; (b) implement a new function in the C++ scanner that is
2640 more efficient than the old way of doing things because it avoids levels of
2641 recursion in the regex matching; (c) add a paragraph to the documentation
2642 for the FullMatch() function.
2643
2644 6. The escape sequence \n was being treated as whatever was defined as
2645 "newline". Not only was this contrary to the documentation, which states
2646 that \n is character 10 (hex 0A), but it also went horribly wrong when
2647 "newline" was defined as CRLF. This has been fixed.
2648
2649 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2650 was being set to -1 for the "end of line" case (supposedly a value that no
2651 character can have). Though this value is never used (the check for end of
2652 line is "zero bytes in current character"), it caused compiler complaints.
2653 I've changed it to 0xffffffff.
2654
2655 8. In pcre_version.c, the version string was being built by a sequence of
2656 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2657 string (as it is for production releases) called a macro with an empty
2658 argument. The C standard says the result of this is undefined. The gcc
2659 compiler treats it as an empty string (which was what was wanted) but it is
2660 reported that Visual C gives an error. The source has been hacked around to
2661 avoid this problem.
2662
2663 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2664 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2665 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2666 of them did).
2667
2668 10. Originally, pcretest opened its input and output without "b"; then I was
2669 told that "b" was needed in some environments, so it was added for release
2670 5.0 to both the input and output. (It makes no difference on Unix-like
2671 systems.) Later I was told that it is wrong for the input on Windows. I've
2672 now abstracted the modes into two macros, to make it easier to fiddle with
2673 them, and removed "b" from the input mode under Windows.
2674
2675 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2676
2677 12. Added -help and --help to pcretest as an official way of being reminded
2678 of the options.
2679
2680 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2681 and pcrecpp.cc because they annoy compilers at high warning levels.
2682
2683 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2684
2685 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2686 scripts are not C programs :-) and which was not noticed because it works
2687 on Linux.
2688
2689 16. pcretest is supposed to handle any length of pattern and data line (as one
2690 line or as a continued sequence of lines) by extending its input buffer if
2691 necessary. This feature was broken for very long pattern lines, leading to
2692 a string of junk being passed to pcre_compile() if the pattern was longer
2693 than about 50K.
2694
2695 17. I have done a major re-factoring of the way pcre_compile() computes the
2696 amount of memory needed for a compiled pattern. Previously, there was code
2697 that made a preliminary scan of the pattern in order to do this. That was
2698 OK when PCRE was new, but as the facilities have expanded, it has become
2699 harder and harder to keep it in step with the real compile phase, and there
2700 have been a number of bugs (see for example, 4 above). I have now found a
2701 cunning way of running the real compile function in a "fake" mode that
2702 enables it to compute how much memory it would need, while actually only
2703 ever using a few hundred bytes of working memory and without too many
2704 tests of the mode. This should make future maintenance and development
2705 easier. A side effect of this work is that the limit of 200 on the nesting
2706 depth of parentheses has been removed (though this was never a serious
2707 limitation, I suspect). However, there is a downside: pcre_compile() now
2708 runs more slowly than before (30% or more, depending on the pattern). I
2709 hope this isn't a big issue. There is no effect on runtime performance.
2710
2711 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2712 newline (only possible for the last line of a file) and it was a
2713 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2714
2715 19. Added additional timing features to pcretest. (1) The -tm option now times
2716 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2717 separate command line item, by a number that specifies the number of
2718 repeats to use when timing. The default is 50000; this gives better
2719 precision, but takes uncomfortably long for very large patterns.
2720
2721 20. Extended pcre_study() to be more clever in cases where a branch of a
2722 subpattern has no definite first character. For example, (a*|b*)[cd] would
2723 previously give no result from pcre_study(). Now it recognizes that the
2724 first character must be a, b, c, or d.
2725
2726 21. There was an incorrect error "recursive call could loop indefinitely" if
2727 a subpattern (or the entire pattern) that was being tested for matching an
2728 empty string contained only one non-empty item after a nested subpattern.
2729 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2730 incorrectly, because the \d was being skipped in the check.
2731
2732 22. The pcretest program now has a new pattern option /B and a command line
2733 option -b, which is equivalent to adding /B to every pattern. This causes
2734 it to show the compiled bytecode, without the additional information that
2735 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2736 is the same as /B/I).
2737
2738 23. A new optimization is now able automatically to treat some sequences such
2739 as a*b as a*+b. More specifically, if something simple (such as a character
2740 or a simple class like \d) has an unlimited quantifier, and is followed by
2741 something that cannot possibly match the quantified thing, the quantifier
2742 is automatically "possessified".
2743
2744 24. A recursive reference to a subpattern whose number was greater than 39
2745 went wrong under certain circumstances in UTF-8 mode. This bug could also
2746 have affected the operation of pcre_study().
2747
2748 25. Realized that a little bit of performance could be had by replacing
2749 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2750
2751 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2752
2753 27. Possessive quantifiers such as a++ were previously implemented by turning
2754 them into atomic groups such as ($>a+). Now they have their own opcodes,
2755 which improves performance. This includes the automatically created ones
2756 from 23 above.
2757
2758 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2759 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2760 the first matched character to be a colon. This applied both to named and
2761 numbered groups.
2762
2763 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2764
2765 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2766 building PCRE on Windows easier, so I have included it in the distribution.
2767
2768 31. There is now a check in pcretest against a ridiculously large number being
2769 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2770 loop, the loop is abandoned.
2771
2772 32. Forward references to subpatterns in conditions such as (?(2)...) where
2773 subpattern 2 is defined later cause pcre_compile() to search forwards in
2774 the pattern for the relevant set of parentheses. This search went wrong
2775 when there were unescaped parentheses in a character class, parentheses
2776 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2777
2778 33. "Subroutine" calls and backreferences were previously restricted to
2779 referencing subpatterns earlier in the regex. This restriction has now
2780 been removed.
2781
2782 34. Added a number of extra features that are going to be in Perl 5.10. On the
2783 whole, these are just syntactic alternatives for features that PCRE had
2784 previously implemented using the Python syntax or my own invention. The
2785 other formats are all retained for compatibility.
2786
2787 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2788 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2789 also .NET compatible.
2790
2791 (b) A recursion or subroutine call to a named group can now be defined as
2792 (?&name) as well as (?P>name).
2793
2794 (c) A backreference to a named group can now be defined as \k<name> or
2795 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2796 5.10, are also .NET compatible.
2797
2798 (d) A conditional reference to a named group can now use the syntax
2799 (?(<name>) or (?('name') as well as (?(name).
2800
2801 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2802 groups (named and numbered) that are never evaluated inline, but can be
2803 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2804 is always false. There may be only one alternative in such a group.
2805
2806 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2807 as the simple (?(R). The condition is true only if the most recent
2808 recursion is that of the given number or name. It does not search out
2809 through the entire recursion stack.
2810
2811 (g) The escape \gN or \g{N} has been added, where N is a positive or
2812 negative number, specifying an absolute or relative reference.
2813
2814 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2815 some "unreachable code" warnings.
2816
2817 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2818 things, this adds five new scripts.
2819
2820 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2821 There were also incompatibilities regarding the handling of \Q..\E inside
2822 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2823 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2824
2825 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2826 matches an empty string, and forcibly breaks the loop. There were bugs in
2827 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2828 against aaaa the result was just "a" rather than "aaaa", for example. Two
2829 separate and independent bugs (that affected different cases) have been
2830 fixed.
2831
2832 39. Refactored the code to abolish the use of different opcodes for small
2833 capturing bracket numbers. This is a tidy that I avoided doing when I
2834 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2835 The new approach is not only tidier, it makes it possible to reduce the
2836 memory needed to fix the previous bug (38).
2837
2838 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2839 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2840 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2841 mode.
2842
2843 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2844 report.
2845
2846 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2847 copy construction and assignment in the C++ wrapper.
2848
2849 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2850 couple of bugs that could have given wrong results in the "--newline=crlf"
2851 case.
2852
2853 44. Added a number of casts and did some reorganization of signed/unsigned int
2854 variables following suggestions from Dair Grant. Also renamed the variable
2855 "this" as "item" because it is a C++ keyword.
2856
2857 45. Arranged for dftables to add
2858
2859 #include "pcre_internal.h"
2860
2861 to pcre_chartables.c because without it, gcc 4.x may remove the array
2862 definition from the final binary if PCRE is built into a static library and
2863 dead code stripping is activated.
2864
2865 46. For an unanchored pattern, if a match attempt fails at the start of a
2866 newline sequence, and the newline setting is CRLF or ANY, and the next two
2867 characters are CRLF, advance by two characters instead of one.
2868
2869
2870 Version 6.7 04-Jul-06
2871 ---------------------
2872
2873 1. In order to handle tests when input lines are enormously long, pcretest has
2874 been re-factored so that it automatically extends its buffers when
2875 necessary. The code is crude, but this _is_ just a test program. The
2876 default size has been increased from 32K to 50K.
2877
2878 2. The code in pcre_study() was using the value of the re argument before
2879 testing it for NULL. (Of course, in any sensible call of the function, it
2880 won't be NULL.)
2881
2882 3. The memmove() emulation function in pcre_internal.h, which is used on
2883 systems that lack both memmove() and bcopy() - that is, hardly ever -
2884 was missing a "static" storage class specifier.
2885
2886 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2887 containing an extended class (one that cannot be represented by a bitmap
2888 because it contains high-valued characters or Unicode property items, e.g.
2889 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2890 pattern, but PCRE should not loop if you do not (it no longer does).
2891 [Detail: two cases were found: (a) a repeated subpattern containing an
2892 extended class; (b) a recursive reference to a subpattern that followed a
2893 previous extended class. It wasn't skipping over the extended class
2894 correctly when UTF-8 mode was not set.]
2895
2896 5. A negated single-character class was not being recognized as fixed-length
2897 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2898 compile error "lookbehind assertion is not fixed length".
2899
2900 6. The RunPerlTest auxiliary script was showing an unexpected difference
2901 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2902 write a Perl script that can interpret lines of an input file either as
2903 byte characters or as UTF-8, which is what "perltest" was being required to
2904 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2905 can't do is switch easily at run time between having the "use utf8;" pragma
2906 or not. In the end, I fudged it by using the RunPerlTest script to insert
2907 "use utf8;" explicitly for the UTF-8 tests.
2908
2909 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2910 the end of the subject string, contrary to the documentation and to what
2911 Perl does. This was true of both matching functions. Now it matches only at
2912 the start of the subject and immediately after *internal* newlines.
2913
2914 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2915 a pointer to an int instead of a pointer to an unsigned long int. This
2916 caused problems on 64-bit systems.
2917
2918 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2919 instance of the 'standard' template library not being so standard".
