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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1327 - (show annotations)
Fri May 10 16:22:40 2013 UTC (6 years, 6 months ago) by ph10
File size: 247011 byte(s)
Error occurred while calculating annotation data.
Fix pcretest crash with a data line longer than 65536 bytes.
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
162 Version 8.32 30-November-2012
163 -----------------------------
164
165 1. Improved JIT compiler optimizations for first character search and single
166 character iterators.
167
168 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
169 Patch by Daniel Richard G.
170
171 3. Single character iterator optimizations in the JIT compiler.
172
173 4. Improved JIT compiler optimizations for character ranges.
174
175 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
176 Reported by Giuseppe D'Angelo.
177
178 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
179 of a line, was being set incorrectly in cases where .* appeared inside
180 atomic brackets at the start of a pattern, or where there was a subsequent
181 *PRUNE or *SKIP.
182
183 7. Improved instruction cache flush for POWER/PowerPC.
184 Patch by Daniel Richard G.
185
186 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
187 grep:
188
189 (a) There is now no limit to the number of patterns to be matched.
190
191 (b) An error is given if a pattern is too long.
192
193 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
194 are now supported.
195
196 (d) --exclude-from and --include-from (multiple use) have been added.
197
198 (e) Exclusions and inclusions now apply to all files and directories, not
199 just to those obtained from scanning a directory recursively.
200
201 (f) Multiple uses of -f and --file-list are now supported.
202
203 (g) In a Windows environment, the default for -d has been changed from
204 "read" (the GNU grep default) to "skip", because otherwise the presence
205 of a directory in the file list provokes an error.
206
207 (h) The documentation has been revised and clarified in places.
208
209 9. Improve the matching speed of capturing brackets.
210
211 10. Changed the meaning of \X so that it now matches a Unicode extended
212 grapheme cluster.
213
214 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
215 out POSIX threads when JIT support is configured.
216
217 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
218
219 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
220 structure could go wrong in environments where size_t is not the same size
221 as int.
222
223 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
224 set.
225
226 15. The EBCDIC support had decayed; later updates to the code had included
227 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
228 general tidy up of EBCDIC-related issues, and the documentation was also
229 not quite right. There is now a test that can be run on ASCII systems to
230 check some of the EBCDIC-related things (but is it not a full test).
231
232 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
233 in a small tidy to the code.
234
235 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
236
237 18. If the --only-matching (-o) option in pcregrep is specified multiple
238 times, each one causes appropriate output. For example, -o1 -o2 outputs the
239 substrings matched by the 1st and 2nd capturing parentheses. A separating
240 string can be specified by --om-separator (default empty).
241
242 19. Improving the first n character searches.
243
244 20. Turn case lists for horizontal and vertical white space into macros so that
245 they are defined only once.
246
247 21. This set of changes together give more compatible Unicode case-folding
248 behaviour for characters that have more than one other case when UCP
249 support is available.
250
251 (a) The Unicode property table now has offsets into a new table of sets of
252 three or more characters that are case-equivalent. The MultiStage2.py
253 script that generates these tables (the pcre_ucd.c file) now scans
254 CaseFolding.txt instead of UnicodeData.txt for character case
255 information.
256
257 (b) The code for adding characters or ranges of characters to a character
258 class has been abstracted into a generalized function that also handles
259 case-independence. In UTF-mode with UCP support, this uses the new data
260 to handle characters with more than one other case.
261
262 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
263 whose other case is greater than 256 are now correctly matched
264 caselessly. Previously, the high codepoint matched the low one, but not
265 vice versa.
266
267 (d) The processing of \h, \H, \v, and \ in character classes now makes use
268 of the new class addition function, using character lists defined as
269 macros alongside the case definitions of 20 above.
270
271 (e) Caseless back references now work with characters that have more than
272 one other case.
273
274 (f) General caseless matching of characters with more than one other case
275 is supported.
276
277 22. Unicode character properties were updated from Unicode 6.2.0
278
279 23. Improved CMake support under Windows. Patch by Daniel Richard G.
280
281 24. Add support for 32-bit character strings, and UTF-32
282
283 25. Major JIT compiler update (code refactoring and bugfixing).
284 Experimental Sparc 32 support is added.
285
286 26. Applied a modified version of Daniel Richard G's patch to create
287 pcre.h.generic and config.h.generic by "make" instead of in the
288 PrepareRelease script.
289
290 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
291 pcre_compile.c when checking for a zero character.
292
293 28. Introducing a native interface for JIT. Through this interface, the compiled
294 machine code can be directly executed. The purpose of this interface is to
295 provide fast pattern matching, so several sanity checks are not performed.
296 However, feature tests are still performed. The new interface provides
297 1.4x speedup compared to the old one.
298
299 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
300 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
301 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
302
303 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
304 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
305 the "old" RFC 2279). Instead, it ended up passing a negative length to
306 pcre_exec().
307
308 31. Add support for GCC's visibility feature to hide internal functions.
309
310 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
311 "unknown -C option" after outputting 0 or 1.
312
313 33. There is now support for generating a code coverage report for the test
314 suite in environments where gcc is the compiler and lcov is installed. This
315 is mainly for the benefit of the developers.
316
317 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
318 unaddressable using valgrind annotations, allowing valgrind to detect
319 invalid memory accesses. This is mainly for the benefit of the developers.
320
321 25. (*UTF) can now be used to start a pattern in any of the three libraries.
322
323 26. Give configure error if --enable-cpp but no C++ compiler found.
324
325
326 Version 8.31 06-July-2012
327 -------------------------
328
329 1. Fixing a wrong JIT test case and some compiler warnings.
330
331 2. Removed a bashism from the RunTest script.
332
333 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
334 to unsigned type, result still unsigned" that was given by an MS compiler
335 on encountering the code "-sizeof(xxx)".
336
337 4. Partial matching support is added to the JIT compiler.
338
339 5. Fixed several bugs concerned with partial matching of items that consist
340 of more than one character:
341
342 (a) /^(..)\1/ did not partially match "aba" because checking references was
343 done on an "all or nothing" basis. This also applied to repeated
344 references.
345
346 (b) \R did not give a hard partial match if \r was found at the end of the
347 subject.
348
349 (c) \X did not give a hard partial match after matching one or more
350 characters at the end of the subject.
351
352 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
353 a partial match for the string "\r".
354
355 (e) When newline was set to CRLF, the metacharacter "." did not recognize
356 a partial match for a CR character at the end of the subject string.
357
358 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
359 running pcretest, the text "(JIT)" added to the output whenever JIT is
360 actually used to run the match.
361
362 7. Individual JIT compile options can be set in pcretest by following -s+[+]
363 or /S+[+] with a digit between 1 and 7.
364
365 8. OP_NOT now supports any UTF character not just single-byte ones.
366
367 9. (*MARK) control verb is now supported by the JIT compiler.
368
369 10. The command "./RunTest list" lists the available tests without actually
370 running any of them. (Because I keep forgetting what they all are.)
371
372 11. Add PCRE_INFO_MAXLOOKBEHIND.
373
374 12. Applied a (slightly modified) user-supplied patch that improves performance
375 when the heap is used for recursion (compiled with --disable-stack-for-
376 recursion). Instead of malloc and free for each heap frame each time a
377 logical recursion happens, frames are retained on a chain and re-used where
378 possible. This sometimes gives as much as 30% improvement.
379
380 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
381 call.
382
383 14. As documented, (*COMMIT) is now confined to within a positive assertion.
384
385 15. It is now possible to link pcretest with libedit as an alternative to
386 libreadline.
387
388 16. (*COMMIT) control verb is now supported by the JIT compiler.
389
390 17. The Unicode data tables have been updated to Unicode 6.1.0.
391
392 18. Added --file-list option to pcregrep.
393
394 19. Added binary file support to pcregrep, including the -a, --binary-files,
395 -I, and --text options.
396
397 20. The madvise function is renamed for posix_madvise for QNX compatibility
398 reasons. Fixed by Giuseppe D'Angelo.
399
400 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
401
402 22. Changed the option for creating symbolic links for 16-bit man pages from
403 -s to -sf so that re-installing does not cause issues.
404
405 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
406
407 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
408 matches in certain environments (the workspace was not being correctly
409 retained). Also added to pcre_dfa_exec() a simple plausibility check on
410 some of the workspace data at the beginning of a restart.
411
412 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
413 was not doing so when it should - probably a typo introduced by SVN 528
414 (change 8.10/14).
415
416 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
417 \w+ when the character tables indicated that \x{c4} was a word character.
418 There were several related cases, all because the tests for doing a table
419 lookup were testing for characters less than 127 instead of 255.
420
421 27. If a pattern contains capturing parentheses that are not used in a match,
422 their slots in the ovector are set to -1. For those that are higher than
423 any matched groups, this happens at the end of processing. In the case when
424 there were back references that the ovector was too small to contain
425 (causing temporary malloc'd memory to be used during matching), and the
426 highest capturing number was not used, memory off the end of the ovector
427 was incorrectly being set to -1. (It was using the size of the temporary
428 memory instead of the true size.)
429
430 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
431 ovector size, it uses memory at the end of the block that it has got.
432
433 29. Check for an overlong MARK name and give an error at compile time. The
434 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
435
436 30. JIT compiler update.
437
438 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
439 Rill for the patch.
440
441 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
442
443 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
444
445 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
446 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
447 was enabled.
448
449 35. Improve JIT code generation for greedy plus quantifier.
450
451 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
452 1 to "aa" instead of to an empty string. The bug affected repeated groups
453 that could potentially match an empty string.
454
455 37. Optimizing single character iterators in JIT.
456
457 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
458 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
459 codepoints that are too big for the mode are faulted, and in a UTF mode,
460 disallowed codepoints are also faulted.
461
462 39. If PCRE was compiled with UTF support, in three places in the DFA
463 matcher there was code that should only have been obeyed in UTF mode, but
464 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
465 processing when bytes with values greater than 127 were present. In 16-bit
466 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
467 both cases the values are those that cannot be the first data item in a UTF
468 character. The three items that might have provoked this were recursions,
469 possessively repeated groups, and atomic groups.
470
471 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
472 and pcregrep, because some OS require shared objects to be explicitly
473 passed to ld, causing the link step to fail if they are not.
474
475 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
476 mode, patterns that started with \h* or \R* might be incorrectly matched.
477
478
479 Version 8.30 04-February-2012
480 -----------------------------
481
482 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
483 name is defined in ctype.h.
484
485 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
486 only in quite long subpatterns.
487
488 3. Removed the function pcre_info(), which has been obsolete and deprecated
489 since it was replaced by pcre_fullinfo() in February 2000.
490
491 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
492 match a (*MARK), and the match failed at the start of the subject, a
493 reference to memory before the start of the subject could occur. This bug
494 was introduced by fix 17 of release 8.21.
495
496 5. A reference to an unset group with zero minimum repetition was giving
497 totally wrong answers (in non-JavaScript-compatibility mode). For example,
498 /(another)?(\1?)test/ matched against "hello world test". This bug was
499 introduced in release 8.13.
500
501 6. Add support for 16-bit character strings (a large amount of work involving
502 many changes and refactorings).
503
504 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
505 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
506 from a file.
507
508 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
509 rounding is not applied in this particular case).
510
511 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
512 if they appear, or are escaped, in patterns.
513
514 10. Get rid of a number of -Wunused-but-set-variable warnings.
515
516 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
517 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
518 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
519 also returns the mark "x". This bug applied to capturing parentheses,
520 non-capturing parentheses, and atomic parentheses. It also applied to some
521 assertions.
522
523 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
524 information out of configure.ac instead of relying on pcre.h.generic, which
525 is not stored in the repository.
526
527 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
528 -lreadline.
529
530 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
531
532 15. Applied Graycode's patch to put the top-level frame on the stack rather
533 than the heap when not using the stack for recursion. This gives a
534 performance improvement in many cases when recursion is not deep.
535
536 16. Experimental code added to "pcretest -C" to output the stack frame size.
537
538
539 Version 8.21 12-Dec-2011
540 ------------------------
541
542 1. Updating the JIT compiler.
543
544 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
545 are added as well.
546
547 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
548 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
549 calling _pcre_jit_exec. Some extra comments are added.
550
551 4. (*MARK) settings inside atomic groups that do not contain any capturing
552 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
553 was introduced by change 18 for 8.20.
554
555 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
556 ECMA-262 standard.
557
558 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
559 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
560 This bug was probably introduced by change 9 of 8.13.
561
562 7. While fixing 6 above, I noticed that a number of other items were being
563 incorrectly rejected as "not fixed length". This arose partly because newer
564 opcodes had not been added to the fixed-length checking code. I have (a)
565 corrected the bug and added tests for these items, and (b) arranged for an
566 error to occur if an unknown opcode is encountered while checking for fixed
567 length instead of just assuming "not fixed length". The items that were
568 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
569 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
570 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
571
572 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
573 being incorrectly compiled and would have given unpredicatble results.
574
575 9. A possessively repeated subpattern with minimum repeat count greater than
576 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
577 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
578 the first (A) could occur when it should not.
579
580 10. Add a cast and remove a redundant test from the code.
581
582 11. JIT should use pcre_malloc/pcre_free for allocation.
583
584 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
585 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
586 is anything other than /usr, -L is still shown).