2920
2921 10. There was no check on the number of named subpatterns nor the maximum
2922 length of a subpattern name. The product of these values is used to compute
2923 the size of the memory block for a compiled pattern. By supplying a very
2924 long subpattern name and a large number of named subpatterns, the size
2925 computation could be caused to overflow. This is now prevented by limiting
2926 the length of names to 32 characters, and the number of named subpatterns
2927 to 10,000.
2928
2929 11. Subpatterns that are repeated with specific counts have to be replicated in
2930 the compiled pattern. The size of memory for this was computed from the
2931 length of the subpattern and the repeat count. The latter is limited to
2932 65535, but there was no limit on the former, meaning that integer overflow
2933 could in principle occur. The compiled length of a repeated subpattern is
2934 now limited to 30,000 bytes in order to prevent this.
2935
2936 12. Added the optional facility to have named substrings with the same name.
2937
2938 13. Added the ability to use a named substring as a condition, using the
2939 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2940 are numbers (not recommended). Forward references are permitted.
2941
2942 14. Added forward references in named backreferences (if you see what I mean).
2943
2944 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2945 pattern could run off the end of the subject. For example, the pattern
2946 "(?s)(.{1,5})"8 did this with the subject "ab".
2947
2948 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2949 PCRE_CASELESS was set when matching characters that were quantified with ?
2950 or *.
2951
2952 17. A character class other than a single negated character that had a minimum
2953 but no maximum quantifier - for example [ab]{6,} - was not handled
2954 correctly by pce_dfa_exec(). It would match only one character.
2955
2956 18. A valid (though odd) pattern that looked like a POSIX character
2957 class but used an invalid character after [ (for example [[,abc,]]) caused
2958 pcre_compile() to give the error "Failed: internal error: code overflow" or
2959 in some cases to crash with a glibc free() error. This could even happen if
2960 the pattern terminated after [[ but there just happened to be a sequence of
2961 letters, a binary zero, and a closing ] in the memory that followed.
2962
2963 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2964 over the years. Originally (before any Unicode support), just the bottom 8
2965 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2966 output from "man perlunicode" includes this:
2967
2968 The regular expression compiler produces polymorphic opcodes. That
2969 is, the pattern adapts to the data and automatically switches to
2970 the Unicode character scheme when presented with Unicode data--or
2971 instead uses a traditional byte scheme when presented with byte
2972 data.
2973
2974 Sadly, a wide octal escape does not cause a switch, and in a string with
2975 no other multibyte characters, these octal escapes are treated as before.
2976 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2977 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2978 Unicode string.
2979
2980 I have not perpetrated such confusion in PCRE. Up till now, it took just
2981 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2982 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2983 translate to the appropriate multibyte character.
2984
2985 29. Applied some refactoring to reduce the number of warnings from Microsoft
2986 and Borland compilers. This has included removing the fudge introduced
2987 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2988 a warning about an unused variable.
2989
2990 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2991 characters since release 4.0, because Perl (from release 5.004) does not.
2992 [Or at least, is documented not to: some releases seem to be in conflict
2993 with the documentation.] However, when a pattern was studied with
2994 pcre_study() and all its branches started with \s, PCRE still included VT
2995 as a possible starting character. Of course, this did no harm; it just
2996 caused an unnecessary match attempt.
2997
2998 22. Removed a now-redundant internal flag bit that recorded the fact that case
2999 dependency changed within the pattern. This was once needed for "required
3000 byte" processing, but is no longer used. This recovers a now-scarce options
3001 bit. Also moved the least significant internal flag bit to the most-
3002 significant bit of the word, which was not previously used (hangover from
3003 the days when it was an int rather than a uint) to free up another bit for
3004 the future.
3005
3006 23. Added support for CRLF line endings as well as CR and LF. As well as the
3007 default being selectable at build time, it can now be changed at runtime
3008 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3009 specify that it is scanning data with non-default line endings.
3010
3011 24. Changed the definition of CXXLINK to make it agree with the definition of
3012 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3013
3014 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3015 recursions. This makes a big different to stack usage for some patterns.
3016
3017 26. If a subpattern containing a named recursion or subroutine reference such
3018 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3019 the space required for the compiled pattern went wrong and gave too small a
3020 value. Depending on the environment, this could lead to "Failed: internal
3021 error: code overflow at offset 49" or "glibc detected double free or
3022 corruption" errors.
3023
3024 27. Applied patches from Google (a) to support the new newline modes and (b) to
3025 advance over multibyte UTF-8 characters in GlobalReplace.
3026
3027 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3028 difference for some implementation of PCRE in some Windows version.
3029
3030 29. Added some extra testing facilities to pcretest:
3031
3032 \q<number> in a data line sets the "match limit" value
3033 \Q<number> in a data line sets the "match recursion limt" value
3034 -S <number> sets the stack size, where <number> is in megabytes
3035
3036 The -S option isn't available for Windows.
3037
3038
3039 Version 6.6 06-Feb-06
3040 ---------------------
3041
3042 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3043 in pcreposix.h. I have copied the definition from pcre.h.
3044
3045 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3046 because pcre.h is no longer a built file.
3047
3048 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3049 not normally included in the compiled code.
3050
3051
3052 Version 6.5 01-Feb-06
3053 ---------------------
3054
3055 1. When using the partial match feature with pcre_dfa_exec(), it was not
3056 anchoring the second and subsequent partial matches at the new starting
3057 point. This could lead to incorrect results. For example, with the pattern
3058 /1234/, partially matching against "123" and then "a4" gave a match.
3059
3060 2. Changes to pcregrep:
3061
3062 (a) All non-match returns from pcre_exec() were being treated as failures
3063 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3064 error message is output. Some extra information is given for the
3065 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3066 probably the only errors that are likely to be caused by users (by
3067 specifying a regex that has nested indefinite repeats, for instance).
3068 If there are more than 20 of these errors, pcregrep is abandoned.
3069
3070 (b) A binary zero was treated as data while matching, but terminated the
3071 output line if it was written out. This has been fixed: binary zeroes
3072 are now no different to any other data bytes.
3073
3074 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3075 used to set a locale for matching. The --locale=xxxx long option has
3076 been added (no short equivalent) to specify a locale explicitly on the
3077 pcregrep command, overriding the environment variables.
3078
3079 (d) When -B was used with -n, some line numbers in the output were one less
3080 than they should have been.
3081
3082 (e) Added the -o (--only-matching) option.
3083
3084 (f) If -A or -C was used with -c (count only), some lines of context were
3085 accidentally printed for the final match.
3086
3087 (g) Added the -H (--with-filename) option.
3088
3089 (h) The combination of options -rh failed to suppress file names for files
3090 that were found from directory arguments.
3091
3092 (i) Added the -D (--devices) and -d (--directories) options.
3093
3094 (j) Added the -F (--fixed-strings) option.
3095
3096 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3097
3098 (l) Added the --colo(u)r option.
3099
3100 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3101 is not present by default.
3102
3103 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3104 items such as (?R) or (?1), when the recursion could match a number of
3105 alternatives. If it matched one of the alternatives, but subsequently,
3106 outside the recursion, there was a failure, the code tried to back up into
3107 the recursion. However, because of the way PCRE is implemented, this is not
3108 possible, and the result was an incorrect result from the match.
3109
3110 In order to prevent this happening, the specification of recursion has
3111 been changed so that all such subpatterns are automatically treated as
3112 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3113
3114 4. I had overlooked the fact that, in some locales, there are characters for
3115 which isalpha() is true but neither isupper() nor islower() are true. In
3116 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3117 and ordfeminine) are like this. This affected the treatment of \w and \W
3118 when they appeared in character classes, but not when they appeared outside
3119 a character class. The bit map for "word" characters is now created
3120 separately from the results of isalnum() instead of just taking it from the
3121 upper, lower, and digit maps. (Plus the underscore character, of course.)
3122
3123 5. The above bug also affected the handling of POSIX character classes such as
3124 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3125 permanent tables. Instead, the bit maps for such a class were previously
3126 created as the appropriate unions of the upper, lower, and digit bitmaps.
3127 Now they are created by subtraction from the [[:word:]] class, which has
3128 its own bitmap.
3129
3130 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3131 It is created by subtracting the vertical space characters (\x09, \x0a,
3132 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3133 subtraction was done in the overall bitmap for a character class, meaning
3134 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3135 be recognized. This bug has been fixed.
3136
3137 7. Patches from the folks at Google:
3138
3139 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3140 real life, but is still worth protecting against".
3141
3142 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3143 regular expressions".
3144
3145 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3146 have it.
3147
3148 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3149 "configure" and the latter not, in order to fix a problem somebody had
3150 with compiling the Arg class on HP-UX.
3151
3152 (e) Improve the error-handling of the C++ wrapper a little bit.
3153
3154 (f) New tests for checking recursion limiting.
3155
3156 8. The pcre_memmove() function, which is used only if the environment does not
3157 have a standard memmove() function (and is therefore rarely compiled),
3158 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3159 returning a result (though PCRE never actually uses the result).
3160
3161 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3162 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3163 returned instead of calling malloc() with an overflowing number that would
3164 most likely cause subsequent chaos.
3165
3166 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3167
3168 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3169 with this option is matched, the nmatch and pmatch options of regexec() are
3170 ignored.
3171
3172 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3173 provided in case anyone wants to the the POSIX interface with UTF-8
3174 strings.
3175
3176 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3177 C++ linking (needed for some HP-UX environments).
3178
3179 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3180 (unused parameter) and in the pcre_printint() function (omitted "default"
3181 switch label when the default is to do nothing).
3182
3183 15. Added some code to make it possible, when PCRE is compiled as a C++
3184 library, to replace subject pointers for pcre_exec() with a smart pointer
3185 class, thus making it possible to process discontinuous strings.
3186
3187 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3188 much the same function. They were added by different people who were trying
3189 to make PCRE easy to compile on non-Unix systems. It has been suggested
3190 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3191 for compiling on Windows systems. I have therefore replaced it with
3192 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3193 defaults to "extern" for C or "extern C" for C++, which works fine on
3194 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3195 SCOPE with something explicit in config.h. In addition:
3196
3197 (a) pcreposix.h still had just "extern" instead of either of these macros;
3198 I have replaced it with PCRE_DATA_SCOPE.
3199
3200 (b) Functions such as _pcre_xclass(), which are internal to the library,
3201 but external in the C sense, all had PCRE_EXPORT in their definitions.
3202 This is apparently wrong for the Windows case, so I have removed it.
3203 (It makes no difference on Unix-like systems.)
3204
3205 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3206 of recursive calls to match(). This is different to MATCH_LIMIT because
3207 that limits the total number of calls to match(), not all of which increase
3208 the depth of recursion. Limiting the recursion depth limits the amount of
3209 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3210 when PCRE is compiled, and changed at run time. A patch from Google adds
3211 this functionality to the C++ interface.
3212
3213 18. Changes to the handling of Unicode character properties:
3214
3215 (a) Updated the table to Unicode 4.1.0.
3216
3217 (b) Recognize characters that are not in the table as "Cn" (undefined).
3218
3219 (c) I revised the way the table is implemented to a much improved format
3220 which includes recognition of ranges. It now supports the ranges that
3221 are defined in UnicodeData.txt, and it also amalgamates other
3222 characters into ranges. This has reduced the number of entries in the
3223 table from around 16,000 to around 3,000, thus reducing its size
3224 considerably. I realized I did not need to use a tree structure after
3225 all - a binary chop search is just as efficient. Having reduced the
3226 number of entries, I extended their size from 6 bytes to 8 bytes to
3227 allow for more data.
3228
3229 (d) Added support for Unicode script names via properties such as \p{Han}.
3230
3231 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3232 matching that character.
3233
3234 20. When matching a repeated Unicode property with a minimum greater than zero,
3235 (for example \pL{2,}), PCRE could look past the end of the subject if it
3236 reached it while seeking the minimum number of characters. This could
3237 happen only if some of the characters were more than one byte long, because
3238 there is a check for at least the minimum number of bytes.
3239
3240 21. Refactored the implementation of \p and \P so as to be more general, to
3241 allow for more different types of property in future. This has changed the
3242 compiled form incompatibly. Anybody with saved compiled patterns that use
3243 \p or \P will have to recompile them.
3244
3245 22. Added "Any" and "L&" to the supported property types.
3246
3247 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3248 but give a compile time error if the value is greater than 0xff.
3249
3250 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3251 accidentally not being installed or uninstalled.
3252
3253 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3254 made were to insert the current release number. This seemed silly, because
3255 it made things harder for people building PCRE on systems that don't run
3256 "configure". I have turned pcre.h into a distributed file, no longer built
3257 by "configure", with the version identification directly included. There is
3258 no longer a pcre.h.in file.
3259
3260 However, this change necessitated a change to the pcre-config script as
3261 well. It is built from pcre-config.in, and one of the substitutions was the
3262 release number. I have updated configure.ac so that ./configure now finds
3263 the release number by grepping pcre.h.
3264
3265 26. Added the ability to run the tests under valgrind.
3266
3267
3268 Version 6.4 05-Sep-05
3269 ---------------------
3270
3271 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3272 "--" to be printed when multiple files were scanned, even when none of the
3273 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3274 consider it to be a bug, and have restored the previous behaviour.
3275
3276 2. A couple of code tidies to get rid of compiler warnings.
3277
3278 3. The pcretest program used to cheat by referring to symbols in the library
3279 whose names begin with _pcre_. These are internal symbols that are not
3280 really supposed to be visible externally, and in some environments it is
3281 possible to suppress them. The cheating is now confined to including
3282 certain files from the library's source, which is a bit cleaner.
3283
3284 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3285 file's purpose clearer.
3286
3287 5. Reorganized pcre_ucp_findchar().
3288
3289
3290 Version 6.3 15-Aug-05
3291 ---------------------
3292
3293 1. The file libpcre.pc.in did not have general read permission in the tarball.
3294
3295 2. There were some problems when building without C++ support:
3296
3297 (a) If C++ support was not built, "make install" and "make test" still
3298 tried to test it.
3299
3300 (b) There were problems when the value of CXX was explicitly set. Some
3301 changes have been made to try to fix these, and ...
3302
3303 (c) --disable-cpp can now be used to explicitly disable C++ support.
3304
3305 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3306 backslash in a target when C++ was disabled. This confuses some
3307 versions of "make", apparently. Using an intermediate variable solves
3308 this. (Same for CPP_LOBJ.)
3309
3310 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3311 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3312 necessary on certain architectures.
3313
3314 4. Added a setting of -export-symbols-regex to the link command to remove
3315 those symbols that are exported in the C sense, but actually are local
3316 within the library, and not documented. Their names all begin with
3317 "_pcre_". This is not a perfect job, because (a) we have to except some
3318 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3319 available (and never for static libraries). I have made a note to try to
3320 find a way round (a) in the future.
3321
3322
3323 Version 6.2 01-Aug-05
3324 ---------------------
3325
3326 1. There was no test for integer overflow of quantifier values. A construction
3327 such as {1111111111111111} would give undefined results. What is worse, if
3328 a minimum quantifier for a parenthesized subpattern overflowed and became
3329 negative, the calculation of the memory size went wrong. This could have
3330 led to memory overwriting.
3331
3332 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3333
3334 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3335 operating environments where this matters.
3336
3337 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3338 PCRE options from within the C++ wrapper.
3339
3340 5. Named capturing subpatterns were not being correctly counted when a pattern
3341 was compiled. This caused two problems: (a) If there were more than 100
3342 such subpatterns, the calculation of the memory needed for the whole
3343 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3344 back references of the form \12, where the number was greater than 9, were
3345 not recognized as back references, even though there were sufficient
3346 previous subpatterns.
3347
3348 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3349 versions of gcc, e.g. 2.95.4.
3350
3351
3352 Version 6.1 21-Jun-05
3353 ---------------------
3354
3355 1. There was one reference to the variable "posix" in pcretest.c that was not
3356 surrounded by "#if !defined NOPOSIX".
3357
3358 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3359 the cross-check on the old pcre_info() function, for the benefit of the
3360 cut-down version of PCRE that is currently imported into Exim.
3361
3362 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3363 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3364 patterns that start (?i) but I don't think that matters. The use of (?i) is
3365 just an example; this all applies to the other options as well.
3366
3367 4. Since libtool seems to echo the compile commands it is issuing, the output
3368 from "make" can be reduced a bit by putting "@" in front of each libtool
3369 compile command.
3370
3371 5. Patch from the folks at Google for configure.in to be a bit more thorough
3372 in checking for a suitable C++ installation before trying to compile the
3373 C++ stuff. This should fix a reported problem when a compiler was present,
3374 but no suitable headers.
3375
3376 6. The man pages all had just "PCRE" as their title. I have changed them to
3377 be the relevant file name. I have also arranged that these names are
3378 retained in the file doc/pcre.txt, which is a concatenation in text format
3379 of all the man pages except the little individual ones for each function.
3380
3381 7. The NON-UNIX-USE file had not been updated for the different set of source
3382 files that come with release 6. I also added a few comments about the C++
3383 wrapper.
3384
3385
3386 Version 6.0 07-Jun-05
3387 ---------------------
3388
3389 1. Some minor internal re-organization to help with my DFA experiments.
3390
3391 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3392 didn't matter for the library itself when fully configured, but did matter
3393 when compiling without UCP support, or within Exim, where the ucp files are
3394 not imported.
3395
3396 3. Refactoring of the library code to split up the various functions into
3397 different source modules. The addition of the new DFA matching code (see
3398 below) to a single monolithic source would have made it really too
3399 unwieldy, quite apart from causing all the code to be include in a
3400 statically linked application, when only some functions are used. This is
3401 relevant even without the DFA addition now that patterns can be compiled in
3402 one application and matched in another.
3403
3404 The downside of splitting up is that there have to be some external
3405 functions and data tables that are used internally in different modules of
3406 the library but which are not part of the API. These have all had their
3407 names changed to start with "_pcre_" so that they are unlikely to clash
3408 with other external names.
3409
3410 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3411 a different (DFA) algorithm. Although it is slower than the original
3412 function, it does have some advantages for certain types of matching
3413 problem.
3414
3415 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3416 including restarting after a partial match.
3417
3418 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3419 defined when compiling for Windows was sent to me. I have put it into the
3420 code, though I have no means of testing or verifying it.
3421
3422 7. Added the pcre_refcount() auxiliary function.
3423
3424 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3425 match before or at the first newline in the subject string. In pcretest,
3426 the /f option on a pattern can be used to set this.
3427
3428 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3429 would behave wrongly. This has been present in PCRE since release 4.0.
3430
3431 10. A number of changes to the pcregrep command:
3432
3433 (a) Refactored how -x works; insert ^(...)$ instead of setting
3434 PCRE_ANCHORED and checking the length, in preparation for adding
3435 something similar for -w.
3436
3437 (b) Added the -w (match as a word) option.
3438
3439 (c) Refactored the way lines are read and buffered so as to have more
3440 than one at a time available.
3441
3442 (d) Implemented a pcregrep test script.
3443
3444 (e) Added the -M (multiline match) option. This allows patterns to match
3445 over several lines of the subject. The buffering ensures that at least
3446 8K, or the rest of the document (whichever is the shorter) is available
3447 for matching (and similarly the previous 8K for lookbehind assertions).
3448
3449 (f) Changed the --help output so that it now says
3450
3451 -w, --word-regex(p)
3452
3453 instead of two lines, one with "regex" and the other with "regexp"
3454 because that confused at least one person since the short forms are the
3455 same. (This required a bit of code, as the output is generated
3456 automatically from a table. It wasn't just a text change.)
3457
3458 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3459 option but starts with a hyphen. Could be a pattern or a path name
3460 starting with a hyphen, for instance.
3461
3462 (h) "-" can be given as a file name to represent stdin.
3463
3464 (i) When file names are being printed, "(standard input)" is used for
3465 the standard input, for compatibility with GNU grep. Previously
3466 "<stdin>" was used.
3467
3468 (j) The option --label=xxx can be used to supply a name to be used for
3469 stdin when file names are being printed. There is no short form.
3470
3471 (k) Re-factored the options decoding logic because we are going to add
3472 two more options that take data. Such options can now be given in four
3473 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3474
3475 (l) Added the -A, -B, and -C options for requesting that lines of context
3476 around matches be printed.
3477
3478 (m) Added the -L option to print the names of files that do not contain
3479 any matching lines, that is, the complement of -l.
3480
3481 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3482 continue to scan other files.
3483
3484 (o) The -s option was incorrectly implemented. For compatibility with other
3485 greps, it now suppresses the error message for a non-existent or non-
3486 accessible file (but not the return code). There is a new option called
3487 -q that suppresses the output of matching lines, which was what -s was
3488 previously doing.
3489
3490 (p) Added --include and --exclude options to specify files for inclusion
3491 and exclusion when recursing.
3492
3493 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3494 Hopefully, it now does.
3495
3496 12. Missing cast in pcre_study().
3497
3498 13. Added an "uninstall" target to the makefile.
3499
3500 14. Replaced "extern" in the function prototypes in Makefile.in with
3501 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3502 world, but is set differently for Windows.
3503
3504 15. Added a second compiling function called pcre_compile2(). The only
3505 difference is that it has an extra argument, which is a pointer to an
3506 integer error code. When there is a compile-time failure, this is set
3507 non-zero, in addition to the error test pointer being set to point to an
3508 error message. The new argument may be NULL if no error number is required
3509 (but then you may as well call pcre_compile(), which is now just a
3510 wrapper). This facility is provided because some applications need a
3511 numeric error indication, but it has also enabled me to tidy up the way
3512 compile-time errors are handled in the POSIX wrapper.