587
588 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
589
590 14. Perl does not support \N without a following name in a [] class; PCRE now
591 also gives an error.
592
593 15. If a forward reference was repeated with an upper limit of around 2000,
594 it caused the error "internal error: overran compiling workspace". The
595 maximum number of forward references (including repeats) was limited by the
596 internal workspace, and dependent on the LINK_SIZE. The code has been
597 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
598 the default depends on LINK_SIZE. There is a new upper limit (for safety)
599 of around 200,000 forward references. While doing this, I also speeded up
600 the filling in of repeated forward references.
601
602 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
603 incorrectly expecting the subject to contain another "a" after the start.
604
605 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
606 in the match, the SKIP should be ignored. This was not happening; instead
607 the SKIP was being treated as NOMATCH. For patterns such as
608 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
609 tested.
610
611 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
612 now much more compatible with Perl, in particular in cases where the result
613 is a non-match for a non-anchored pattern. For example, if
614 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
615 "m", where previously it did not return a name. A side effect of this
616 change is that for partial matches, the last encountered mark name is
617 returned, as for non matches. A number of tests that were previously not
618 Perl-compatible have been moved into the Perl-compatible test files. The
619 refactoring has had the pleasing side effect of removing one argument from
620 the match() function, thus reducing its stack requirements.
621
622 19. If the /S+ option was used in pcretest to study a pattern using JIT,
623 subsequent uses of /S (without +) incorrectly behaved like /S+.
624
625 21. Retrieve executable code size support for the JIT compiler and fixing
626 some warnings.
627
628 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
629 not work when the shorter character appeared right at the end of the
630 subject string.
631
632 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
633 systems.
634
635 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
636 output it when the /M option is used in pcretest.
637
638 25. The CheckMan script was not being included in the distribution. Also, added
639 an explicit "perl" to run Perl scripts from the PrepareRelease script
640 because this is reportedly needed in Windows.
641
642 26. If study data was being save in a file and studying had not found a set of
643 "starts with" bytes for the pattern, the data written to the file (though
644 never used) was taken from uninitialized memory and so caused valgrind to
645 complain.
646
647 27. Updated RunTest.bat as provided by Sheri Pierce.
648
649 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
650
651 29. Computation of memory usage for the table of capturing group names was
652 giving an unnecessarily large value.
653
654
655 Version 8.20 21-Oct-2011
656 ------------------------
657
658 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
659 a POSIX class. After further experiments with Perl, which convinced me that
660 Perl has bugs and confusions, a closing square bracket is no longer allowed
661 in a POSIX name. This bug also affected patterns with classes that started
662 with full stops.
663
664 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
665 captured substring, but while checking the failing first alternative,
666 substring 1 is temporarily captured. If the output vector supplied to
667 pcre_exec() was not big enough for this capture, the yield of the function
668 was still zero ("insufficient space for captured substrings"). This cannot
669 be totally fixed without adding another stack variable, which seems a lot
670 of expense for a edge case. However, I have improved the situation in cases
671 such as /(a)(b)x|abc/ matched against "abc", where the return code
672 indicates that fewer than the maximum number of slots in the ovector have
673 been set.
674
675 3. Related to (2) above: when there are more back references in a pattern than
676 slots in the output vector, pcre_exec() uses temporary memory during
677 matching, and copies in the captures as far as possible afterwards. It was
678 using the entire output vector, but this conflicts with the specification
679 that only 2/3 is used for passing back captured substrings. Now it uses
680 only the first 2/3, for compatibility. This is, of course, another edge
681 case.
682
683 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
684 main code base, and can be used by building with --enable-jit. When this is
685 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
686 runtime --no-jit option is given.
687
688 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
689 ovector, the return from the function was zero, implying that there were
690 other matches that did not fit. The correct "exactly full" value is now
691 returned.
692
693 6. If a subpattern that was called recursively or as a subroutine contained
694 (*PRUNE) or any other control that caused it to give a non-standard return,
695 invalid errors such as "Error -26 (nested recursion at the same subject
696 position)" or even infinite loops could occur.
697
698 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
699 computing the minimum length on reaching *ACCEPT, and so ended up with the
700 wrong value of 1 rather than 0. Further investigation indicates that
701 computing a minimum subject length in the presence of *ACCEPT is difficult
702 (think back references, subroutine calls), and so I have changed the code
703 so that no minimum is registered for a pattern that contains *ACCEPT.
704
705 8. If (*THEN) was present in the first (true) branch of a conditional group,
706 it was not handled as intended. [But see 16 below.]
707
708 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
709 Sheri Pierce.
710
711 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
712 the first byte in a match must be "a".
713
714 11. Change 17 for 8.13 increased the recursion depth for patterns like
715 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
716 pattern contains any instances of (*THEN). If it does not, the old
717 optimizations are restored. It would be nice to do this on a per-group
718 basis, but at the moment that is not feasible.
719
720 12. In some environments, the output of pcretest -C is CRLF terminated. This
721 broke RunTest's code that checks for the link size. A single white space
722 character after the value is now allowed for.
723
724 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
725 For "fr", it uses the Windows-specific input and output files.
726
727 14. If (*THEN) appeared in a group that was called recursively or as a
728 subroutine, it did not work as intended. [But see next item.]
729
730 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
731 pattern fragments (but not containing any | characters). If A and B are
732 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
733 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
734 D. In other words, Perl considers parentheses that do not contain any |
735 characters to be part of a surrounding alternative, whereas PCRE was
736 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
737 differently. PCRE now behaves in the same way as Perl, except in the case
738 of subroutine/recursion calls such as (?1) which have in any case always
739 been different (but PCRE had them first :-).
740
741 16. Related to 15 above: Perl does not treat the | in a conditional group as
742 creating alternatives. Such a group is treated in the same way as an
743 ordinary group without any | characters when processing (*THEN). PCRE has
744 been changed to match Perl's behaviour.
745
746 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
747 RunGrepTest script failed.
748
749 18. Change 22 for version 13 caused atomic groups to use more stack. This is
750 inevitable for groups that contain captures, but it can lead to a lot of
751 stack use in large patterns. The old behaviour has been restored for atomic
752 groups that do not contain any capturing parentheses.
753
754 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
755 suppress the check for a minimum subject length at run time. (If it was
756 given to pcre_exec() or pcre_dfa_exec() it did work.)
757
758 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
759 fail to work when decoding hex characters in data strings in EBCDIC
760 environments.
761
762 21. It appears that in at least one Mac OS environment, the isxdigit() function
763 is implemented as a macro that evaluates to its argument more than once,
764 contravening the C 90 Standard (I haven't checked a later standard). There
765 was an instance in pcretest which caused it to go wrong when processing
766 \x{...} escapes in subject strings. The has been rewritten to avoid using
767 things like p++ in the argument of isxdigit().
768
769
770 Version 8.13 16-Aug-2011
771 ------------------------
772
773 1. The Unicode data tables have been updated to Unicode 6.0.0.
774
775 2. Two minor typos in pcre_internal.h have been fixed.
776
777 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
778 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
779 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
780
781 4. There were a number of related bugs in the code for matching backrefences
782 caselessly in UTF-8 mode when codes for the characters concerned were
783 different numbers of bytes. For example, U+023A and U+2C65 are an upper
784 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
785 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
786 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
787 2-byte code at the end of the subject (it thought there wasn't enough data
788 left).
789
790 5. Comprehensive information about what went wrong is now returned by
791 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
792 as the output vector has at least 2 elements. The offset of the start of
793 the failing character and a reason code are placed in the vector.
794
795 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
796 now returned is for the first byte of the failing character, instead of the
797 last byte inspected. This is an incompatible change, but I hope it is small
798 enough not to be a problem. It makes the returned offset consistent with
799 pcre_exec() and pcre_dfa_exec().
800
801 7. pcretest now gives a text phrase as well as the error number when
802 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
803 failure, the offset and reason code are output.
804
805 8. When \R was used with a maximizing quantifier it failed to skip backwards
806 over a \r\n pair if the subsequent match failed. Instead, it just skipped
807 back over a single character (\n). This seems wrong (because it treated the
808 two characters as a single entity when going forwards), conflicts with the
809 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
810 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
811 has been changed.
812
813 9. Some internal refactoring has changed the processing so that the handling
814 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
815 time (the PCRE_DOTALL option was changed this way some time ago: version
816 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
817 which was always a bit of a fudge. It also means that there is one less
818 argument for the match() function, which reduces its stack requirements
819 slightly. This change also fixes an incompatibility with Perl: the pattern
820 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
821
822 10. More internal refactoring has drastically reduced the number of recursive
823 calls to match() for possessively repeated groups such as (abc)++ when
824 using pcre_exec().
825
826 11. While implementing 10, a number of bugs in the handling of groups were
827 discovered and fixed:
828
829 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
830 (a|)*(?1) gave a compile-time internal error.
831 ((a|)+)+ did not notice that the outer group could match an empty string.
832 (^a|^)+ was not marked as anchored.
833 (.*a|.*)+ was not marked as matching at start or after a newline.
834
835 12. Yet more internal refactoring has removed another argument from the match()
836 function. Special calls to this function are now indicated by setting a
837 value in a variable in the "match data" data block.
838
839 13. Be more explicit in pcre_study() instead of relying on "default" for
840 opcodes that mean there is no starting character; this means that when new
841 ones are added and accidentally left out of pcre_study(), testing should
842 pick them up.
843
844 14. The -s option of pcretest has been documented for ages as being an old
845 synonym of -m (show memory usage). I have changed it to mean "force study
846 for every regex", that is, assume /S for every regex. This is similar to -i
847 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
848 using it. It makes it easier to run collections of tests with and without
849 study enabled, and thereby test pcre_study() more easily. All the standard
850 tests are now run with and without -s (but some patterns can be marked as
851 "never study" - see 20 below).
852
853 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
854 restoration of the capturing data to the outer values was not happening
855 correctly.
856
857 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
858 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
859 pattern had matched an empty string, and so incorrectly returned a no
860 match.
861
862 17. There was optimizing code for the last branch of non-capturing parentheses,
863 and also for the obeyed branch of a conditional subexpression, which used
864 tail recursion to cut down on stack usage. Unfortunately, now that there is
865 the possibility of (*THEN) occurring in these branches, tail recursion is
866 no longer possible because the return has to be checked for (*THEN). These
867 two optimizations have therefore been removed. [But see 8.20/11 above.]
868
869 18. If a pattern containing \R was studied, it was assumed that \R always
870 matched two bytes, thus causing the minimum subject length to be
871 incorrectly computed because \R can also match just one byte.
872
873 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
874 was incorrectly computed.
875
876 20. If /S is present twice on a test pattern in pcretest input, it now
877 *disables* studying, thereby overriding the use of -s on the command line
878 (see 14 above). This is necessary for one or two tests to keep the output
879 identical in both cases.
880
881 21. When (*ACCEPT) was used in an assertion that matched an empty string and
882 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
883
884 22. When an atomic group that contained a capturing parenthesis was
885 successfully matched, but the branch in which it appeared failed, the
886 capturing was not being forgotten if a higher numbered group was later
887 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
888 group 1 to "a", when in fact it should be unset. This applied to multi-
889 branched capturing and non-capturing groups, repeated or not, and also to
890 positive assertions (capturing in negative assertions does not happen
891 in PCRE) and also to nested atomic groups.
892
893 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
894 subject after a captured substring, to make it easier to tell which of a
895 number of identical substrings has been captured.
896
897 24. The way atomic groups are processed by pcre_exec() has been changed so that
898 if they are repeated, backtracking one repetition now resets captured
899 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
900 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
901 "aaa". Previously, it would have been "a". As part of this code
902 refactoring, the way recursive calls are handled has also been changed.
903
904 25. If an assertion condition captured any substrings, they were not passed
905 back unless some other capturing happened later. For example, if
906 (?(?=(a))a) was matched against "a", no capturing was returned.
907
908 26. When studying a pattern that contained subroutine calls or assertions,
909 the code for finding the minimum length of a possible match was handling
910 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
911 group 1 called group 2 while simultaneously a separate group 2 called group
912 1). A stack overflow occurred in this case. I have fixed this by limiting
913 the recursion depth to 10.
914
915 27. Updated RunTest.bat in the distribution to the version supplied by Tom
916 Fortmann. This supports explicit test numbers on the command line, and has
917 argument validation and error reporting.
918
919 28. An instance of \X with an unlimited repeat could fail if at any point the
920 first character it looked at was a mark character.
921
922 29. Some minor code refactoring concerning Unicode properties and scripts
923 should reduce the stack requirement of match() slightly.
924
925 30. Added the '=' option to pcretest to check the setting of unused capturing
926 slots at the end of the pattern, which are documented as being -1, but are
927 not included in the return count.
928
929 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
930 compiled something random. Now it gives a compile-time error (as does
931 Perl).
932
933 32. A *MARK encountered during the processing of a positive assertion is now
934 recorded and passed back (compatible with Perl).
935
936 33. If --only-matching or --colour was set on a pcregrep call whose pattern
937 had alternative anchored branches, the search for a second match in a line
938 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
939 matched the line "0102" twice. The same bug affected patterns that started
940 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
941 twice.
942
943 34. Previously, PCRE did not allow quantification of assertions. However, Perl
944 does, and because of capturing effects, quantifying parenthesized
945 assertions may at times be useful. Quantifiers are now allowed for
946 parenthesized assertions.