3513
3514 16. Added VPATH=.libs to the makefile; this should help when building with one
3515 prefix path and installing with another. (Or so I'm told by someone who
3516 knows more about this stuff than I do.)
3517
3518 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3519 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3520 match everything, including newlines. This is not POSIX-compatible, but
3521 somebody wanted the feature. From pcretest it can be activated by using
3522 both the P and the s flags.
3523
3524 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3525
3526 19. libpcre.pc was being incorrectly installed as executable.
3527
3528 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3529 it now also looks for '\r' so that it will work unmodified on Windows.
3530
3531 21. Added Google's contributed C++ wrapper to the distribution.
3532
3533 22. Added some untidy missing memory free() calls in pcretest, to keep
3534 Electric Fence happy when testing.
3535
3536
3537
3538 Version 5.0 13-Sep-04
3539 ---------------------
3540
3541 1. Internal change: literal characters are no longer packed up into items
3542 containing multiple characters in a single byte-string. Each character
3543 is now matched using a separate opcode. However, there may be more than one
3544 byte in the character in UTF-8 mode.
3545
3546 2. The pcre_callout_block structure has two new fields: pattern_position and
3547 next_item_length. These contain the offset in the pattern to the next match
3548 item, and its length, respectively.
3549
3550 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3551 insertion of callouts before each pattern item. Added the /C option to
3552 pcretest to make use of this.
3553
3554 4. On the advice of a Windows user, the lines
3555
3556 #if defined(_WIN32) || defined(WIN32)
3557 _setmode( _fileno( stdout ), 0x8000 );
3558 #endif /* defined(_WIN32) || defined(WIN32) */
3559
3560 have been added to the source of pcretest. This apparently does useful
3561 magic in relation to line terminators.
3562
3563 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3564 for the benefit of those environments where the "b" makes a difference.
3565
3566 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3567 to know about it. I have put a hack into configure.in that adds in code
3568 to set GCC=yes if CC=icc. This seems to end up at a point in the
3569 generated configure script that is early enough to affect the setting of
3570 compiler options, which is what is needed, but I have no means of testing
3571 whether it really works. (The user who reported this had patched the
3572 generated configure script, which of course I cannot do.)
3573
3574 LATER: After change 22 below (new libtool files), the configure script
3575 seems to know about icc (and also ecc). Therefore, I have commented out
3576 this hack in configure.in.
3577
3578 7. Added support for pkg-config (2 patches were sent in).
3579
3580 8. Negated POSIX character classes that used a combination of internal tables
3581 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3582 [[:^ascii]]. Typically, they would match almost any characters. The other
3583 POSIX classes were not broken in this way.
3584
3585 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3586 to find the match, as PCRE was deluded into thinking that the match had to
3587 start at the start point or following a newline. The same bug applied to
3588 patterns with negative forward assertions or any backward assertions
3589 preceding ".*" at the start, unless the pattern required a fixed first
3590 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3591
3592 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3593 starting at the last subject character, bytes beyond the end of the subject
3594 string were read.
3595
3596 11. Renamed the variable "class" as "classbits" to make life easier for C++
3597 users. (Previously there was a macro definition, but it apparently wasn't
3598 enough.)
3599
3600 12. Added the new field "tables" to the extra data so that tables can be passed
3601 in at exec time, or the internal tables can be re-selected. This allows
3602 a compiled regex to be saved and re-used at a later time by a different
3603 program that might have everything at different addresses.
3604
3605 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3606 -R library as well as a -L library.
3607
3608 14. The debugging options of pcretest (-d on the command line or D on a
3609 pattern) showed incorrect output for anything following an extended class
3610 that contained multibyte characters and which was followed by a quantifier.
3611
3612 15. Added optional support for general category Unicode character properties
3613 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3614 support. It adds about 90K to the size of the library. The meanings of the
3615 inbuilt class escapes such as \d and \s have NOT been changed.
3616
3617 16. Updated pcredemo.c to include calls to free() to release the memory for the
3618 compiled pattern.
3619
3620 17. The generated file chartables.c was being created in the source directory
3621 instead of in the building directory. This caused the build to fail if the
3622 source directory was different from the building directory, and was
3623 read-only.
3624
3625 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3626 file. No doubt somebody will tell me if they don't make sense... Also added
3627 Dan Mooney's comments about building on OpenVMS.
3628
3629 19. Added support for partial matching via the PCRE_PARTIAL option for
3630 pcre_exec() and the \P data escape in pcretest.
3631
3632 20. Extended pcretest with 3 new pattern features:
3633
3634 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3635 write the compiled pattern to the file whose name is "rest-of-line".
3636 This is a straight binary dump of the data, with the saved pointer to
3637 the character tables forced to be NULL. The study data, if any, is
3638 written too. After writing, pcretest reads a new pattern.
3639
3640 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3641 compiled pattern from the given file. There must not be any
3642 occurrences of "<" in the file name (pretty unlikely); if there are,
3643 pcretest will instead treat the initial "<" as a pattern delimiter.
3644 After reading in the pattern, pcretest goes on to read data lines as
3645 usual.
3646
3647 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3648 and 16-bit fields in a compiled pattern, to simulate a pattern that
3649 was compiled on a host of opposite endianness.
3650
3651 21. The pcre-exec() function can now cope with patterns that were compiled on
3652 hosts of opposite endianness, with this restriction:
3653
3654 As for any compiled expression that is saved and used later, the tables
3655 pointer field cannot be preserved; the extra_data field in the arguments
3656 to pcre_exec() should be used to pass in a tables address if a value
3657 other than the default internal tables were used at compile time.
3658
3659 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3660 now diagnosed as an error. Previously, most of the time, a negative number
3661 would have been treated as zero, but if in addition "ovector" was passed as
3662 NULL, a crash could occur.
3663
3664 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3665 new versions from the libtool 1.5 distribution (the last one is a copy of
3666 a file called libtool.m4). This seems to have fixed the need to patch
3667 "configure" to support Darwin 1.3 (which I used to do). However, I still
3668 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3669 workstation).
3670
3671 24. Changed the PCRE licence to be the more standard "BSD" licence.
3672
3673
3674 Version 4.5 01-Dec-03
3675 ---------------------
3676
3677 1. There has been some re-arrangement of the code for the match() function so
3678 that it can be compiled in a version that does not call itself recursively.
3679 Instead, it keeps those local variables that need separate instances for
3680 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3681 needs to "recurse". Keeping track of where control must go is done by means
3682 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3683 hide most of the details from the main code, and operates only if
3684 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3685 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3686 operating.
3687
3688 To make it easier for callers to provide specially tailored get/free
3689 functions for this usage, two new functions, pcre_stack_malloc, and
3690 pcre_stack_free, are used. They are always called in strict stacking order,
3691 and the size of block requested is always the same.
3692
3693 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3694 PCRE has been compiled to use the stack or the heap for recursion. The
3695 -C option of pcretest uses this to show which version is compiled.
3696
3697 A new data escape \S, is added to pcretest; it causes the amounts of store
3698 obtained and freed by both kinds of malloc/free at match time to be added
3699 to the output.
3700
3701 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3702 what's available on my current Linux desktop machine.
3703
3704 3. When matching a UTF-8 string, the test for a valid string at the start has
3705 been extended. If start_offset is not zero, PCRE now checks that it points
3706 to a byte that is the start of a UTF-8 character. If not, it returns
3707 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3708 this is necessary because there may be backward assertions in the pattern.
3709 When matching the same subject several times, it may save resources to use
3710 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3711
3712 4. The code for checking the validity of UTF-8 strings has been tightened so
3713 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3714 containing "overlong sequences".
3715
3716 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3717 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3718 should have been "&", but it just so happened that all the cases this let
3719 through by mistake were picked up later in the function.
3720
3721 6. I had used a variable called "isblank" - this is a C99 function, causing
3722 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3723
3724 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3725 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3726 so the version is shown; (c) stop "make test" echoing "./RunTest".
3727
3728 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3729
3730 9. The prototype for memmove() for systems that don't have it was using
3731 size_t, but the inclusion of the header that defines size_t was later. I've
3732 moved the #includes for the C headers earlier to avoid this.
3733
3734 10. Added some adjustments to the code to make it easier to compiler on certain
3735 special systems:
3736
3737 (a) Some "const" qualifiers were missing.
3738 (b) Added the macro EXPORT before all exported functions; by default this
3739 is defined to be empty.
3740 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3741 that it reads its output file name as an argument instead of writing
3742 to the standard output and assuming this can be redirected.
3743
3744 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3745 class containing characters with values greater than 255, PCRE compilation
3746 went into a loop.
3747
3748 12. A recursive reference to a subpattern that was within another subpattern
3749 that had a minimum quantifier of zero caused PCRE to crash. For example,
3750 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3751 recursion. If the recursively-called subpattern itself had a zero repeat,
3752 that was OK.
3753
3754 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3755 buffer into which it was copied (for escape processing) was still set at
3756 1024, so long lines caused crashes.
3757
3758 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3759 "internal error: code overflow...". This applied to any character class
3760 that was followed by a possessive quantifier.
3761
3762 15. Modified the Makefile to add libpcre.la as a prerequisite for
3763 libpcreposix.la because I was told this is needed for a parallel build to
3764 work.
3765
3766 16. If a pattern that contained .* following optional items at the start was
3767 studied, the wrong optimizing data was generated, leading to matching
3768 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3769 matching string must start with a or b or c. The correct conclusion for
3770 this pattern is that a match can start with any character.
3771
3772
3773 Version 4.4 13-Aug-03
3774 ---------------------
3775
3776 1. In UTF-8 mode, a character class containing characters with values between
3777 127 and 255 was not handled correctly if the compiled pattern was studied.
3778 In fixing this, I have also improved the studying algorithm for such
3779 classes (slightly).
3780
3781 2. Three internal functions had redundant arguments passed to them. Removal
3782 might give a very teeny performance improvement.
3783
3784 3. Documentation bug: the value of the capture_top field in a callout is *one
3785 more than* the number of the hightest numbered captured substring.
3786
3787 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3788 in incorrectly linking with a previously installed version. They now link
3789 explicitly with libpcre.la.
3790
3791 5. configure.in no longer needs to recognize Cygwin specially.
3792
3793 6. A problem in pcre.in for Windows platforms is fixed.
3794
3795 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3796 pcretest, it used to include the size of the study block as part of its
3797 output. Unfortunately, the structure contains a field that has a different
3798 size on different hardware architectures. This meant that the tests that
3799 showed this size failed. As the block is currently always of a fixed size,
3800 this information isn't actually particularly useful in pcretest output, so
3801 I have just removed it.
3802
3803 8. Three pre-processor statements accidentally did not start in column 1.