947
948 35. A minor code tidy in pcre_compile() when checking options for \R usage.
949
950 36. \g was being checked for fancy things in a character class, when it should
951 just be a literal "g".
952
953 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
954 appearance of a nested POSIX class supersedes an apparent external class.
955 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
956 unescaped square brackets may also appear as part of class names. For
957 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
958 more like Perl. (But see 8.20/1 above.)
959
960 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
961 was because it thought it was \N{name}, which is not supported).
962
963 39. Add minix to OS list not supporting the -S option in pcretest.
964
965 40. PCRE tries to detect cases of infinite recursion at compile time, but it
966 cannot analyze patterns in sufficient detail to catch mutual recursions
967 such as ((?1))((?2)). There is now a runtime test that gives an error if a
968 subgroup is called recursively as a subpattern for a second time at the
969 same position in the subject string. In previous releases this might have
970 been caught by the recursion limit, or it might have run out of stack.
971
972 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
973 happen only once. PCRE was, however incorrectly giving a compile time error
974 "recursive call could loop indefinitely" because it cannot analyze the
975 pattern in sufficient detail. The compile time test no longer happens when
976 PCRE is compiling a conditional subpattern, but actual runaway loops are
977 now caught at runtime (see 40 above).
978
979 42. It seems that Perl allows any characters other than a closing parenthesis
980 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
981 has been changed to be the same.
982
983 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
984 as not to get warnings when autogen.sh is called. Also changed
985 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
986
987 44. To help people who use pcregrep to scan files containing exceedingly long
988 lines, the following changes have been made:
989
990 (a) The default value of the buffer size parameter has been increased from
991 8K to 20K. (The actual buffer used is three times this size.)
992
993 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
994 PCRE is built.
995
996 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
997 to be set at run time.
998
999 (d) Numerical values in pcregrep options can be followed by K or M, for
1000 example --buffer-size=50K.
1001
1002 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1003 given and the return code is set to 2.
1004
1005 45. Add a pointer to the latest mark to the callout data block.
1006
1007 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1008 partial match of an empty string instead of no match. This was specific to
1009 the use of ".".
1010
1011 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1012 complete match instead of a partial match. This bug was dependent on both
1013 the PCRE_UTF8 and PCRE_DOTALL options being set.
1014
1015 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1016 starting byte set, because \b was not being ignored.
1017
1018
1019 Version 8.12 15-Jan-2011
1020 ------------------------
1021
1022 1. Fixed some typos in the markup of the man pages, and wrote a script that
1023 checks for such things as part of the documentation building process.
1024
1025 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1026 --match-limit and --recursion-limit options (added for 8.11). In
1027 particular, this made one of the standard tests fail. (The integer value
1028 went into the wrong half of a long int.)
1029
1030 3. If the --colour option was given to pcregrep with -v (invert match), it
1031 did strange things, either producing crazy output, or crashing. It should,
1032 of course, ignore a request for colour when reporting lines that do not
1033 match.
1034
1035 4. Another pcregrep bug caused similar problems if --colour was specified with
1036 -M (multiline) and the pattern match finished with a line ending.
1037
1038 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1039 matched in multiline mode, the following line was shown as part of the
1040 match. This seems wrong, so I have changed it.
1041
1042 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1043 the check for further matches in the same line (so they could be coloured)
1044 to overrun the end of the current line. If another match was found, it was
1045 incorrectly shown (and then shown again when found in the next line).
1046
1047 7. If pcregrep was compiled under Windows, there was a reference to the
1048 function pcregrep_exit() before it was defined. I am assuming this was
1049 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1050 reported by a user. I've moved the definition above the reference.
1051
1052
1053 Version 8.11 10-Dec-2010
1054 ------------------------
1055
1056 1. (*THEN) was not working properly if there were untried alternatives prior
1057 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1058 backtracked to try for "b" instead of moving to the next alternative branch
1059 at the same level (in this case, to look for "c"). The Perl documentation
1060 is clear that when (*THEN) is backtracked onto, it goes to the "next
1061 alternative in the innermost enclosing group".
1062
1063 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1064 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1065 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1066 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1067 (*THEN).
1068
1069 3. If \s appeared in a character class, it removed the VT character from
1070 the class, even if it had been included by some previous item, for example
1071 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1072 of \s, but is part of the POSIX "space" class.)
1073
1074 4. A partial match never returns an empty string (because you can always
1075 match an empty string at the end of the subject); however the checking for
1076 an empty string was starting at the "start of match" point. This has been
1077 changed to the "earliest inspected character" point, because the returned
1078 data for a partial match starts at this character. This means that, for
1079 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1080 (previously it gave "no match").
1081
1082 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1083 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1084 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1085 has an implication that the given string is incomplete (because a partial
1086 match is preferred over a full match). For this reason, these items now
1087 give a partial match in this situation. [Aside: previously, the one case
1088 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1089 match rather than a full match, which was wrong by the old rules, but is
1090 now correct.]
1091
1092 6. There was a bug in the handling of #-introduced comments, recognized when
1093 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1094 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1095 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1096 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1097 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1098 places in pcre_compile().
1099
1100 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1101 comments when looking ahead for named forward references to subpatterns,
1102 the only newline sequence it recognized was NL. It now handles newlines
1103 according to the set newline convention.
1104
1105 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1106 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1107 cater for a lack of strerror(). These oversights have been fixed.
1108
1109 9. Added --match-limit and --recursion-limit to pcregrep.
1110
1111 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1112
1113 11. When the -o option was used, pcregrep was setting a return code of 1, even
1114 when matches were found, and --line-buffered was not being honoured.
1115
1116 12. Added an optional parentheses number to the -o and --only-matching options
1117 of pcregrep.
1118
1119 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1120 can match an empty string. The code to do it in pcretest and pcredemo
1121 needed fixing:
1122
1123 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1124 only one byte after an empty string match just before CRLF (this case
1125 just got forgotten; "any" and "anycrlf" were OK).
1126
1127 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1128 mode when an empty string match preceded an ASCII character followed by
1129 a non-ASCII character. (The code for advancing by one character rather
1130 than one byte was nonsense.)
1131
1132 (c) The pcredemo.c sample program did not have any code at all to handle
1133 the cases when CRLF is a valid newline sequence.
1134
1135 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1136 as a starting offset was within the subject string. There is now a new
1137 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1138 negative or greater than the length of the string. In order to test this,
1139 pcretest is extended to allow the setting of negative starting offsets.
1140
1141 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1142 starting offset points to the beginning of a UTF-8 character was
1143 unnecessarily clumsy. I tidied it up.
1144
1145 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1146 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1147
1148 17. Nobody had reported that the --include_dir option, which was added in
1149 release 7.7 should have been called --include-dir (hyphen, not underscore)
1150 for compatibility with GNU grep. I have changed it to --include-dir, but
1151 left --include_dir as an undocumented synonym, and the same for
1152 --exclude-dir, though that is not available in GNU grep, at least as of
1153 release 2.5.4.
1154
1155 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1156 characters from a string of bytes) have been redefined so as not to use
1157 loops, in order to improve performance in some environments. At the same
1158 time, I abstracted some of the common code into auxiliary macros to save
1159 repetition (this should not affect the compiled code).
1160
1161 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1162 compile-time error is now given if \c is not followed by an ASCII
1163 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1164 different, and any byte value is allowed.)
1165
1166 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1167 START_OPTIMIZE option, which is now allowed at compile time - but just
1168 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1169 to pcregrep and other applications that have no direct access to PCRE
1170 options. The new /Y option in pcretest sets this option when calling
1171 pcre_compile().
1172
1173 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1174 back references. Groups containing recursive back references were forced to
1175 be atomic by that change, but in the case of named groups, the amount of
1176 memory required was incorrectly computed, leading to "Failed: internal
1177 error: code overflow". This has been fixed.
1178
1179 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1180 pcretest.c, to avoid build problems in some Borland environments.
1181
1182
1183 Version 8.10 25-Jun-2010
1184 ------------------------
1185
1186 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1187 THEN.
1188
1189 2. (*ACCEPT) was not working when inside an atomic group.
1190
1191 3. Inside a character class, \B is treated as a literal by default, but
1192 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1193 causes the error). The code is unchanged, but I tidied the documentation.
1194
1195 4. Inside a character class, PCRE always treated \R and \X as literals,
1196 whereas Perl faults them if its -w option is set. I have changed PCRE so
1197 that it faults them when PCRE_EXTRA is set.
1198
1199 5. Added support for \N, which always matches any character other than
1200 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1201
1202 6. When compiling pcregrep with newer versions of gcc which may have
1203 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1204 declared with attribute warn_unused_result" were given. Just casting the
1205 result to (void) does not stop the warnings; a more elaborate fudge is
1206 needed. I've used a macro to implement this.
1207
1208 7. Minor change to pcretest.c to avoid a compiler warning.
1209
1210 8. Added four artifical Unicode properties to help with an option to make
1211 \s etc use properties (see next item). The new properties are: Xan
1212 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1213
1214 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1215 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1216 this option. Modified pcretest to add /W to test this facility. Added
1217 REG_UCP to make it available via the POSIX interface.
1218
1219 10. Added --line-buffered to pcregrep.
1220
1221 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1222 studied, and the match started with a letter with a code point greater than
1223 127 whose first byte was different to the first byte of the other case of
1224 the letter, the other case of this starting letter was not recognized
1225 (#976).
1226
1227 12. If a pattern that was studied started with a repeated Unicode property
1228 test, for example, \p{Nd}+, there was the theoretical possibility of
1229 setting up an incorrect bitmap of starting bytes, but fortunately it could
1230 not have actually happened in practice until change 8 above was made (it
1231 added property types that matched character-matching opcodes).
1232
1233 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1234 possible starting bytes for non-anchored patterns.
1235
1236 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1237 \R, and also a number of cases that involve Unicode properties, both
1238 explicit and implicit when PCRE_UCP is set.
1239
1240 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1241 input, it could crash or give wrong results if characters with values
1242 greater than 0xc0 were present in the subject string. (Detail: it assumed
1243 UTF-8 input when processing these items.)
1244
1245 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1246 size_t is 64-bit (#991).
1247
1248 17. Added a check for running out of memory when PCRE is compiled with
1249 --disable-stack-for-recursion (#990).
1250
1251 18. If the last data line in a file for pcretest does not have a newline on
1252 the end, a newline was missing in the output.
1253
1254 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1255 less than 128) in its various bitmaps. However, there is a facility for
1256 generating tables according to the current locale when PCRE is compiled. It
1257 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1258 characters, are recognized by isspace() and therefore were getting set in
1259 these tables, and indeed these tables seem to approximate to ISO 8859. This
1260 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1261 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1262 which of course cannot start UTF-8 characters. I have changed the code so
1263 that only real ASCII characters (less than 128) and the correct starting
1264 bytes for UTF-8 encodings are set for characters greater than 127 when in
1265 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1266 altogether.)
1267
1268 20. Added the /T option to pcretest so as to be able to run tests with non-
1269 standard character tables, thus making it possible to include the tests
1270 used for 19 above in the standard set of tests.
1271
1272 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1273 reference to a subpattern the other side of a comment that contains an
1274 opening parenthesis caused either an internal compiling error, or a
1275 reference to the wrong subpattern.
1276
1277
1278 Version 8.02 19-Mar-2010
1279 ------------------------
1280
1281 1. The Unicode data tables have been updated to Unicode 5.2.0.
1282
1283 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1284 configured.
1285
1286 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1287 original author of that file, following a query about its status.
1288
1289 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1290 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1291
1292 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1293 quantifier applied to a forward-referencing subroutine call, could compile
1294 incorrect code or give the error "internal error: previously-checked
1295 referenced subpattern not found".
1296
1297 6. Both MS Visual Studio and Symbian OS have problems with initializing
1298 variables to point to external functions. For these systems, therefore,
1299 pcre_malloc etc. are now initialized to local functions that call the
1300 relevant global functions.
1301
1302 7. There were two entries missing in the vectors called coptable and poptable
1303 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1304 I've fixed the data, and added a kludgy way of testing at compile time that
1305 the lengths are correct (equal to the number of opcodes).
1306
1307 8. Following on from 7, I added a similar kludge to check the length of the
1308 eint vector in pcreposix.c.
1309
1310 9. Error texts for pcre_compile() are held as one long string to avoid too
1311 much relocation at load time. To find a text, the string is searched,
1312 counting zeros. There was no check for running off the end of the string,
1313 which could happen if a new error number was added without updating the
1314 string.
1315
1316 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1317
1318 11. \K was not working if it appeared in an atomic group or in a group that
1319 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1320 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1321 the assertion is positive, but not if it is negative.
1322
1323 12. Change 11 fortuitously reduced the size of the stack frame used in the
1324 "match()" function of pcre_exec.c by one pointer. Forthcoming
1325 implementation of support for (*MARK) will need an extra pointer on the
1326 stack; I have reserved it now, so that the stack frame size does not
1327 decrease.
1328
1329 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1330 item in branch that calls a recursion is a subroutine call - as in the
1331 second branch in the above example - was incorrectly given the compile-
1332 time error "recursive call could loop indefinitely" because pcre_compile()
1333 was not correctly checking the subroutine for matching a non-empty string.