3804 Sadly, there are *still* compilers around that complain, even though
3805 standard C has not required this for well over a decade. Sigh.
3806
3807 9. In pcretest, the code for checking callouts passed small integers in the
3808 callout_data field, which is a void * field. However, some picky compilers
3809 complained about the casts involved for this on 64-bit systems. Now
3810 pcretest passes the address of the small integer instead, which should get
3811 rid of the warnings.
3812
3813 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3814 both compile and run time, and gives an error if an invalid UTF-8 sequence
3815 is found. There is a option for disabling this check in cases where the
3816 string is known to be correct and/or the maximum performance is wanted.
3817
3818 11. In response to a bug report, I changed one line in Makefile.in from
3819
3820 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3821 to
3822 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3823
3824 to look similar to other lines, but I have no way of telling whether this
3825 is the right thing to do, as I do not use Windows. No doubt I'll get told
3826 if it's wrong...
3827
3828
3829 Version 4.3 21-May-03
3830 ---------------------
3831
3832 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3833 Makefile.
3834
3835 2. Some refactoring to improve the quality of the code:
3836
3837 (i) The utf8_table... variables are now declared "const".
3838
3839 (ii) The code for \cx, which used the "case flipping" table to upper case
3840 lower case letters, now just substracts 32. This is ASCII-specific,
3841 but the whole concept of \cx is ASCII-specific, so it seems
3842 reasonable.
3843
3844 (iii) PCRE was using its character types table to recognize decimal and
3845 hexadecimal digits in the pattern. This is silly, because it handles
3846 only 0-9, a-f, and A-F, but the character types table is locale-
3847 specific, which means strange things might happen. A private
3848 table is now used for this - though it costs 256 bytes, a table is
3849 much faster than multiple explicit tests. Of course, the standard
3850 character types table is still used for matching digits in subject
3851 strings against \d.
3852
3853 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3854 ending in _t are). So I've renamed it as ESC_tee.
3855
3856 3. The first argument for regexec() in the POSIX wrapper should have been
3857 defined as "const".
3858
3859 4. Changed pcretest to use malloc() for its buffers so that they can be
3860 Electric Fenced for debugging.
3861
3862 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3863 to read one or more bytes before the start of the subject string. Often this
3864 had no effect on PCRE's behaviour, but in some circumstances it could
3865 provoke a segmentation fault.
3866
3867 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3868 to try to read one or more bytes before the start of the subject string.
3869
3870 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3871 UTF-8 support could misbehave in various ways if the subject string
3872 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3873 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3874 back over UTF-8 characters.)
3875
3876
3877 Version 4.2 14-Apr-03
3878 ---------------------
3879
3880 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3881
3882 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3883 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3884 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3885 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3886 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3887 and BUILD_EXEEXT
3888 Note: automatic setting of the BUILD variables is not yet working
3889 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3890 compile-time but not at link-time
3891 [LINK]: use for linking executables only
3892 make different versions for Windows and non-Windows
3893 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3894 libraries
3895 [LINK_FOR_BUILD]: new variable
3896 [OBJEXT]: use throughout
3897 [EXEEXT]: use throughout
3898 <winshared>: new target
3899 <wininstall>: new target
3900 <dftables.o>: use native compiler
3901 <dftables>: use native linker
3902 <install>: handle Windows platform correctly
3903 <clean>: ditto
3904 <check>: ditto
3905 copy DLL to top builddir before testing
3906
3907 As part of these changes, -no-undefined was removed again. This was reported
3908 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3909 in any case.
3910
3911 3. Some tidies to get rid of compiler warnings:
3912
3913 . In the match_data structure, match_limit was an unsigned long int, whereas
3914 match_call_count was an int. I've made them both unsigned long ints.
3915
3916 . In pcretest the fact that a const uschar * doesn't automatically cast to
3917 a void * provoked a warning.
3918
3919 . Turning on some more compiler warnings threw up some "shadow" variables
3920 and a few more missing casts.
3921
3922 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3923 option, a class that contained a single character with a value between 128
3924 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3925
3926 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3927 option, a class that contained several characters, but with at least one
3928 whose value was between 128 and 255 caused PCRE to crash.
3929
3930
3931 Version 4.1 12-Mar-03
3932 ---------------------
3933
3934 1. Compiling with gcc -pedantic found a couple of places where casts were
3935 needed, and a string in dftables.c that was longer than standard compilers are
3936 required to support.
3937
3938 2. Compiling with Sun's compiler found a few more places where the code could
3939 be tidied up in order to avoid warnings.
3940
3941 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3942 first of these names is deprecated in the latest Autoconf in favour of the name
3943 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3944 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3945 analogy I have changed it to CFLAGS_FOR_BUILD.
3946
3947 4. Added -no-undefined to the linking command in the Makefile, because this is
3948 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3949 linking step for the pcreposix library.
3950
3951 5. PCRE was failing to diagnose the case of two named groups with the same
3952 name.
3953
3954 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3955 literal character that is needed in the subject for a match, and scans along to
3956 ensure that it is present before embarking on the full matching process. This
3957 saves time in cases of nested unlimited repeats that are never going to match.
3958 Problem: the scan can take a lot of time if the subject is very long (e.g.
3959 megabytes), thus penalizing straightforward matches. It is now done only if the
3960 amount of subject to be scanned is less than 1000 bytes.
3961
3962 7. A lesser problem with the same optimization is that it was recording the
3963 first character of an anchored pattern as "needed", thus provoking a search
3964 right along the subject, even when the first match of the pattern was going to
3965 fail. The "needed" character is now not set for anchored patterns, unless it
3966 follows something in the pattern that is of non-fixed length. Thus, it still
3967 fulfils its original purpose of finding quick non-matches in cases of nested
3968 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3969
3970
3971 Version 4.0 17-Feb-03
3972 ---------------------
3973
3974 1. If a comment in an extended regex that started immediately after a meta-item
3975 extended to the end of string, PCRE compiled incorrect data. This could lead to
3976 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3977
3978 2. Moved to autoconf 2.53 and libtool 1.4.2.
3979
3980 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3981 the special perltest8 script is no longer needed - all the tests can be run
3982 from a single perltest script.
3983
3984 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3985 by \s. It has now been removed in PCRE. This means it isn't recognized as
3986 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3987 class [:space:] *does* include VT, thereby creating a mess.
3988
3989 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3990 space and tab.
3991
3992 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3993 its new features into the main test script, reducing the number of scripts.
3994
3995 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3996 were backward compatible, and made the (?i) apply to the whole pattern, as if
3997 /i were given. Now it behaves more logically, and applies the option setting
3998 only to what follows. PCRE has been changed to follow suit. However, if it
3999 finds options settings right at the start of the pattern, it extracts them into
4000 the global options, as before. Thus, they show up in the info data.
4001
4002 8. Added support for the \Q...\E escape sequence. Characters in between are
4003 treated as literals. This is slightly different from Perl in that $ and @ are
4004 also handled as literals inside the quotes. In Perl, they will cause variable
4005 interpolation. Note the following examples:
4006
4007 Pattern PCRE matches Perl matches
4008
4009 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4010 \Qabc\$xyz\E abc\$xyz abc\$xyz
4011 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4012
4013 For compatibility with Perl, \Q...\E sequences are recognized inside character
4014 classes as well as outside them.
4015
4016 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4017 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4018 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4019 signed/unsigned warnings.
4020
4021 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4022 option for pcretest, so I've replaced it by a simple function that does just
4023 that job.
4024
4025 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4026 "pcregrep -".
4027
4028 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4029 Java package. This provides some syntactic sugar for simple cases of what my
4030 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4031 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4032 item, you can use this simplified notation. Note that only makes sense with
4033 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4034 greediness, whatever the setting of the PCRE_UNGREEDY option.
4035
4036 13. A change of greediness default within a pattern was not taking effect at
4037 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4038 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4039 was abstracted outside.
4040
4041 14. PCRE now supports the \G assertion. It is true when the current matching
4042 position is at the start point of the match. This differs from \A when the
4043 starting offset is non-zero. Used with the /g option of pcretest (or similar
4044 code), it works in the same way as it does for Perl's /g option. If all
4045 alternatives of a regex begin with \G, the expression is anchored to the start
4046 match position, and the "anchored" flag is set in the compiled expression.
4047
4048 15. Some bugs concerning the handling of certain option changes within patterns
4049 have been fixed. These applied to options other than (?ims). For example,
4050 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4051 been the other way round. Some of this was related to change 7 above.
4052
4053 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4054 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4055 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4056 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4057
4058 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4059 mode. Unlike ".", it always matches newline, whatever the setting of
4060 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4061 assertions. Perl allows it, but it doesn't (in general) work because it can't
4062 calculate the length of the lookbehind. At least, that's the case for Perl
4063 5.8.0 - I've been told they are going to document that it doesn't work in
4064 future.
4065
4066 18. Added an error diagnosis for escapes that PCRE does not support: these are
4067 \L, \l, \N, \P, \p, \U, \u, and \X.
4068
4069 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4070 reading past the end of the pattern in cases such as /[abcd/.
4071
4072 20. PCRE was getting more memory than necessary for patterns with classes that
4073 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4074
4075 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4076 compiling PCRE for use with Virtual Pascal.
4077
4078 22. Small fix to the Makefile to make it work properly if the build is done
4079 outside the source tree.
4080
4081 23. Added a new extension: a condition to go with recursion. If a conditional
4082 subpattern starts with (?(R) the "true" branch is used if recursion has
4083 happened, whereas the "false" branch is used only at the top level.
4084
4085 24. When there was a very long string of literal characters (over 255 bytes
4086 without UTF support, over 250 bytes with UTF support), the computation of how
4087 much memory was required could be incorrect, leading to segfaults or other
4088 strange effects.
4089
4090 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4091 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4092 there was a subsequent back reference to those brackets. This meant that, for
4093 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4094 possible to check for precisely this case. All we can do is abandon the
4095 optimization if .* occurs inside capturing brackets when there are any back
4096 references whatsoever. (See below for a better fix that came later.)
4097
4098 26. The handling of the optimization for finding the first character of a
4099 non-anchored pattern, and for finding a character that is required later in the
4100 match were failing in some cases. This didn't break the matching; it just
4101 failed to optimize when it could. The way this is done has been re-implemented.
4102
4103 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4104
4105 28. Added a new feature that provides some of the functionality that Perl
4106 provides with (?{...}). The facility is termed a "callout". The way it is done
4107 in PCRE is for the caller to provide an optional function, by setting
4108 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4109 global variable. By default it is unset, which disables all calling out. To get
4110 the function called, the regex must include (?C) at appropriate points. This
4111 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4112 This provides a means of identifying different callout points. When PCRE
4113 reaches such a point in the regex, if pcre_callout has been set, the external
4114 function is called. It is provided with data in a structure called
4115 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4116 matching continues; if it returns a non-zero value, the match at the current
4117 point fails. However, backtracking will occur if possible. [This was changed
4118 later and other features added - see item 49 below.]
4119
4120 29. pcretest is upgraded to test the callout functionality. It provides a
4121 callout function that displays information. By default, it shows the start of
4122 the match and the current position in the text. There are some new data escapes
4123 to vary what happens:
4124
4125 \C+ in addition, show current contents of captured substrings
4126 \C- do not supply a callout function
4127 \C!n return 1 when callout number n is reached
4128 \C!n!m return 1 when callout number n is reached for the mth time
4129
4130 30. If pcregrep was called with the -l option and just a single file name, it
4131 output "<stdin>" if a match was found, instead of the file name.
4132
4133 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4134 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4135 pcre_exec(). This saves a malloc/free per call. The default value of
4136 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4137 when configuring.
4138
4139 32. The default maximum size of a compiled pattern is 64K. There have been a
4140 few cases of people hitting this limit. The code now uses macros to handle the
4141 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4142 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4143 configuring. Tests 2 and 5 work only with 2-byte links because they output
4144 debugging information about compiled patterns.
4145
4146 33. Internal code re-arrangements:
4147
4148 (a) Moved the debugging function for printing out a compiled regex into
4149 its own source file (printint.c) and used #include to pull it into
4150 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4151 separate copies.
4152
4153 (b) Defined the list of op-code names for debugging as a macro in
4154 internal.h so that it is next to the definition of the opcodes.
4155
4156 (c) Defined a table of op-code lengths for simpler skipping along compiled
4157 code. This is again a macro in internal.h so that it is next to the
4158 definition of the opcodes.
4159
4160 34. Added support for recursive calls to individual subpatterns, along the
4161 lines of Robin Houston's patch (but implemented somewhat differently).
4162
4163 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4164 allow it to read and process whole directories in Win32. This code was
4165 contributed by Lionel Fourquaux; it has not been tested by me.
4166
4167 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4168 used to name a group. Names consist of alphanumerics and underscores, and must
4169 be unique. Back references use the syntax (?P=name) and recursive calls use
4170 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4171 numbers. The function pcre_fullinfo() can be used after compilation to extract
4172 a name/number map. There are three relevant calls:
4173
4174 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4175 PCRE_INFO_NAMECOUNT yields the number of entries
4176 PCRE_INFO_NAMETABLE yields a pointer to the map.
4177
4178 The map is a vector of fixed-size entries. The size of each entry depends on
4179 the length of the longest name used. The first two bytes of each entry are the
4180 group number, most significant byte first. There follows the corresponding
4181 name, zero terminated. The names are in alphabetical order.
4182
4183 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4184 case instead of 255. Making it the same both with and without UTF-8 support
4185 means that the same test output works with both.
4186
4187 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4188 calling malloc() with a zero argument.
4189
4190 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4191 optimization. I've improved things by keeping a bitmap of backreferences with
4192 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4193 fact referenced, the optimization can be applied. It is unlikely that a
4194 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4195 the match to follow \n) will appear inside brackets with a number greater than
4196 31, but if it does, any back reference > 31 suppresses the optimization.
4197
4198 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4199 of disabling numbered capturing parentheses. Any opening parenthesis that is
4200 not followed by ? behaves as if it were followed by ?: but named parentheses
4201 can still be used for capturing (and they will acquire numbers in the usual
4202 way).
4203
4204 41. Redesigned the return codes from the match() function into yes/no/error so
4205 that errors can be passed back from deep inside the nested calls. A malloc
4206 failure while inside a recursive subpattern call now causes the
4207 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4208
4209 42. It is now possible to set a limit on the number of times the match()
4210 function is called in a call to pcre_exec(). This facility makes it possible to
4211 limit the amount of recursion and backtracking, though not in a directly
4212 obvious way, because the match() function is used in a number of different
4213 circumstances. The count starts from zero for each position in the subject
4214 string (for non-anchored patterns). The default limit is, for compatibility, a
4215 large number, namely 10 000 000. You can change this in two ways:
4216
4217 (a) When configuring PCRE before making, you can use --with-match-limit=n
4218 to set a default value for the compiled library.
4219
4220 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4221 a different value is set. See 45 below.
4222
4223 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4224
4225 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4226 of things that can be changed at compile time. The first argument specifies
4227 what is wanted and the second points to where the information is to be placed.
4228 The current list of available information is:
4229
4230 PCRE_CONFIG_UTF8
4231
4232 The output is an integer that is set to one if UTF-8 support is available;
4233 otherwise it is set to zero.
4234
4235 PCRE_CONFIG_NEWLINE
4236
4237 The output is an integer that it set to the value of the code that is used for
4238 newline. It is either LF (10) or CR (13).
4239
4240 PCRE_CONFIG_LINK_SIZE
4241
4242 The output is an integer that contains the number of bytes used for internal
4243 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4244
4245 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4246
4247 The output is an integer that contains the threshold above which the POSIX
4248 interface uses malloc() for output vectors. See item 31 above.
4249
4250 PCRE_CONFIG_MATCH_LIMIT
4251
4252 The output is an unsigned integer that contains the default limit of the number
4253 of match() calls in a pcre_exec() execution. See 42 above.
4254
4255 44. pcretest has been upgraded by the addition of the -C option. This causes it
4256 to extract all the available output from the new pcre_config() function, and to
4257 output it. The program then exits immediately.
4258
4259 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4260 order to support additional features. One way would have been to define
4261 pcre_exec2() (for example) with extra arguments, but this would not have been
4262 extensible, and would also have required all calls to the original function to
4263 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4264 is used for passing in "extra" data from pcre_study().
4265
4266 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4267 contains the following fields:
4268
4269 flags a bitmap indicating which of the following fields are set
4270 study_data opaque data from pcre_study()
4271 match_limit a way of specifying a limit on match() calls for a specific
4272 call to pcre_exec()
4273 callout_data data for callouts (see 49 below)
4274
4275 The flag bits are also defined in pcre.h, and are
4276
4277 PCRE_EXTRA_STUDY_DATA
4278 PCRE_EXTRA_MATCH_LIMIT
4279 PCRE_EXTRA_CALLOUT_DATA
4280
4281 The pcre_study() function now returns one of these new pcre_extra blocks, with
4282 the actual study data pointed to by the study_data field, and the
4283 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4284 before. That is, this change is entirely upwards-compatible and requires no
4285 change to existing code.
4286
4287 If you want to pass in additional data to pcre_exec(), you can either place it
4288 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4289 block.
4290
4291 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4292 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4293 times with different match limits, until it finds the minimum value needed for
4294 pcre_exec() to complete. The value is then output. This can be instructive; for
4295 most simple matches the number is quite small, but for pathological cases it
4296 gets very large very quickly.
4297
4298 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4299 returns the size of the data block pointed to by the study_data field in a
4300 pcre_extra block, that is, the value that was passed as the argument to
4301 pcre_malloc() when PCRE was getting memory in which to place the information
4302 created by pcre_study(). The fourth argument should point to a size_t variable.
4303 pcretest has been extended so that this information is shown after a successful
4304 pcre_study() call when information about the compiled regex is being displayed.
4305
4306 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4307 because what follows is always an absolute path. (Later: it turns out that this
4308 is more than cosmetic for MinGW, because it doesn't like empty path
4309 components.)
4310
4311 49. Some changes have been made to the callout feature (see 28 above):
4312
4313 (i) A callout function now has three choices for what it returns:
4314
4315 0 => success, carry on matching
4316 > 0 => failure at this point, but backtrack if possible
4317 < 0 => serious error, return this value from pcre_exec()
4318
4319 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4320 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4321 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4322 use by callout functions. It will never be used by PCRE itself.
4323
4324 (ii) The pcre_extra structure (see 45 above) has a void * field called
4325 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4326 pcre_callout_block structure has a field of the same name. The contents of
4327 the field passed in the pcre_extra structure are passed to the callout
4328 function in the corresponding field in the callout block. This makes it
4329 easier to use the same callout-containing regex from multiple threads. For
4330 testing, the pcretest program has a new data escape
4331
4332 \C*n pass the number n (may be negative) as callout_data
4333
4334 If the callout function in pcretest receives a non-zero value as
4335 callout_data, it returns that value.
4336
4337 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4338 there were some redundant $(CFLAGS) in commands that are now specified as
4339 $(LINK), which already includes $(CFLAGS).
4340
4341 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4342 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4343 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4344 one-byte characters throughout. Note that case-insensitive matching applies
4345 only to characters whose values are less than 256. PCRE doesn't support the
4346 notion of cases for higher-valued characters.
4347
4348 (i) A character class whose characters are all within 0-255 is handled as
4349 a bit map, and the map is inverted for negative classes. Previously, a
4350 character > 255 always failed to match such a class; however it should
4351 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4352
4353 (ii) A negated character class with a single character < 255 is coded as
4354 "not this character" (OP_NOT). This wasn't working properly when the test
4355 character was multibyte, either singly or repeated.
4356
4357 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4358 mode, for example: \x{100}{2,3}.
4359
4360 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4361 singly or repeated) now correctly test multibyte characters. However,
4362 PCRE doesn't recognize any characters with values greater than 255 as
4363 digits, spaces, or word characters. Such characters always match \D, \S,
4364 and \W, and never match \d, \s, or \w.
4365
4366 (v) Classes may now contain characters and character ranges with values
4367 greater than 255. For example: [ab\x{100}-\x{400}].
4368
4369 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4370 PCRE in UTF-8 mode.
4371
4372 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4373 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4374 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4375 value.)
4376
4377 53. The single man page has become too large. I have therefore split it up into
4378 a number of separate man pages. These also give rise to individual HTML pages;
4379 these are now put in a separate directory, and there is an index.html page that
4380 lists them all. Some hyperlinking between the pages has been installed.
4381
4382 54. Added convenience functions for handling named capturing parentheses.
4383
4384 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4385 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4386 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4387 are faulted.
4388
4389 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4390 calling configure. These values are used when compiling the dftables.c program
4391 which is run to generate the source of the default character tables. They
4392 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4393 you will need to set these values.
4394
4395 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4396
4397
4398 Version 3.9 02-Jan-02
4399 ---------------------
4400
4401 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4402
4403 2. If --disable-static was given, the building process failed when trying to
4404 build pcretest and pcregrep. (For some reason it was using libtool to compile
4405 them, which is not right, as they aren't part of the library.)
4406
4407
4408 Version 3.8 18-Dec-01
4409 ---------------------
4410
4411 1. The experimental UTF-8 code was completely screwed up. It was packing the
4412 bytes in the wrong order. How dumb can you get?
4413
4414
4415 Version 3.7 29-Oct-01
4416 ---------------------
4417
4418 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4419 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4420 this didn't happen under Solaris 8, where I normally test things.