1334
1335 14. The checks for overrunning compiling workspace could trigger after an
1336 overrun had occurred. This is a "should never occur" error, but it can be
1337 triggered by pathological patterns such as hundreds of nested parentheses.
1338 The checks now trigger 100 bytes before the end of the workspace.
1339
1340 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1341
1342
1343 Version 8.01 19-Jan-2010
1344 ------------------------
1345
1346 1. If a pattern contained a conditional subpattern with only one branch (in
1347 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1348 computed the wrong minimum data length (which is of course zero for such
1349 subpatterns). This could cause incorrect "no match" results.
1350
1351 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1352 the pattern is reset in the first branch, pcre_compile() failed with
1353 "internal error: code overflow at offset...". This happened only when
1354 the reset was to the original external option setting. (An optimization
1355 abstracts leading options settings into an external setting, which was the
1356 cause of this.)
1357
1358 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1359 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1360 assertion pattern did not match (meaning that the assertion was true), it
1361 was incorrectly treated as false if the SKIP had been reached during the
1362 matching. This also applied to assertions used as conditions.
1363
1364 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1365 assertion subpattern, including such a pattern used as a condition,
1366 unpredictable results occurred, instead of the error return
1367 PCRE_ERROR_DFA_UITEM.
1368
1369 5. The C++ GlobalReplace function was not working like Perl for the special
1370 situation when an empty string is matched. It now does the fancy magic
1371 stuff that is necessary.
1372
1373 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1374 removed. (These were left over from very, very early versions of PCRE.)
1375
1376 7. Some cosmetic changes to the code to make life easier when compiling it
1377 as part of something else:
1378
1379 (a) Change DEBUG to PCRE_DEBUG.
1380
1381 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1382 called "current" as "current_branch", to prevent a collision with the
1383 Linux macro when compiled as a kernel module.
1384
1385 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1386 prevent a collision with the Linux macro when compiled as a kernel
1387 module.
1388
1389 8. In pcre_compile() there are some checks for integer overflows that used to
1390 cast potentially large values to (double). This has been changed to that
1391 when building, a check for int64_t is made, and if it is found, it is used
1392 instead, thus avoiding the use of floating point arithmetic. (There is no
1393 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1394 double.
1395
1396 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1397 2005 (difference between two addresses compared to an unsigned value).
1398
1399 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1400 custom one, because of the following reported problem in Windows:
1401
1402 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1403 under Win32.
1404 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1405 therefore missing the function definition.
1406 - The compiler thus generates a "C" signature for the test function.
1407 - The linker fails to find the "C" function.
1408 - PCRE fails to configure if asked to do so against libbz2.
1409
1410 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1411 messages were output:
1412
1413 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1414 rerunning libtoolize, to keep the correct libtool macros in-tree.
1415 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1416
1417 I have done both of these things.
1418
1419 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1420 most of the time, it *can* run out if it is given a pattern that contains a
1421 runaway infinite recursion. I updated the discussion in the pcrestack man
1422 page.
1423
1424 13. Now that we have gone to the x.xx style of version numbers, the minor
1425 version may start with zero. Using 08 or 09 is a bad idea because users
1426 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1427 interpreted as invalid octal numbers. I've updated the previous comment in
1428 configure.ac, and also added a check that gives an error if 08 or 09 are
1429 used.
1430
1431 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1432 causing partial matching to fail when the end of the subject matched \W
1433 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1434
1435 15. There were some discrepancies between the declarations in pcre_internal.h
1436 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1437 their definitions. The declarations used "const uschar *" and the
1438 definitions used USPTR. Even though USPTR is normally defined as "const
1439 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1440 reported that: "This difference in casting confuses some C++ compilers, for
1441 example, SunCC recognizes above declarations as different functions and
1442 generates broken code for hbpcre." I have changed the declarations to use
1443 USPTR.
1444
1445 16. GNU libtool is named differently on some systems. The autogen.sh script now
1446 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1447 (FreeBSD).
1448
1449 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1450 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1451 comment: "Figure out how to create a longlong from a string: strtoll and
1452 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1453 instance, but it only takes 2 args instead of 3!"
1454
1455 18. A subtle bug concerned with back references has been fixed by a change of
1456 specification, with a corresponding code fix. A pattern such as
1457 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1458 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1459 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1460 same bug. Such groups have to be quantified to be useful, or contained
1461 inside another quantified group. (If there's no repetition, the reference
1462 can never match.) The problem arises because, having left the group and
1463 moved on to the rest of the pattern, a later failure that backtracks into
1464 the group uses the captured value from the final iteration of the group
1465 rather than the correct earlier one. I have fixed this in PCRE by forcing
1466 any group that contains a reference to itself to be an atomic group; that
1467 is, there cannot be any backtracking into it once it has completed. This is
1468 similar to recursive and subroutine calls.
1469
1470
1471 Version 8.00 19-Oct-09
1472 ----------------------
1473
1474 1. The table for translating pcre_compile() error codes into POSIX error codes
1475 was out-of-date, and there was no check on the pcre_compile() error code
1476 being within the table. This could lead to an OK return being given in
1477 error.
1478
1479 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1480 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1481 in a Windows environment.
1482
1483 3. The pcregrep --count option prints the count for each file even when it is
1484 zero, as does GNU grep. However, pcregrep was also printing all files when
1485 --files-with-matches was added. Now, when both options are given, it prints
1486 counts only for those files that have at least one match. (GNU grep just
1487 prints the file name in this circumstance, but including the count seems
1488 more useful - otherwise, why use --count?) Also ensured that the
1489 combination -clh just lists non-zero counts, with no names.
1490
1491 4. The long form of the pcregrep -F option was incorrectly implemented as
1492 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1493 but it seems right to fix it, and I didn't think it was worth preserving
1494 the old behaviour.
1495
1496 5. The command line items --regex=pattern and --regexp=pattern were not
1497 recognized by pcregrep, which required --regex pattern or --regexp pattern
1498 (with a space rather than an '='). The man page documented the '=' forms,
1499 which are compatible with GNU grep; these now work.
1500
1501 6. No libpcreposix.pc file was created for pkg-config; there was just
1502 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1503
1504 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1505 when UCP support is not needed, by modifying the Python script that
1506 generates it from Unicode data files. This should not matter if the module
1507 is correctly used as a library, but I received one complaint about 50K of
1508 unwanted data. My guess is that the person linked everything into his
1509 program rather than using a library. Anyway, it does no harm.
1510
1511 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1512 was a minimum greater than 1 for a wide character in a possessive
1513 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1514 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1515 character. Chaos in the form of incorrect output or a compiling loop could
1516 result.
1517
1518 9. The restrictions on what a pattern can contain when partial matching is
1519 requested for pcre_exec() have been removed. All patterns can now be
1520 partially matched by this function. In addition, if there are at least two
1521 slots in the offset vector, the offset of the earliest inspected character
1522 for the match and the offset of the end of the subject are set in them when
1523 PCRE_ERROR_PARTIAL is returned.
1524
1525 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1526 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1527 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1528 and may be more useful for multi-segment matching.
1529
1530 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1531 used to be given if ever the end of the subject was reached; now it is
1532 given only if matching could not proceed because another character was
1533 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1534 string "Z", which now yields "no match" instead of "partial match". In the
1535 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1536 final character ended with (*FAIL).
1537
1538 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1539 if the pattern had a "must contain" character that was already found in the
1540 earlier partial match, unless partial matching was again requested. For
1541 example, with the pattern /dog.(body)?/, the "must contain" character is
1542 "g". If the first part-match was for the string "dog", restarting with
1543 "sbody" failed. This bug has been fixed.
1544
1545 13. The string returned by pcre_dfa_exec() after a partial match has been
1546 changed so that it starts at the first inspected character rather than the
1547 first character of the match. This makes a difference only if the pattern
1548 starts with a lookbehind assertion or \b or \B (\K is not supported by
1549 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1550 matching functions compatible, and I think it's the right thing to do.
1551
1552 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1553 so that the demonstration program is easily available in environments where
1554 PCRE has not been installed from source.
1555
1556 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1557 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1558 library.
1559
1560 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1561 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1562 is not the first non-POSIX option to be added. Clearly some people find
1563 these options useful.
1564
1565 17. If a caller to the POSIX matching function regexec() passes a non-zero
1566 value for nmatch with a NULL value for pmatch, the value of
1567 nmatch is forced to zero.
1568
1569 18. RunGrepTest did not have a test for the availability of the -u option of
1570 the diff command, as RunTest does. It now checks in the same way as
1571 RunTest, and also checks for the -b option.
1572
1573 19. If an odd number of negated classes containing just a single character
1574 interposed, within parentheses, between a forward reference to a named
1575 subpattern and the definition of the subpattern, compilation crashed with
1576 an internal error, complaining that it could not find the referenced
1577 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1578 [The bug was that it was starting one character too far in when skipping
1579 over the character class, thus treating the ] as data rather than
1580 terminating the class. This meant it could skip too much.]
1581
1582 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1583 /g option in pcretest when the pattern contains \K, which makes it possible
1584 to have an empty string match not at the start, even when the pattern is
1585 anchored. Updated pcretest and pcredemo to use this option.
1586
1587 21. If the maximum number of capturing subpatterns in a recursion was greater
1588 than the maximum at the outer level, the higher number was returned, but
1589 with unset values at the outer level. The correct (outer level) value is
1590 now given.
1591
1592 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1593 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1594 make it do so. The string so far is captured, making this feature
1595 compatible with Perl.
1596
1597 23. The tests have been re-organized, adding tests 11 and 12, to make it
1598 possible to check the Perl 5.10 features against Perl 5.10.
1599
1600 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1601 pattern matches a fixed length string. PCRE did not allow this; now it
1602 does. Neither allows recursion.
1603
1604 25. I finally figured out how to implement a request to provide the minimum
1605 length of subject string that was needed in order to match a given pattern.
1606 (It was back references and recursion that I had previously got hung up
1607 on.) This code has now been added to pcre_study(); it finds a lower bound
1608 to the length of subject needed. It is not necessarily the greatest lower
1609 bound, but using it to avoid searching strings that are too short does give
1610 some useful speed-ups. The value is available to calling programs via
1611 pcre_fullinfo().
1612
1613 26. While implementing 25, I discovered to my embarrassment that pcretest had
1614 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1615 study optimizations had never been tested with that matching function.
1616 Oops. What is worse, even when it was passed study data, there was a bug in
1617 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1618 were also very few tests of studied patterns with pcre_dfa_exec().
1619
1620 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1621 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1622 on the other side of the coin, they are no longer allowed to have different
1623 names, because these cannot be distinguished in PCRE, and this has caused
1624 confusion. (This is a difference from Perl.)
1625
1626 28. When duplicate subpattern names are present (necessarily with different
1627 numbers, as required by 27 above), and a test is made by name in a
1628 conditional pattern, either for a subpattern having been matched, or for
1629 recursion in such a pattern, all the associated numbered subpatterns are
1630 tested, and the overall condition is true if the condition is true for any
1631 one of them. This is the way Perl works, and is also more like the way
1632 testing by number works.
1633
1634
1635 Version 7.9 11-Apr-09
1636 ---------------------
1637
1638 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1639 (pcretest), all targets were linked against these libraries. This included
1640 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1641 libraries. This caused unwanted dependencies to be created. This problem
1642 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1643 pcretest is linked with readline.
1644
1645 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1646 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1647 moved outside it again, because FALSE and TRUE are already defined in AIX,
1648 but BOOL is not.
1649
1650 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1651 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1652
1653 4. The pcregrep documentation said spaces were inserted as well as colons (or
1654 hyphens) following file names and line numbers when outputting matching
1655 lines. This is not true; no spaces are inserted. I have also clarified the
1656 wording for the --colour (or --color) option.
1657
1658 5. In pcregrep, when --colour was used with -o, the list of matching strings
1659 was not coloured; this is different to GNU grep, so I have changed it to be
1660 the same.
1661
1662 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1663 each matching line was coloured. Now it goes on to look for further matches
1664 of any of the test patterns, which is the same behaviour as GNU grep.
1665
1666 7. A pattern that could match an empty string could cause pcregrep to loop; it
1667 doesn't make sense to accept an empty string match in pcregrep, so I have
1668 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1669 seems to be how GNU grep behaves. [But see later change 40 for release
1670 8.33.]
1671
1672 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1673 start or after a newline", because the conditional assertion was not being
1674 correctly handled. The rule now is that both the assertion and what follows
1675 in the first alternative must satisfy the test.
1676
1677 9. If auto-callout was enabled in a pattern with a conditional group whose
1678 condition was an assertion, PCRE could crash during matching, both with
1679 pcre_exec() and pcre_dfa_exec().
1680
1681 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1682 used for matching.
1683
1684 11. Unicode property support in character classes was not working for
1685 characters (bytes) greater than 127 when not in UTF-8 mode.
1686
1687 12. Added the -M command line option to pcretest.
1688
1689 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1690
1691 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1692
1693 16. Added comments and documentation about mis-use of no_arg in the C++
1694 wrapper.
1695
1696 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1697 from Martin Jerabek that uses macro names for all relevant character and
1698 string constants.
1699
1700 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1701 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1702 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1703 these, but not everybody uses configure.