4421
4422 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4423 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4424 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4425 this fix an hour or so after the initial 3.7 release.)
4426
4427
4428 Version 3.6 23-Oct-01
4429 ---------------------
4430
4431 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4432 offsets passed as NULL with zero offset count.
4433
4434 2. The config.guess and config.sub files had not been updated when I moved to
4435 the latest autoconf.
4436
4437
4438 Version 3.5 15-Aug-01
4439 ---------------------
4440
4441 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4442 had been forgotten.
4443
4444 2. By using declared but undefined structures, we can avoid using "void"
4445 definitions in pcre.h while keeping the internal definitions of the structures
4446 private.
4447
4448 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4449 user point of view, this means that both static and shared libraries are built
4450 by default, but this can be individually controlled. More of the work of
4451 handling this static/shared cases is now inside libtool instead of PCRE's make
4452 file.
4453
4454 4. The pcretest utility is now installed along with pcregrep because it is
4455 useful for users (to test regexs) and by doing this, it automatically gets
4456 relinked by libtool. The documentation has been turned into a man page, so
4457 there are now .1, .txt, and .html versions in /doc.
4458
4459 5. Upgrades to pcregrep:
4460 (i) Added long-form option names like gnu grep.
4461 (ii) Added --help to list all options with an explanatory phrase.
4462 (iii) Added -r, --recursive to recurse into sub-directories.
4463 (iv) Added -f, --file to read patterns from a file.
4464
4465 6. pcre_exec() was referring to its "code" argument before testing that
4466 argument for NULL (and giving an error if it was NULL).
4467
4468 7. Upgraded Makefile.in to allow for compiling in a different directory from
4469 the source directory.
4470
4471 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4472 options bits, the pointer it was passed was to an int instead of to an unsigned
4473 long int. This mattered only on 64-bit systems.
4474
4475 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4476 generated) instead of pcre.in, which it its source. Also made the same change
4477 in several of the .c files.
4478
4479 10. A new release of gcc defines printf() as a macro, which broke pcretest
4480 because it had an ifdef in the middle of a string argument for printf(). Fixed
4481 by using separate calls to printf().
4482
4483 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4484 script, to force use of CR or LF instead of \n in the source. On non-Unix
4485 systems, the value can be set in config.h.
4486
4487 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4488 absolute limit. Changed the text of the error message to make this clear, and
4489 likewise updated the man page.
4490
4491 13. The limit of 99 on the number of capturing subpatterns has been removed.
4492 The new limit is 65535, which I hope will not be a "real" limit.
4493
4494
4495 Version 3.4 22-Aug-00
4496 ---------------------
4497
4498 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4499
4500 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4501
4502
4503 Version 3.3 01-Aug-00
4504 ---------------------
4505
4506 1. If an octal character was given, but the value was greater than \377, it
4507 was not getting masked to the least significant bits, as documented. This could
4508 lead to crashes in some systems.
4509
4510 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4511 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4512
4513 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4514 These just pass their arguments on to (pcre_free)(), but they are provided
4515 because some uses of PCRE bind it to non-C systems that can call its functions,
4516 but cannot call free() or pcre_free() directly.
4517
4518 4. Add "make test" as a synonym for "make check". Corrected some comments in
4519 the Makefile.
4520
4521 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4522 Makefile.
4523
4524 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4525 command called pgrep for grepping around the active processes.
4526
4527 7. Added the beginnings of support for UTF-8 character strings.
4528
4529 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4530 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4531 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4532 out for the ar command.)
4533
4534
4535 Version 3.2 12-May-00
4536 ---------------------
4537
4538 This is purely a bug fixing release.
4539
4540 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4541 of ZA. This was just one example of several cases that could provoke this bug,
4542 which was introduced by change 9 of version 2.00. The code for breaking
4543 infinite loops after an iteration that matches an empty string was't working
4544 correctly.
4545
4546 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4547 when matched against abbab (for example). After matching an empty string, it
4548 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4549 caused it to match further down the string than it should.
4550
4551 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4552 was there because it doesn't seem to be needed, and it causes trouble on some
4553 systems, as it is not a Standard C header. It has been removed.
4554
4555 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4556 were reported on the Macintosh. The changes were from
4557
4558 while ((c = *(++ptr)) != 0 && c != '\n');
4559 to
4560 while ((c = *(++ptr)) != 0 && c != '\n') ;
4561
4562 Totally extraordinary, but if that's what it takes...
4563
4564 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4565 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4566 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4567 assumes the way PCRE uses memmove() (always moving upwards).
4568
4569 6. PCRE is being used in one environment where strchr() is not available. There
4570 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4571 faster code anyway.
4572
4573
4574 Version 3.1 09-Feb-00
4575 ---------------------
4576
4577 The only change in this release is the fixing of some bugs in Makefile.in for
4578 the "install" target:
4579
4580 (1) It was failing to install pcreposix.h.
4581
4582 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4583
4584
4585 Version 3.0 01-Feb-00
4586 ---------------------
4587
4588 1. Add support for the /+ modifier to perltest (to output $` like it does in
4589 pcretest).
4590
4591 2. Add support for the /g modifier to perltest.
4592
4593 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4594 matches null strings.
4595
4596 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4597 pattern. Perl treats empty patterns specially - it reuses the most recent
4598 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4599 effect.
4600
4601 5. The POSIX interface was broken in that it was just handing over the POSIX
4602 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4603 required a bigger vector, with some working space on the end. This means that
4604 the POSIX wrapper now has to get and free some memory, and copy the results.
4605
4606 6. Added some simple autoconf support, placing the test data and the
4607 documentation in separate directories, re-organizing some of the
4608 information files, and making it build pcre-config (a GNU standard). Also added
4609 libtool support for building PCRE as a shared library, which is now the
4610 default.
4611
4612 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4613 09 are not valid octal constants. Single digits will be used for minor values
4614 less than 10.
4615
4616 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4617 existing programs that set these in the POSIX interface can use PCRE without
4618 modification.
4619
4620 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4621 return all that pcre_info() returns, plus additional data. The pcre_info()
4622 function is retained for compatibility, but is considered to be obsolete.
4623
4624 10. Added experimental recursion feature (?R) to handle one common case that
4625 Perl 5.6 will be able to do with (?p{...}).
4626
4627 11. Added support for POSIX character classes like [:alpha:], which Perl is
4628 adopting.
4629
4630
4631 Version 2.08 31-Aug-99
4632 ----------------------
4633
4634 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4635 trying to match at the startoffset position, but instead was moving forward to
4636 the next newline as if a previous match had failed.
4637
4638 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4639 and could get into a loop if a null string was matched other than at the start
4640 of the subject.
4641
4642 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4643 be distinguished at compile time, and for completeness also added PCRE_DATE.
4644
4645 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4646 in GnuWin32 environments.
4647
4648
4649 Version 2.07 29-Jul-99
4650 ----------------------
4651
4652 1. The documentation is now supplied in plain text form and HTML as well as in
4653 the form of man page sources.
4654
4655 2. C++ compilers don't like assigning (void *) values to other pointer types.
4656 In particular this affects malloc(). Although there is no problem in Standard
4657 C, I've put in casts to keep C++ compilers happy.
4658
4659 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4660 should be (const char *).
4661
4662 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4663 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4664 However, I haven't made this a standard facility. The documentation doesn't
4665 mention it, and the Makefile doesn't support it.
4666
4667 5. The Makefile now contains an "install" target, with editable destinations at
4668 the top of the file. The pcretest program is not installed.
4669
4670 6. pgrep -V now gives the PCRE version number and date.
4671
4672 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4673 causing the entire string to be ignored, instead of just the last character.
4674
4675 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4676 non-matching string, it can take a very, very long time, even for strings of
4677 quite modest length, because of the nested recursion. PCRE now does better in
4678 some of these cases. It does this by remembering the last required literal
4679 character in the pattern, and pre-searching the subject to ensure it is present
4680 before running the real match. In other words, it applies a heuristic to detect
4681 some types of certain failure quickly, and in the above example, if presented
4682 with a string that has no trailing " it gives "no match" very quickly.
4683
4684 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4685 other alternatives are tried instead.
4686
4687
4688 Version 2.06 09-Jun-99
4689 ----------------------
4690
4691 1. Change pcretest's output for amount of store used to show just the code
4692 space, because the remainder (the data block) varies in size between 32-bit and
4693 64-bit systems.
4694
4695 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4696 start matching at. This allows lookbehinds to work when searching for multiple
4697 occurrences in a string.
4698
4699 3. Added additional options to pcretest for testing multiple occurrences:
4700
4701 /+ outputs the rest of the string that follows a match
4702 /g loops for multiple occurrences, using the new startoffset argument
4703 /G loops for multiple occurrences by passing an incremented pointer
4704
4705 4. PCRE wasn't doing the "first character" optimization for patterns starting
4706 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4707 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4708 the letter 'x'. On long subject strings, this gives a significant speed-up.
4709
4710
4711 Version 2.05 21-Apr-99
4712 ----------------------
4713
4714 1. Changed the type of magic_number from int to long int so that it works
4715 properly on 16-bit systems.
4716
4717 2. Fixed a bug which caused patterns starting with .* not to work correctly
4718 when the subject string contained newline characters. PCRE was assuming
4719 anchoring for such patterns in all cases, which is not correct because .* will
4720 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4721 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4722 must be retried after every newline in the subject.
4723
4724
4725 Version 2.04 18-Feb-99
4726 ----------------------
4727
4728 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4729 computation of the store needed to hold the pattern was incorrect (too large).
4730 If such patterns were nested a few deep, this could multiply and become a real
4731 problem.
4732
4733 2. Added /M option to pcretest to show the memory requirement of a specific
4734 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4735
4736 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4737 compiled in such a way that the backtracking after subsequent failure was
4738 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4739 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4740
4741
4742 Version 2.03 02-Feb-99
4743 ----------------------
4744
4745 1. Fixed typo and small mistake in man page.
4746
4747 2. Added 4th condition (GPL supersedes if conflict) and created separate
4748 LICENCE file containing the conditions.
4749
4750 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4751 Perl, that is the internal \ allows the delimiter to be included in the
4752 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4753 the final delimiter, add \ to the end of the pattern (to test the error).
4754
4755 4. Added the convenience functions for extracting substrings after a successful
4756 match. Updated pcretest to make it able to test these functions.
4757
4758
4759 Version 2.02 14-Jan-99
4760 ----------------------
4761
4762 1. Initialized the working variables associated with each extraction so that
4763 their saving and restoring doesn't refer to uninitialized store.
4764
4765 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4766 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4767 fix the problem.
4768
4769 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4770 calls, and wasn't printing the correct value for compiling calls. Increased the
4771 default value of LOOPREPEAT, and the number of significant figures in the
4772 times.