1704
1705 19. A conditional group that had only one branch was not being correctly
1706 recognized as an item that could match an empty string. This meant that an
1707 enclosing group might also not be so recognized, causing infinite looping
1708 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1709 with the subject "ab", where knowledge that the repeated group can match
1710 nothing is needed in order to break the loop.
1711
1712 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1713 exec(), but without supplying a callout function, matching went wrong.
1714
1715 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1716 leak if the size of the offset vector was greater than 30. When the vector
1717 is smaller, the saved offsets during recursion go onto a local stack
1718 vector, but for larger vectors malloc() is used. It was failing to free
1719 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1720 error, in fact).
1721
1722 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1723 heapframe that is used only when UTF-8 support is enabled. This caused no
1724 problem, but was untidy.
1725
1726 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1727 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1728 included within another project.
1729
1730 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1731 slightly modified by me:
1732
1733 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1734 not building pcregrep.
1735
1736 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1737 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1738
1739 25. Forward references, both numeric and by name, in patterns that made use of
1740 duplicate group numbers, could behave incorrectly or give incorrect errors,
1741 because when scanning forward to find the reference group, PCRE was not
1742 taking into account the duplicate group numbers. A pattern such as
1743 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1744
1745 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1746 the feature of a custom pointer more persuasive (as requested by a user).
1747
1748 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1749 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1750 pre-defined.
1751
1752 28. Added support for (*UTF8) at the start of a pattern.
1753
1754 29. Arrange for flags added by the "release type" setting in CMake to be shown
1755 in the configuration summary.
1756
1757
1758 Version 7.8 05-Sep-08
1759 ---------------------
1760
1761 1. Replaced UCP searching code with optimized version as implemented for Ad
1762 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1763 stage table and inline lookup instead of a function, giving speed ups of 2
1764 to 5 times on some simple patterns that I tested. Permission was given to
1765 distribute the MultiStage2.py script that generates the tables (it's not in
1766 the tarball, but is in the Subversion repository).
1767
1768 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1769 scripts.
1770
1771 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1772 a group with a zero qualifier. The result of the study could be incorrect,
1773 or the function might crash, depending on the pattern.
1774
1775 4. Caseless matching was not working for non-ASCII characters in back
1776 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1777 It now works when Unicode Property Support is available.
1778
1779 5. In pcretest, an escape such as \x{de} in the data was always generating
1780 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1781 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1782 truncation.
1783
1784 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1785
1786 7. Added two (int) casts to pcregrep when printing the difference of two
1787 pointers, in case they are 64-bit values.
1788
1789 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1790 test 2 if it fails.
1791
1792 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1793 and a #define of that name to empty if it is not externally set. This is to
1794 allow users of MSVC to set it if necessary.
1795
1796 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1797 the convenience functions in the pcre_get.c source file.
1798
1799 11. An option change at the start of a pattern that had top-level alternatives
1800 could cause overwriting and/or a crash. This command provoked a crash in
1801 some environments:
1802
1803 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1804
1805 This potential security problem was recorded as CVE-2008-2371.
1806
1807 12. For a pattern where the match had to start at the beginning or immediately
1808 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1809 pcre_dfa_exec() could read past the end of the passed subject if there was
1810 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1811 pcretest so that it places the subject at the end of its malloc-ed buffer.
1812
1813 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1814 exec() might read past the end of the data buffer in UTF-8 mode.
1815
1816 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1817 the data contained the byte 0x85 as part of a UTF-8 character within its
1818 first line. This applied both to normal and DFA matching.
1819
1820 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1821 /^[^d]*?$/8 failed to match "abc".
1822
1823 16. Added a missing copyright notice to pcrecpp_internal.h.
1824
1825 17. Make it more clear in the documentation that values returned from
1826 pcre_exec() in ovector are byte offsets, not character counts.
1827
1828 18. Tidied a few places to stop certain compilers from issuing warnings.
1829
1830 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1831 supplied by Stefan Weber. I made a further small update for 7.8 because
1832 there is a change of source arrangements: the pcre_searchfuncs.c module is
1833 replaced by pcre_ucd.c.
1834
1835
1836 Version 7.7 07-May-08
1837 ---------------------
1838
1839 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1840 a string to a long long, pretend we don't even have a long long." This is
1841 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1842
1843 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1844 pre-7.6 versions, which defined a global no_arg variable instead of putting
1845 it in the RE class. (See also #8 below.)
1846
1847 3. Remove a line of dead code, identified by coverity and reported by Nuno
1848 Lopes.
1849
1850 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1851
1852 (1) The include/exclude patterns were being applied to the whole pathnames
1853 of files, instead of just to the final components.
1854
1855 (2) If there was more than one level of directory, the subdirectories were
1856 skipped unless they satisfied the include/exclude conditions. This is
1857 inconsistent with GNU grep (and could even be seen as contrary to the
1858 pcregrep specification - which I improved to make it absolutely clear).
1859 The action now is always to scan all levels of directory, and just
1860 apply the include/exclude patterns to regular files.
1861
1862 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1863 --exclude_dir in the tests to avoid scanning .svn directories.
1864
1865 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1866 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1867 doesn't support NULs in patterns.
1868
1869 7. Added some missing "const"s to declarations of static tables in
1870 pcre_compile.c and pcre_dfa_exec.c.
1871
1872 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1873 caused by fix #2 above. (Subsequently also a second patch to fix the
1874 first patch. And a third patch - this was a messy problem.)
1875
1876 9. Applied Craig's patch to remove the use of push_back().
1877
1878 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1879 matching function regexec().
1880
1881 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1882 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1883 references. PCRE supports relative numbers with this syntax (I don't think
1884 Oniguruma does).
1885
1886 12. Previously, a group with a zero repeat such as (...){0} was completely
1887 omitted from the compiled regex. However, this means that if the group
1888 was called as a subroutine from elsewhere in the pattern, things went wrong
1889 (an internal error was given). Such groups are now left in the compiled
1890 pattern, with a new opcode that causes them to be skipped at execution
1891 time.
1892
1893 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1894 to the way PCRE behaves:
1895
1896 (a) A lone ] character is dis-allowed (Perl treats it as data).
1897
1898 (b) A back reference to an unmatched subpattern matches an empty string
1899 (Perl fails the current match path).
1900
1901 (c) A data ] in a character class must be notated as \] because if the
1902 first data character in a class is ], it defines an empty class. (In
1903 Perl it is not possible to have an empty class.) The empty class []
1904 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1905 The negative empty class [^] matches any one character, independently
1906 of the DOTALL setting.
1907
1908 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1909 non-existent subpattern following a character class starting with ']' and
1910 containing () gave an internal compiling error instead of "reference to
1911 non-existent subpattern". Fortunately, when the pattern did exist, the
1912 compiled code was correct. (When scanning forwards to check for the
1913 existence of the subpattern, it was treating the data ']' as terminating
1914 the class, so got the count wrong. When actually compiling, the reference
1915 was subsequently set up correctly.)
1916
1917 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1918 it was being rejected as not supported by pcre_dfa_exec(), even though
1919 other assertions are supported. I have made pcre_dfa_exec() support
1920 (*FAIL).
1921
1922 16. The implementation of 13c above involved the invention of a new opcode,
1923 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1924 cannot be changed at match time, I realized I could make a small
1925 improvement to matching performance by compiling OP_ALLANY instead of
1926 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1927 on the OP_ANY path.
1928
1929 17. Compiling pcretest on Windows with readline support failed without the
1930 following two fixes: (1) Make the unistd.h include conditional on
1931 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1932
1933 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1934 ncurses library to be included for pcretest when ReadLine support is
1935 requested, but also to allow for it to be overridden. This patch came from
1936 Daniel Bergström.
1937
1938 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1939 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1940 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1941 spotting this.
1942
1943
1944 Version 7.6 28-Jan-08
1945 ---------------------
1946
1947 1. A character class containing a very large number of characters with
1948 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1949 overflow.
1950
1951 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1952 HAVE_LONG_LONG is not defined.
1953
1954 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1955 bring them up to date and include new features. This patch includes:
1956
1957 - Fixed PH's badly added libz and libbz2 support.
1958 - Fixed a problem with static linking.
1959 - Added pcredemo. [But later removed - see 7 below.]
1960 - Fixed dftables problem and added an option.
1961 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1962 HAVE_LONG_LONG.
1963 - Added readline support for pcretest.
1964 - Added an listing of the option settings after cmake has run.
1965
1966 4. A user submitted a patch to Makefile that makes it easy to create
1967 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1968 Makefile.am that cause it to include this special target, without
1969 affecting anything else. Note that the same mingw target plus all
1970 the other distribution libraries and programs are now supported
1971 when configuring with CMake (see 6 below) instead of with
1972 Configure/Make.
1973
1974 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1975 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1976 exported in the Windows port". It has not yet been confirmed that the patch
1977 solves the problem, but it does no harm.
1978
1979 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1980 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1981 with CMake, and also correct the comment about stack recursion.
1982
1983 7. Remove the automatic building of pcredemo from the ./configure system and
1984 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1985 of a program that users should build themselves after PCRE is installed, so
1986 building it automatically is not really right. What is more, it gave
1987 trouble in some build environments.
1988
1989 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1990
1991
1992 Version 7.5 10-Jan-08
1993 ---------------------
1994
1995 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1996 values in parens when parsing an RE using the C++ wrapper."
1997
1998 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1999 Characters greater than 255 were excluded from the class instead of being
2000 included.
2001
2002 3. The same bug as (2) above applied to negated POSIX classes such as
2003 [:^space:].
2004
2005 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2006 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2007 I have changed it.
2008
2009 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2010 first named subpattern) and a construct such as (?&a) would reference the
2011 first named subpattern whose name started with "a" (in other words, the
2012 length check was missing). Both these problems are fixed. "Subpattern name
2013 expected" is now given for (?&) (a zero-length name), and this patch also
2014 makes it give the same error for \k'' (previously it complained that that
2015 was a reference to a non-existent subpattern).
2016
2017 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2018 this is right because (?- can be followed by option settings as well as by
2019 digits. I have, however, made the messages clearer.
2020
2021 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2022 than the number used in the conditional) now cause a compile-time error.
2023 This is actually not compatible with Perl, which accepts such patterns, but
2024 treats the conditional as always being FALSE (as PCRE used to), but it
2025 seems to me that giving a diagnostic is better.
2026
2027 8. Change "alphameric" to the more common word "alphanumeric" in comments
2028 and messages.
2029
2030 9. Fix two occurrences of "backslash" in comments that should have been
2031 "backspace".
2032
2033 10. Remove two redundant lines of code that can never be obeyed (their function
2034 was moved elsewhere).
2035
2036 11. The program that makes PCRE's Unicode character property table had a bug
2037 which caused it to generate incorrect table entries for sequences of
2038 characters that have the same character type, but are in different scripts.
2039 It amalgamated them into a single range, with the script of the first of
2040 them. In other words, some characters were in the wrong script. There were
2041 thirteen such cases, affecting characters in the following ranges:
2042
2043 U+002b0 - U+002c1
2044 U+0060c - U+0060d
2045 U+0061e - U+00612
2046 U+0064b - U+0065e
2047 U+0074d - U+0076d
2048 U+01800 - U+01805
2049 U+01d00 - U+01d77
2050 U+01d9b - U+01dbf
2051 U+0200b - U+0200f
2052 U+030fc - U+030fe
2053 U+03260 - U+0327f
2054 U+0fb46 - U+0fbb1
2055 U+10450 - U+1049d
2056
2057 12. The -o option (show only the matching part of a line) for pcregrep was not
2058 compatible with GNU grep in that, if there was more than one match in a
2059 line, it showed only the first of them. It now behaves in the same way as
2060 GNU grep.
2061
2062 13. If the -o and -v options were combined for pcregrep, it printed a blank
2063 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2064 does the same. The return code can be used to tell if there were any
2065 non-matching lines.
2066
2067 14. Added --file-offsets and --line-offsets to pcregrep.
2068
2069 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2070 infinitely looping recursion. The bug was that positive lookaheads were not
2071 being skipped when checking for a possible empty match (negative lookaheads
2072 and both kinds of lookbehind were skipped).
2073
2074 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2075 inclusion of <windows.h> to before rather than after the definition of
2076 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2077
2078 17. Specifying a possessive quantifier with a specific limit for a Unicode
2079 character property caused pcre_compile() to compile bad code, which led at
2080 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2081 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2082 caused the error; without that there was no problem.
2083
2084 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2085
2086 19. Added --enable-pcretest-libreadline.
2087
2088 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2089 RE::GlobalReplace(). As a result, the number of replacements returned was
2090 double what it should be. I removed one of the increments, but Craig sent a
2091 later patch that removed the other one (the right fix) and added unit tests
2092 that check the return values (which was not done before).
2093
2094 21. Several CMake things:
2095
2096 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2097 the names libpcre and libpcreposix, not just pcre and pcreposix.
2098
2099 (2) The above change means that pcretest and pcregrep are now correctly
2100 linked with the newly-built libraries, not previously installed ones.
2101
2102 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2103
2104 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2105 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2106 UTF-8 newline character). The key issue is that the pattern starts .*;
2107 this means that the match must be either at the beginning, or after a
2108 newline. The bug was in the code for advancing after a failed match and
2109 checking that the new position followed a newline. It was not taking
2110 account of UTF-8 characters correctly.
2111
2112 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2113 character classes. PCRE was not treating the sequence [:...:] as a
2114 character class unless the ... were all letters. Perl, however, seems to
2115 allow any characters between [: and :], though of course it rejects as
2116 unknown any "names" that contain non-letters, because all the known class
2117 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2118 for example, whereas PCRE did not - it did not recognize a POSIX character
2119 class. This seemed a bit dangerous, so the code has been changed to be
2120 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2121 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2122 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2123 Perl does, and where it didn't before.
2124
2125 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2126 Windows environments %n is disabled by default.
2127
2128
2129 Version 7.4 21-Sep-07
2130 ---------------------
2131
2132 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2133 means that a class such as [\s] counted as "explicit reference to CR or
2134 LF". That isn't really right - the whole point of the change was to try to
2135 help when there was an actual mention of one of the two characters. So now
2136 the change happens only if \r or \n (or a literal CR or LF) character is
2137 encountered.
2138
2139 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2140 of both had grown to the point where there were only 3 bits left.
2141 Fortunately, there was spare space in the data structure, and so I have
2142 moved the internal flags into a new 16-bit field to free up more option
2143 bits.
2144
2145 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2146 but did not set the internal JCHANGED flag - either of these is enough to
2147 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2148 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2149 start sets both bits.
2150
2151 4. Added options (at build time, compile time, exec time) to change \R from
2152 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2153
2154 5. doc/pcresyntax.html was missing from the distribution.
2155
2156 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2157 compatibility, even though it is no longer used.
2158
2159 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2160 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2161 windows.h file is present (where different names are used). [This was
2162 reversed later after testing - see 16 below.]
2163
2164 8. Changed all #include <config.h> to #include "config.h". There were also
2165 some further <pcre.h> cases that I changed to "pcre.h".
2166
2167 9. When pcregrep was used with the --colour option, it missed the line ending
2168 sequence off the lines that it output.
2169
2170 10. It was pointed out to me that arrays of string pointers cause lots of
2171 relocations when a shared library is dynamically loaded. A technique of
2172 using a single long string with a table of offsets can drastically reduce
2173 these. I have refactored PCRE in four places to do this. The result is
2174 dramatic:
2175
2176 Originally: 290
2177 After changing UCP table: 187
2178 After changing error message table: 43
2179 After changing table of "verbs" 36
2180 After changing table of Posix names 22
2181
2182 Thanks to the folks working on Gregex for glib for this insight.
2183
2184 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2185 unicode-properties was also set.
2186
2187 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2188
2189 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2190 checked only for CRLF.
2191
2192 14. Added casts to pcretest.c to avoid compiler warnings.
2193
2194 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2195
2196 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2197 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2198 entirely. This removes changes made in 7 above.
2199
2200 17. The CMake files have been updated, and there is now more information about
2201 building with CMake in the NON-UNIX-USE document.
2202
2203
2204 Version 7.3 28-Aug-07
2205 ---------------------
2206
2207 1. In the rejigging of the build system that eventually resulted in 7.1, the
2208 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2209 brackets there is not right, since it causes compilers to look for an
2210 installed pcre.h, not the version that is in the source that is being
2211 compiled (which of course may be different). I have changed it back to:
2212
2213 #include "pcre.h"
2214
2215 I have a vague recollection that the change was concerned with compiling in
2216 different directories, but in the new build system, that is taken care of
2217 by the VPATH setting the Makefile.
2218
2219 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2220 when the subject happened to end in the byte 0x85 (e.g. if the last
2221 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2222 characters but of course it shouldn't be taken as a newline when it is part
2223 of another character. The bug was that, for an unlimited repeat of . in
2224 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2225 characters when looking for a newline.
2226
2227 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2228
2229 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2230 in debug output.
2231
2232 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2233 long printing in the pcrecpp unittest when running under MinGW.
2234
2235 6. ESC_K was left out of the EBCDIC table.
2236
2237 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2238 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2239 limit also applies to "virtual nesting" when a pattern is recursive, and in
2240 this case 1000 isn't so big. I have been able to remove this limit at the
2241 expense of backing off one optimization in certain circumstances. Normally,
2242 when pcre_exec() would call its internal match() function recursively and
2243 immediately return the result unconditionally, it uses a "tail recursion"
2244 feature to save stack. However, when a subpattern that can match an empty
2245 string has an unlimited repetition quantifier, it no longer makes this
2246 optimization. That gives it a stack frame in which to save the data for
2247 checking that an empty string has been matched. Previously this was taken
2248 from the 1000-entry workspace that had been reserved. So now there is no
2249 explicit limit, but more stack is used.
2250
2251 8. Applied Daniel's patches to solve problems with the import/export magic
2252 syntax that is required for Windows, and which was going wrong for the
2253 pcreposix and pcrecpp parts of the library. These were overlooked when this
2254 problem was solved for the main library.
2255
2256 9. There were some crude static tests to avoid integer overflow when computing
2257 the size of patterns that contain repeated groups with explicit upper
2258 limits. As the maximum quantifier is 65535, the maximum group length was
2259 set at 30,000 so that the product of these two numbers did not overflow a
2260 32-bit integer. However, it turns out that people want to use groups that
2261 are longer than 30,000 bytes (though not repeat them that many times).
2262 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2263 made it possible to implement the integer overflow checks in a much more
2264 dynamic way, which I have now done. The artificial limitation on group
2265 length has been removed - we now have only the limit on the total length of
2266 the compiled pattern, which depends on the LINK_SIZE setting.
2267
2268 10. Fixed a bug in the documentation for get/copy named substring when
2269 duplicate names are permitted. If none of the named substrings are set, the
2270 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2271 empty string.
2272
2273 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2274 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2275 because the ] is interpreted as the first data character and the
2276 terminating ] is not found. PCRE has been made compatible with Perl in this
2277 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2278 cause memory overwriting.
2279
2280 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2281 string has been matched (to stop an infinite loop). It was not recognizing
2282 a conditional subpattern that could match an empty string if that
2283 subpattern was within another subpattern. For example, it looped when
2284 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2285 condition was not nested. This bug has been fixed.
2286
2287 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2288 past the start of the subject in the presence of bytes with the top bit
2289 set, for example "\x8aBCD".
2290
2291 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2292 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2293
2294 14. Optimized (?!) to (*FAIL).
2295
2296 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2297 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2298 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2299 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2300 does: it's just the validity check that is more restrictive.
2301
2302 16. Inserted checks for integer overflows during escape sequence (backslash)
2303 processing, and also fixed erroneous offset values for syntax errors during
2304 backslash processing.
2305
2306 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2307 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2308
2309 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2310 caused an overrun.
2311
2312 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2313 something other than just ASCII characters) inside a group that had an
2314 unlimited repeat caused a loop at compile time (while checking to see
2315 whether the group could match an empty string).
2316
2317 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2318 [\P{Any}] did so. (Error in the code for printing property names.)
2319
2320 21. An orphan \E inside a character class could cause a crash.
2321
2322 22. A repeated capturing bracket such as (A)? could cause a wild memory
2323 reference during compilation.
2324
2325 23. There are several functions in pcre_compile() that scan along a compiled
2326 expression for various reasons (e.g. to see if it's fixed length for look
2327 behind). There were bugs in these functions when a repeated \p or \P was
2328 present in the pattern. These operators have additional parameters compared
2329 with \d, etc, and these were not being taken into account when moving along
2330 the compiled data. Specifically:
2331
2332 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2333 length.
2334
2335 (b) An item such as \pL+ within a repeated group could cause crashes or
2336 loops.
2337
2338 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2339 "reference to non-existent subpattern" error.
2340
2341 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2342
2343 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2344 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2345
2346 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2347
2348 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2349 character were causing crashes (broken optimization).
2350
2351 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2352 \p or \P) caused a compile-time loop.
2353
2354 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2355 break. For example, the unstudied pattern [\r\n]A does not match the string
2356 "\r\nA" because change 7.0/46 below moves the current point on by two
2357 characters after failing to match at the start. However, the pattern \nA
2358 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2359 the same is true. There doesn't seem any very clean way out of this, but
2360 what I have chosen to do makes the common cases work: PCRE now takes note
2361 of whether there can be an explicit match for \r or \n anywhere in the
2362 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2363 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2364 pattern has explicit CR or LF references.
2365
2366 29. Added (*CR) etc for changing newline setting at start of pattern.
2367
2368
2369 Version 7.2 19-Jun-07
2370 ---------------------
2371
2372 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2373 which is apparently normally available under Windows.
2374
2375 2. Re-jig the pcregrep tests with different newline settings in an attempt
2376 to make them independent of the local environment's newline setting.
2377
2378 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2379
2380 4. Some of the "internals" tests were previously cut out when the link size
2381 was not 2, because the output contained actual offsets. The recent new
2382 "Z" feature of pcretest means that these can be cut out, making the tests
2383 usable with all link sizes.
2384
2385 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2386 stack recursion. This gives a massive performance boost under BSD, but just
2387 a small improvement under Linux. However, it saves one field in the frame
2388 in all cases.
2389
2390 6. Added more features from the forthcoming Perl 5.10:
2391
2392 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2393 recursion call. It refers to the nth most recently opened parentheses.
2394
2395 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2396 to be opened parentheses.
2397
2398 (c) Conditions that refer to capturing parentheses can be specified
2399 relatively, for example, (?(-2)... or (?(+3)...
2400
2401 (d) \K resets the start of the current match so that everything before
2402 is not part of it.
2403
2404 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2405
2406 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2407 reference syntax.
2408
2409 (g) (?| introduces a group in which the numbering of parentheses in each
2410 alternative starts with the same number.
2411
2412 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2413
2414 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2415 PCRE_INFO_JCHANGED.
2416
2417 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2418 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2419 for detecting groups that can match an empty string.
2420
2421 9. A pattern with a very large number of alternatives (more than several
2422 hundred) was running out of internal workspace during the pre-compile
2423 phase, where pcre_compile() figures out how much memory will be needed. A
2424 bit of new cunning has reduced the workspace needed for groups with
2425 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2426 workspace instead of running out of the 4096 that are available.
2427
2428 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2429
2430 11. Applied patch from Google to remove an optimization that didn't quite work.
2431 The report of the bug said:
2432
2433 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2434 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2435 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2436
2437 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2438 it matched the wrong number of bytes.
2439
2440
2441 Version 7.1 24-Apr-07
2442 ---------------------
2443
2444 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2445 that is more "standard", making use of automake and other Autotools. There
2446 is some re-arrangement of the files and adjustment of comments consequent
2447 on this.
2448
2449 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2450 for recursive directory scanning broke on some systems because the files
2451 are not scanned in any specific order and on different systems the order
2452 was different. A call to "sort" has been inserted into RunGrepTest for the
2453 approprate test as a short-term fix. In the longer term there may be an
2454 alternative.
2455
2456 3. I had an email from Eric Raymond about problems translating some of PCRE's
2457 man pages to HTML (despite the fact that I distribute HTML pages, some
2458 people do their own conversions for various reasons). The problems
2459 concerned the use of low-level troff macros .br and .in. I have therefore
2460 removed all such uses from the man pages (some were redundant, some could
2461 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2462 HTML has been updated to handle .nf/.fi and to complain if it encounters
2463 .br or .in.
2464
2465 4. Updated comments in configure.ac that get placed in config.h.in and also
2466 arranged for config.h to be included in the distribution, with the name
2467 config.h.generic, for the benefit of those who have to compile without
2468 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2469
2470 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2471 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2472 makevp.bat for the current PCRE, using the additional files
2473 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2474
2475 6. A Windows user reported a minor discrepancy with test 2, which turned out
2476 to be caused by a trailing space on an input line that had got lost in his
2477 copy. The trailing space was an accident, so I've just removed it.
2478
2479 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2480 that is needed.
2481
2482 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2483 as "const" (a) because they are and (b) because it helps the PHP
2484 maintainers who have recently made a script to detect big data structures
2485 in the php code that should be moved to the .rodata section. I remembered
2486 to update Builducptable as well, so it won't revert if ucptable.h is ever
2487 re-created.
2488
2489 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2490 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2491 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2492 support is not required. This saves 1.5-2K of code, which is important in
2493 some applications.
2494
2495 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2496 so as not to refer to the tables, even though these functions will never be
2497 called when UTF-8 support is disabled. Otherwise there are problems with a
2498 shared library.
2499
2500 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2501
2502 (a) It was defining its arguments as char * instead of void *.
2503
2504 (b) It was assuming that all moves were upwards in memory; this was true
2505 a long time ago when I wrote it, but is no longer the case.
2506
2507 The emulated memove() is provided for those environments that have neither
2508 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2509 is clearly not the case, as these two bugs were recently reported.
2510
2511 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2512 and Detrail to create the HTML documentation, the .txt form of the man
2513 pages, and it removes trailing spaces from listed files. It also creates
2514 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2515 case, it wraps all the #defines with #ifndefs. This script should be run
2516 before "make dist".
2517
2518 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2519 with Unicode property support.
2520
2521 (a) For a maximizing quantifier, if the two different cases of the
2522 character were of different lengths in their UTF-8 codings (there are
2523 some cases like this - I found 11), and the matching function had to
2524 back up over a mixture of the two cases, it incorrectly assumed they
2525 were both the same length.