4773
4774 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4775
4776 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4777 a building problem on Windows NT with a FAT file system.
4778
4779
4780 Version 2.01 21-Oct-98
4781 ----------------------
4782
4783 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4784 to character tables built by pcre_maketables() in the current locale. If NULL
4785 is passed, the default tables are used.
4786
4787
4788 Version 2.00 24-Sep-98
4789 ----------------------
4790
4791 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4792 it any more.
4793
4794 2. Allow quantification of (?>) groups, and make it work correctly.
4795
4796 3. The first character computation wasn't working for (?>) groups.
4797
4798 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4799 end of the subject) and add 5.005's \z, which really does match only at the
4800 very end of the subject.
4801
4802 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4803
4804 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4805 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4806 localized options. All options to pcre_study() were also removed.
4807
4808 7. Add other new features from 5.005:
4809
4810 $(?<= positive lookbehind
4811 $(?<! negative lookbehind
4812 (?imsx-imsx) added the unsetting capability
4813 such a setting is global if at outer level; local otherwise
4814 (?imsx-imsx:) non-capturing groups with option setting
4815 (?(cond)re|re) conditional pattern matching
4816
4817 A backreference to itself in a repeated group matches the previous
4818 captured string.
4819
4820 8. General tidying up of studying (both automatic and via "study")
4821 consequential on the addition of new assertions.
4822
4823 9. As in 5.005, unlimited repeated groups that could match an empty substring
4824 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4825 runtime if any iteration does actually match an empty substring.
4826
4827 10. Include the RunTest script in the distribution.
4828
4829 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4830 discrepancies, some of which were old and were also with respect to 5.004. They
4831 have now been fixed.
4832
4833
4834 Version 1.09 28-Apr-98
4835 ----------------------
4836
4837 1. A negated single character class followed by a quantifier with a minimum
4838 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4839 program crashes, or just wrong answers. This did not apply to negated classes
4840 containing more than one character, or to minima other than one.
4841
4842
4843 Version 1.08 27-Mar-98
4844 ----------------------
4845
4846 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4847
4848 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4849 latter must appear before anything that relies on it in the pattern.
4850
4851
4852 Version 1.07 16-Feb-98
4853 ----------------------
4854
4855 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4856 repeat of a potentially empty string).
4857
4858
4859 Version 1.06 23-Jan-98
4860 ----------------------
4861
4862 1. Added Markus Oberhumer's little patches for C++.
4863
4864 2. Literal strings longer than 255 characters were broken.
4865
4866
4867 Version 1.05 23-Dec-97
4868 ----------------------
4869
4870 1. Negated character classes containing more than one character were failing if
4871 PCRE_CASELESS was set at run time.
4872
4873
4874 Version 1.04 19-Dec-97
4875 ----------------------
4876
4877 1. Corrected the man page, where some "const" qualifiers had been omitted.
4878
4879 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4880 input syntax.
4881
4882 3. Fixed memory leak which occurred when a regex with back references was
4883 matched with an offsets vector that wasn't big enough. The temporary memory
4884 that is used in this case wasn't being freed if the match failed.
4885
4886 4. Tidied pcretest to ensure it frees memory that it gets.
4887
4888 5. Temporary memory was being obtained in the case where the passed offsets
4889 vector was exactly big enough.
4890
4891 6. Corrected definition of offsetof() from change 5 below.
4892
4893 7. I had screwed up change 6 below and broken the rules for the use of
4894 setjmp(). Now fixed.
4895
4896
4897 Version 1.03 18-Dec-97
4898 ----------------------
4899
4900 1. A erroneous regex with a missing opening parenthesis was correctly
4901 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4902 on some systems.
4903
4904 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4905 it was reported that one broken compiler failed on the former because "code" is
4906 also an independent variable.
4907
4908 3. The erroneous regex a[]b caused an array overrun reference.
4909
4910 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4911 fail on data ending with that character. (It was going on too far, and checking
4912 the next character, typically a binary zero.) This was specific to the
4913 optimized code for single-character negative classes.
4914
4915 5. Added a contributed patch from the TIN world which does the following:
4916
4917 + Add an undef for memmove, in case the the system defines a macro for it.
4918
4919 + Add a definition of offsetof(), in case there isn't one. (I don't know
4920 the reason behind this - offsetof() is part of the ANSI standard - but
4921 it does no harm).
4922
4923 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4924 most of the places where whitespace preceded '#'. I have given up and
4925 allowed the remaining 2 cases to be at the margin.
4926
4927 + Rename some variables in pcre to eliminate shadowing. This seems very
4928 pedantic, but does no harm, of course.
4929
4930 6. Moved the call to setjmp() into its own function, to get rid of warnings
4931 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4932
4933 7. Constructs such as \d{8,} were compiling into the equivalent of
4934 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4935 outcome, but in this particular case used more store than had been allocated,
4936 which caused the bug to be discovered because it threw up an internal error.
4937
4938 8. The debugging code in both pcre and pcretest for outputting the compiled
4939 form of a regex was going wrong in the case of back references followed by
4940 curly-bracketed repeats.
4941
4942
4943 Version 1.02 12-Dec-97
4944 ----------------------
4945
4946 1. Typos in pcre.3 and comments in the source fixed.
4947
4948 2. Applied a contributed patch to get rid of places where it used to remove
4949 'const' from variables, and fixed some signed/unsigned and uninitialized
4950 variable warnings.
4951
4952 3. Added the "runtest" target to Makefile.
4953
4954 4. Set default compiler flag to -O2 rather than just -O.
4955
4956
4957 Version 1.01 19-Nov-97
4958 ----------------------
4959
4960 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4961 like /([ab]*)*/, that is, for classes with more than one character in them.
4962
4963 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4964 as /((?>a*))*/ (a PCRE_EXTRA facility).
4965
4966
4967 Version 1.00 18-Nov-97
4968 ----------------------
4969
4970 1. Added compile-time macros to support systems such as SunOS4 which don't have
4971 memmove() or strerror() but have other things that can be used instead.
4972
4973 2. Arranged that "make clean" removes the executables.
4974
4975
4976 Version 0.99 27-Oct-97
4977 ----------------------
4978
4979 1. Fixed bug in code for optimizing classes with only one character. It was
4980 initializing a 32-byte map regardless, which could cause it to run off the end
4981 of the memory it had got.
4982
4983 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4984
4985
4986 Version 0.98 22-Oct-97
4987 ----------------------
4988
4989 1. Fixed bug in code for handling temporary memory usage when there are more
4990 back references than supplied space in the ovector. This could cause segfaults.
4991
4992
4993 Version 0.97 21-Oct-97
4994 ----------------------
4995
4996 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4997
4998 2. Optimized negated single characters not to use a bit map.
4999
5000 3. Brought error texts together as macro definitions; clarified some of them;
5001 fixed one that was wrong - it said "range out of order" when it meant "invalid
5002 escape sequence".
5003
5004 4. Changed some char * arguments to const char *.
5005
5006 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5007
5008 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5009 pcretest.
5010
5011
5012 Version 0.96 16-Oct-97
5013 ----------------------
5014
5015 1. Added a simple "pgrep" utility to the distribution.
5016
5017 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5018 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5019 where "ddd" means "one or more decimal digits".
5020
5021 3. Fixed serious bug. If a pattern had a back reference, but the call to
5022 pcre_exec() didn't supply a large enough ovector to record the related
5023 identifying subpattern, the match always failed. PCRE now remembers the number
5024 of the largest back reference, and gets some temporary memory in which to save
5025 the offsets during matching if necessary, in order to ensure that
5026 backreferences always work.
5027
5028 4. Increased the compatibility with Perl in a number of ways:
5029
5030 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5031 to request this handling. The option can be set at compile or exec time.
5032
5033 (b) $ matches before a terminating newline by default; an option
5034 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5035 mode). The option can be set at compile or exec time.
5036
5037 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5038 the same as Perl's. If the decimal number it represents is less than 10
5039 or there aren't that many previous left capturing parentheses, an octal
5040 escape is read. Inside a character class, it's always an octal escape,
5041 even if it is a single digit.
5042
5043 (d) An escaped but undefined alphabetic character is taken as a literal,
5044 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5045 escapes.
5046
5047 (e) {0} is now permitted. (The previous item is removed from the compiled
5048 pattern).
5049
5050 5. Changed all the names of code files so that the basic parts are no longer
5051 than 10 characters, and abolished the teeny "globals.c" file.
5052
5053 6. Changed the handling of character classes; they are now done with a 32-byte
5054 bit map always.
5055
5056 7. Added the -d and /D options to pcretest to make it possible to look at the
5057 internals of compilation without having to recompile pcre.
5058
5059
5060 Version 0.95 23-Sep-97
5061 ----------------------
5062
5063 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5064 \x20 at the start of a run of normal characters. These were being treated as
5065 real characters, instead of the source characters being re-checked.
5066
5067
5068 Version 0.94 18-Sep-97
5069 ----------------------
5070
5071 1. The functions are now thread-safe, with the caveat that the global variables
5072 containing pointers to malloc() and free() or alternative functions are the
5073 same for all threads.
5074
5075 2. Get pcre_study() to generate a bitmap of initial characters for non-
5076 anchored patterns when this is possible, and use it if passed to pcre_exec().
5077
5078
5079 Version 0.93 15-Sep-97
5080 ----------------------
5081
5082 1. /(b)|(:+)/ was computing an incorrect first character.
5083
5084 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5085 but not actually doing anything yet.
5086
5087 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5088 as Perl does (e.g. [-az] or [az-]).
5089
5090 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5091 all possible positions.
5092
5093 5. Split up into different modules to avoid including unneeded functions in a
5094 compiled binary. However, compile and exec are still in one module. The "study"
5095 function is split off.
5096
5097 6. The character tables are now in a separate module whose source is generated
5098 by an auxiliary program - but can then be edited by hand if required. There are
5099 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5100 toupper() in the code.
5101
5102 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5103 make them global. Abolish the function for setting them, as the caller can now
5104 set them directly.
5105
5106
5107 Version 0.92 11-Sep-97
5108 ----------------------
5109
5110 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5111 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5112
5113 2. Caseless matching was not working in character classes if the characters in
5114 the pattern were in upper case.
5115
5116 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5117
5118 4. Make PCRE_ANCHORED public and accept as a compile option.
5119
5120 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5121 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5122 pass them.
5123
5124 6. Give an error if bad option bits passed at compile or run time.
5125
5126 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5127 pcretest to cause it to pass that flag.
5128
5129 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5130 options, and the first character, if set.
5131
5132 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5133
5134
5135 Version 0.91 10-Sep-97
5136 ----------------------
5137
5138 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5139 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5140
5141 2. PCRE was looping on encountering an indefinitely repeated back reference to
5142 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5143 Perl does - treats the match as successful.
5144
5145 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5