2526
2527 (b) When PCRE was configured to use the heap rather than the stack for
2528 recursion during matching, it was not correctly preserving the data for
2529 the other case of a UTF-8 character when checking ahead for a match
2530 while processing a minimizing repeat. If the check also involved
2531 matching a wide character, but failed, corruption could cause an
2532 erroneous result when trying to check for a repeat of the original
2533 character.
2534
2535 13. Some tidying changes to the testing mechanism:
2536
2537 (a) The RunTest script now detects the internal link size and whether there
2538 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2539 values substituted by "configure". (The RunGrepTest script already did
2540 this for UTF-8.) The configure.ac script no longer substitutes the
2541 relevant variables.
2542
2543 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2544 with length and offset values. This means that the output is different
2545 for different internal link sizes. Test 2 is skipped for link sizes
2546 other than 2 because of this, bypassing the problem. Unfortunately,
2547 there was also a test in test 3 (the locale tests) that used /B and
2548 failed for link sizes other than 2. Rather than cut the whole test out,
2549 I have added a new /Z option to pcretest that replaces the length and
2550 offset values with spaces. This is now used to make test 3 independent
2551 of link size. (Test 2 will be tidied up later.)
2552
2553 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2554 segmentation fault instead of returning the appropriate error message.
2555
2556 15. In multiline mode when the newline sequence was set to "any", the pattern
2557 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2558 This doesn't seem right; it now treats the CRLF combination as the line
2559 ending, and so does not match in that case. It's only a pattern such as ^$
2560 that would hit this one: something like ^ABC$ would have failed after \r
2561 and then tried again after \r\n.
2562
2563 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2564 in an attempt to make files that differ only in their line terminators
2565 compare equal. This works on Linux.
2566
2567 17. Under certain error circumstances pcregrep might try to free random memory
2568 as it exited. This is now fixed, thanks to valgrind.
2569
2570 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2571 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2572 was because its rules for how to advance for /g after matching an empty
2573 string at the end of a line did not allow for this case. They now check for
2574 it specially.
2575
2576 20. pcretest is supposed to handle patterns and data of any length, by
2577 extending its buffers when necessary. It was getting this wrong when the
2578 buffer for a data line had to be extended.
2579
2580 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2581 CRLF as a newline sequence.
2582
2583 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2584 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2585 I have nevertheless tidied it up.
2586
2587 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2588
2589 24. Added a man page for pcre-config.
2590
2591
2592 Version 7.0 19-Dec-06
2593 ---------------------
2594
2595 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2596 moving to gcc 4.1.1.
2597
2598 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2599 sys/time.h, which is documented as needed for this function. It doesn't
2600 seem to matter on Linux, but it showed up on some releases of OS X.
2601
2602 3. It seems that there are systems where bytes whose values are greater than
2603 127 match isprint() in the "C" locale. The "C" locale should be the
2604 default when a C program starts up. In most systems, only ASCII printing
2605 characters match isprint(). This difference caused the output from pcretest
2606 to vary, making some of the tests fail. I have changed pcretest so that:
2607
2608 (a) When it is outputting text in the compiled version of a pattern, bytes
2609 other than 32-126 are always shown as hex escapes.
2610
2611 (b) When it is outputting text that is a matched part of a subject string,
2612 it does the same, unless a different locale has been set for the match
2613 (using the /L modifier). In this case, it uses isprint() to decide.
2614
2615 4. Fixed a major bug that caused incorrect computation of the amount of memory
2616 required for a compiled pattern when options that changed within the
2617 pattern affected the logic of the preliminary scan that determines the
2618 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2619 that the computed length was too small. The symptoms of this bug were
2620 either the PCRE error "internal error: code overflow" from pcre_compile(),
2621 or a glibc crash with a message such as "pcretest: free(): invalid next
2622 size (fast)". Examples of patterns that provoked this bug (shown in
2623 pcretest format) are:
2624
2625 /(?-x: )/x
2626 /(?x)(?-x: \s*#\s*)/
2627 /((?i)[\x{c0}])/8
2628 /(?i:[\x{c0}])/8
2629
2630 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2631 is now done differently.
2632
2633 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2634 wrapper classes; (b) implement a new function in the C++ scanner that is
2635 more efficient than the old way of doing things because it avoids levels of
2636 recursion in the regex matching; (c) add a paragraph to the documentation
2637 for the FullMatch() function.
2638
2639 6. The escape sequence \n was being treated as whatever was defined as
2640 "newline". Not only was this contrary to the documentation, which states
2641 that \n is character 10 (hex 0A), but it also went horribly wrong when
2642 "newline" was defined as CRLF. This has been fixed.
2643
2644 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2645 was being set to -1 for the "end of line" case (supposedly a value that no
2646 character can have). Though this value is never used (the check for end of
2647 line is "zero bytes in current character"), it caused compiler complaints.
2648 I've changed it to 0xffffffff.
2649
2650 8. In pcre_version.c, the version string was being built by a sequence of
2651 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2652 string (as it is for production releases) called a macro with an empty
2653 argument. The C standard says the result of this is undefined. The gcc
2654 compiler treats it as an empty string (which was what was wanted) but it is
2655 reported that Visual C gives an error. The source has been hacked around to
2656 avoid this problem.
2657
2658 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2659 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2660 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2661 of them did).
2662
2663 10. Originally, pcretest opened its input and output without "b"; then I was
2664 told that "b" was needed in some environments, so it was added for release
2665 5.0 to both the input and output. (It makes no difference on Unix-like
2666 systems.) Later I was told that it is wrong for the input on Windows. I've
2667 now abstracted the modes into two macros, to make it easier to fiddle with
2668 them, and removed "b" from the input mode under Windows.
2669
2670 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2671
2672 12. Added -help and --help to pcretest as an official way of being reminded
2673 of the options.
2674
2675 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2676 and pcrecpp.cc because they annoy compilers at high warning levels.
2677
2678 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2679
2680 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2681 scripts are not C programs :-) and which was not noticed because it works
2682 on Linux.
2683
2684 16. pcretest is supposed to handle any length of pattern and data line (as one
2685 line or as a continued sequence of lines) by extending its input buffer if
2686 necessary. This feature was broken for very long pattern lines, leading to
2687 a string of junk being passed to pcre_compile() if the pattern was longer
2688 than about 50K.
2689
2690 17. I have done a major re-factoring of the way pcre_compile() computes the
2691 amount of memory needed for a compiled pattern. Previously, there was code
2692 that made a preliminary scan of the pattern in order to do this. That was
2693 OK when PCRE was new, but as the facilities have expanded, it has become
2694 harder and harder to keep it in step with the real compile phase, and there
2695 have been a number of bugs (see for example, 4 above). I have now found a
2696 cunning way of running the real compile function in a "fake" mode that
2697 enables it to compute how much memory it would need, while actually only
2698 ever using a few hundred bytes of working memory and without too many
2699 tests of the mode. This should make future maintenance and development
2700 easier. A side effect of this work is that the limit of 200 on the nesting
2701 depth of parentheses has been removed (though this was never a serious
2702 limitation, I suspect). However, there is a downside: pcre_compile() now
2703 runs more slowly than before (30% or more, depending on the pattern). I
2704 hope this isn't a big issue. There is no effect on runtime performance.
2705
2706 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2707 newline (only possible for the last line of a file) and it was a
2708 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2709
2710 19. Added additional timing features to pcretest. (1) The -tm option now times
2711 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2712 separate command line item, by a number that specifies the number of
2713 repeats to use when timing. The default is 50000; this gives better
2714 precision, but takes uncomfortably long for very large patterns.
2715
2716 20. Extended pcre_study() to be more clever in cases where a branch of a
2717 subpattern has no definite first character. For example, (a*|b*)[cd] would
2718 previously give no result from pcre_study(). Now it recognizes that the
2719 first character must be a, b, c, or d.
2720
2721 21. There was an incorrect error "recursive call could loop indefinitely" if
2722 a subpattern (or the entire pattern) that was being tested for matching an
2723 empty string contained only one non-empty item after a nested subpattern.
2724 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2725 incorrectly, because the \d was being skipped in the check.
2726
2727 22. The pcretest program now has a new pattern option /B and a command line
2728 option -b, which is equivalent to adding /B to every pattern. This causes
2729 it to show the compiled bytecode, without the additional information that
2730 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2731 is the same as /B/I).
2732
2733 23. A new optimization is now able automatically to treat some sequences such
2734 as a*b as a*+b. More specifically, if something simple (such as a character
2735 or a simple class like \d) has an unlimited quantifier, and is followed by
2736 something that cannot possibly match the quantified thing, the quantifier
2737 is automatically "possessified".
2738
2739 24. A recursive reference to a subpattern whose number was greater than 39
2740 went wrong under certain circumstances in UTF-8 mode. This bug could also
2741 have affected the operation of pcre_study().
2742
2743 25. Realized that a little bit of performance could be had by replacing
2744 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2745
2746 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2747
2748 27. Possessive quantifiers such as a++ were previously implemented by turning
2749 them into atomic groups such as ($>a+). Now they have their own opcodes,
2750 which improves performance. This includes the automatically created ones
2751 from 23 above.
2752
2753 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2754 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2755 the first matched character to be a colon. This applied both to named and
2756 numbered groups.
2757
2758 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2759
2760 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2761 building PCRE on Windows easier, so I have included it in the distribution.
2762
2763 31. There is now a check in pcretest against a ridiculously large number being
2764 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2765 loop, the loop is abandoned.
2766
2767 32. Forward references to subpatterns in conditions such as (?(2)...) where
2768 subpattern 2 is defined later cause pcre_compile() to search forwards in
2769 the pattern for the relevant set of parentheses. This search went wrong
2770 when there were unescaped parentheses in a character class, parentheses
2771 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2772
2773 33. "Subroutine" calls and backreferences were previously restricted to
2774 referencing subpatterns earlier in the regex. This restriction has now
2775 been removed.
2776
2777 34. Added a number of extra features that are going to be in Perl 5.10. On the
2778 whole, these are just syntactic alternatives for features that PCRE had
2779 previously implemented using the Python syntax or my own invention. The
2780 other formats are all retained for compatibility.
2781
2782 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2783 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2784 also .NET compatible.
2785
2786 (b) A recursion or subroutine call to a named group can now be defined as
2787 (?&name) as well as (?P>name).
2788
2789 (c) A backreference to a named group can now be defined as \k<name> or
2790 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2791 5.10, are also .NET compatible.
2792
2793 (d) A conditional reference to a named group can now use the syntax
2794 (?(<name>) or (?('name') as well as (?(name).
2795
2796 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2797 groups (named and numbered) that are never evaluated inline, but can be
2798 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2799 is always false. There may be only one alternative in such a group.
2800
2801 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2802 as the simple (?(R). The condition is true only if the most recent
2803 recursion is that of the given number or name. It does not search out
2804 through the entire recursion stack.
2805
2806 (g) The escape \gN or \g{N} has been added, where N is a positive or
2807 negative number, specifying an absolute or relative reference.
2808
2809 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2810 some "unreachable code" warnings.
2811
2812 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2813 things, this adds five new scripts.
2814
2815 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2816 There were also incompatibilities regarding the handling of \Q..\E inside
2817 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2818 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2819
2820 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2821 matches an empty string, and forcibly breaks the loop. There were bugs in
2822 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2823 against aaaa the result was just "a" rather than "aaaa", for example. Two
2824 separate and independent bugs (that affected different cases) have been
2825 fixed.
2826
2827 39. Refactored the code to abolish the use of different opcodes for small
2828 capturing bracket numbers. This is a tidy that I avoided doing when I
2829 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2830 The new approach is not only tidier, it makes it possible to reduce the
2831 memory needed to fix the previous bug (38).
2832
2833 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2834 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2835 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2836 mode.
2837
2838 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2839 report.
2840
2841 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2842 copy construction and assignment in the C++ wrapper.
2843
2844 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2845 couple of bugs that could have given wrong results in the "--newline=crlf"
2846 case.
2847
2848 44. Added a number of casts and did some reorganization of signed/unsigned int
2849 variables following suggestions from Dair Grant. Also renamed the variable
2850 "this" as "item" because it is a C++ keyword.
2851
2852 45. Arranged for dftables to add
2853
2854 #include "pcre_internal.h"
2855
2856 to pcre_chartables.c because without it, gcc 4.x may remove the array
2857 definition from the final binary if PCRE is built into a static library and
2858 dead code stripping is activated.
2859
2860 46. For an unanchored pattern, if a match attempt fails at the start of a
2861 newline sequence, and the newline setting is CRLF or ANY, and the next two
2862 characters are CRLF, advance by two characters instead of one.
2863
2864
2865 Version 6.7 04-Jul-06
2866 ---------------------
2867
2868 1. In order to handle tests when input lines are enormously long, pcretest has
2869 been re-factored so that it automatically extends its buffers when
2870 necessary. The code is crude, but this _is_ just a test program. The
2871 default size has been increased from 32K to 50K.
2872
2873 2. The code in pcre_study() was using the value of the re argument before
2874 testing it for NULL. (Of course, in any sensible call of the function, it
2875 won't be NULL.)
2876
2877 3. The memmove() emulation function in pcre_internal.h, which is used on
2878 systems that lack both memmove() and bcopy() - that is, hardly ever -
2879 was missing a "static" storage class specifier.
2880
2881 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2882 containing an extended class (one that cannot be represented by a bitmap
2883 because it contains high-valued characters or Unicode property items, e.g.
2884 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2885 pattern, but PCRE should not loop if you do not (it no longer does).
2886 [Detail: two cases were found: (a) a repeated subpattern containing an
2887 extended class; (b) a recursive reference to a subpattern that followed a
2888 previous extended class. It wasn't skipping over the extended class
2889 correctly when UTF-8 mode was not set.]
2890
2891 5. A negated single-character class was not being recognized as fixed-length
2892 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2893 compile error "lookbehind assertion is not fixed length".
2894
2895 6. The RunPerlTest auxiliary script was showing an unexpected difference
2896 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2897 write a Perl script that can interpret lines of an input file either as
2898 byte characters or as UTF-8, which is what "perltest" was being required to
2899 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2900 can't do is switch easily at run time between having the "use utf8;" pragma
2901 or not. In the end, I fudged it by using the RunPerlTest script to insert
2902 "use utf8;" explicitly for the UTF-8 tests.
2903
2904 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2905 the end of the subject string, contrary to the documentation and to what
2906 Perl does. This was true of both matching functions. Now it matches only at
2907 the start of the subject and immediately after *internal* newlines.
2908
2909 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2910 a pointer to an int instead of a pointer to an unsigned long int. This
2911 caused problems on 64-bit systems.
2912
2913 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2914 instance of the 'standard' template library not being so standard".
2915
2916 10. There was no check on the number of named subpatterns nor the maximum
2917 length of a subpattern name. The product of these values is used to compute
2918 the size of the memory block for a compiled pattern. By supplying a very
2919 long subpattern name and a large number of named subpatterns, the size
2920 computation could be caused to overflow. This is now prevented by limiting
2921 the length of names to 32 characters, and the number of named subpatterns
2922 to 10,000.
2923
2924 11. Subpatterns that are repeated with specific counts have to be replicated in
2925 the compiled pattern. The size of memory for this was computed from the
2926 length of the subpattern and the repeat count. The latter is limited to
2927 65535, but there was no limit on the former, meaning that integer overflow
2928 could in principle occur. The compiled length of a repeated subpattern is
2929 now limited to 30,000 bytes in order to prevent this.
2930
2931 12. Added the optional facility to have named substrings with the same name.
2932
2933 13. Added the ability to use a named substring as a condition, using the
2934 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2935 are numbers (not recommended). Forward references are permitted.
2936
2937 14. Added forward references in named backreferences (if you see what I mean).
2938
2939 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2940 pattern could run off the end of the subject. For example, the pattern
2941 "(?s)(.{1,5})"8 did this with the subject "ab".
2942
2943 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2944 PCRE_CASELESS was set when matching characters that were quantified with ?
2945 or *.
2946
2947 17. A character class other than a single negated character that had a minimum
2948 but no maximum quantifier - for example [ab]{6,} - was not handled
2949 correctly by pce_dfa_exec(). It would match only one character.
2950
2951 18. A valid (though odd) pattern that looked like a POSIX character
2952 class but used an invalid character after [ (for example [[,abc,]]) caused
2953 pcre_compile() to give the error "Failed: internal error: code overflow" or
2954 in some cases to crash with a glibc free() error. This could even happen if
2955 the pattern terminated after [[ but there just happened to be a sequence of
2956 letters, a binary zero, and a closing ] in the memory that followed.
2957
2958 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2959 over the years. Originally (before any Unicode support), just the bottom 8
2960 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2961 output from "man perlunicode" includes this:
2962
2963 The regular expression compiler produces polymorphic opcodes. That
2964 is, the pattern adapts to the data and automatically switches to
2965 the Unicode character scheme when presented with Unicode data--or
2966 instead uses a traditional byte scheme when presented with byte
2967 data.
2968
2969 Sadly, a wide octal escape does not cause a switch, and in a string with
2970 no other multibyte characters, these octal escapes are treated as before.
2971 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2972 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2973 Unicode string.
2974
2975 I have not perpetrated such confusion in PCRE. Up till now, it took just
2976 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2977 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2978 translate to the appropriate multibyte character.
2979
2980 29. Applied some refactoring to reduce the number of warnings from Microsoft
2981 and Borland compilers. This has included removing the fudge introduced
2982 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2983 a warning about an unused variable.
2984
2985 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2986 characters since release 4.0, because Perl (from release 5.004) does not.
2987 [Or at least, is documented not to: some releases seem to be in conflict
2988 with the documentation.] However, when a pattern was studied with
2989 pcre_study() and all its branches started with \s, PCRE still included VT
2990 as a possible starting character. Of course, this did no harm; it just
2991 caused an unnecessary match attempt.
2992
2993 22. Removed a now-redundant internal flag bit that recorded the fact that case
2994 dependency changed within the pattern. This was once needed for "required
2995 byte" processing, but is no longer used. This recovers a now-scarce options
2996 bit. Also moved the least significant internal flag bit to the most-
2997 significant bit of the word, which was not previously used (hangover from
2998 the days when it was an int rather than a uint) to free up another bit for
2999 the future.
3000
3001 23. Added support for CRLF line endings as well as CR and LF. As well as the
3002 default being selectable at build time, it can now be changed at runtime
3003 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3004 specify that it is scanning data with non-default line endings.
3005
3006 24. Changed the definition of CXXLINK to make it agree with the definition of
3007 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3008
3009 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3010 recursions. This makes a big different to stack usage for some patterns.
3011
3012 26. If a subpattern containing a named recursion or subroutine reference such
3013 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3014 the space required for the compiled pattern went wrong and gave too small a
3015 value. Depending on the environment, this could lead to "Failed: internal
3016 error: code overflow at offset 49" or "glibc detected double free or
3017 corruption" errors.
3018
3019 27. Applied patches from Google (a) to support the new newline modes and (b) to
3020 advance over multibyte UTF-8 characters in GlobalReplace.
3021
3022 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3023 difference for some implementation of PCRE in some Windows version.
3024
3025 29. Added some extra testing facilities to pcretest:
3026
3027 \q<number> in a data line sets the "match limit" value
3028 \Q<number> in a data line sets the "match recursion limt" value
3029 -S <number> sets the stack size, where <number> is in megabytes
3030
3031 The -S option isn't available for Windows.
3032
3033
3034 Version 6.6 06-Feb-06
3035 ---------------------
3036
3037 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3038 in pcreposix.h. I have copied the definition from pcre.h.
3039
3040 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3041 because pcre.h is no longer a built file.
3042
3043 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3044 not normally included in the compiled code.
3045
3046
3047 Version 6.5 01-Feb-06
3048 ---------------------
3049
3050 1. When using the partial match feature with pcre_dfa_exec(), it was not
3051 anchoring the second and subsequent partial matches at the new starting
3052 point. This could lead to incorrect results. For example, with the pattern
3053 /1234/, partially matching against "123" and then "a4" gave a match.
3054
3055 2. Changes to pcregrep:
3056
3057 (a) All non-match returns from pcre_exec() were being treated as failures
3058 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3059 error message is output. Some extra information is given for the
3060 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3061 probably the only errors that are likely to be caused by users (by
3062 specifying a regex that has nested indefinite repeats, for instance).
3063 If there are more than 20 of these errors, pcregrep is abandoned.
3064
3065 (b) A binary zero was treated as data while matching, but terminated the
3066 output line if it was written out. This has been fixed: binary zeroes
3067 are now no different to any other data bytes.
3068
3069 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3070 used to set a locale for matching. The --locale=xxxx long option has
3071 been added (no short equivalent) to specify a locale explicitly on the
3072 pcregrep command, overriding the environment variables.
3073
3074 (d) When -B was used with -n, some line numbers in the output were one less
3075 than they should have been.
3076
3077 (e) Added the -o (--only-matching) option.
3078
3079 (f) If -A or -C was used with -c (count only), some lines of context were
3080 accidentally printed for the final match.
3081
3082 (g) Added the -H (--with-filename) option.
3083
3084 (h) The combination of options -rh failed to suppress file names for files
3085 that were found from directory arguments.
3086
3087 (i) Added the -D (--devices) and -d (--directories) options.
3088
3089 (j) Added the -F (--fixed-strings) option.
3090
3091 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3092
3093 (l) Added the --colo(u)r option.
3094
3095 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3096 is not present by default.
3097
3098 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3099 items such as (?R) or (?1), when the recursion could match a number of
3100 alternatives. If it matched one of the alternatives, but subsequently,
3101 outside the recursion, there was a failure, the code tried to back up into
3102 the recursion. However, because of the way PCRE is implemented, this is not
3103 possible, and the result was an incorrect result from the match.
3104
3105 In order to prevent this happening, the specification of recursion has
3106 been changed so that all such subpatterns are automatically treated as
3107 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3108
3109 4. I had overlooked the fact that, in some locales, there are characters for
3110 which isalpha() is true but neither isupper() nor islower() are true. In
3111 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3112 and ordfeminine) are like this. This affected the treatment of \w and \W
3113 when they appeared in character classes, but not when they appeared outside
3114 a character class. The bit map for "word" characters is now created
3115 separately from the results of isalnum() instead of just taking it from the
3116 upper, lower, and digit maps. (Plus the underscore character, of course.)
3117
3118 5. The above bug also affected the handling of POSIX character classes such as
3119 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3120 permanent tables. Instead, the bit maps for such a class were previously
3121 created as the appropriate unions of the upper, lower, and digit bitmaps.
3122 Now they are created by subtraction from the [[:word:]] class, which has
3123 its own bitmap.
3124
3125 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3126 It is created by subtracting the vertical space characters (\x09, \x0a,
3127 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3128 subtraction was done in the overall bitmap for a character class, meaning
3129 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3130 be recognized. This bug has been fixed.
3131
3132 7. Patches from the folks at Google:
3133
3134 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3135 real life, but is still worth protecting against".
3136
3137 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3138 regular expressions".
3139
3140 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3141 have it.
3142
3143 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3144 "configure" and the latter not, in order to fix a problem somebody had
3145 with compiling the Arg class on HP-UX.
3146
3147 (e) Improve the error-handling of the C++ wrapper a little bit.
3148
3149 (f) New tests for checking recursion limiting.
3150
3151 8. The pcre_memmove() function, which is used only if the environment does not
3152 have a standard memmove() function (and is therefore rarely compiled),
3153 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3154 returning a result (though PCRE never actually uses the result).
3155
3156 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3157 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3158 returned instead of calling malloc() with an overflowing number that would
3159 most likely cause subsequent chaos.
3160
3161 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3162
3163 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3164 with this option is matched, the nmatch and pmatch options of regexec() are
3165 ignored.
3166
3167 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3168 provided in case anyone wants to the the POSIX interface with UTF-8
3169 strings.
3170
3171 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3172 C++ linking (needed for some HP-UX environments).
3173
3174 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3175 (unused parameter) and in the pcre_printint() function (omitted "default"
3176 switch label when the default is to do nothing).
3177
3178 15. Added some code to make it possible, when PCRE is compiled as a C++
3179 library, to replace subject pointers for pcre_exec() with a smart pointer
3180 class, thus making it possible to process discontinuous strings.
3181
3182 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3183 much the same function. They were added by different people who were trying
3184 to make PCRE easy to compile on non-Unix systems. It has been suggested
3185 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3186 for compiling on Windows systems. I have therefore replaced it with
3187 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3188 defaults to "extern" for C or "extern C" for C++, which works fine on
3189 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3190 SCOPE with something explicit in config.h. In addition:
3191
3192 (a) pcreposix.h still had just "extern" instead of either of these macros;
3193 I have replaced it with PCRE_DATA_SCOPE.
3194
3195 (b) Functions such as _pcre_xclass(), which are internal to the library,
3196 but external in the C sense, all had PCRE_EXPORT in their definitions.
3197 This is apparently wrong for the Windows case, so I have removed it.
3198 (It makes no difference on Unix-like systems.)
3199
3200 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3201 of recursive calls to match(). This is different to MATCH_LIMIT because
3202 that limits the total number of calls to match(), not all of which increase
3203 the depth of recursion. Limiting the recursion depth limits the amount of
3204 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3205 when PCRE is compiled, and changed at run time. A patch from Google adds
3206 this functionality to the C++ interface.
3207
3208 18. Changes to the handling of Unicode character properties:
3209
3210 (a) Updated the table to Unicode 4.1.0.
3211
3212 (b) Recognize characters that are not in the table as "Cn" (undefined).
3213
3214 (c) I revised the way the table is implemented to a much improved format
3215 which includes recognition of ranges. It now supports the ranges that
3216 are defined in UnicodeData.txt, and it also amalgamates other
3217 characters into ranges. This has reduced the number of entries in the
3218 table from around 16,000 to around 3,000, thus reducing its size
3219 considerably. I realized I did not need to use a tree structure after
3220 all - a binary chop search is just as efficient. Having reduced the
3221 number of entries, I extended their size from 6 bytes to 8 bytes to
3222 allow for more data.
3223
3224 (d) Added support for Unicode script names via properties such as \p{Han}.
3225
3226 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3227 matching that character.
3228
3229 20. When matching a repeated Unicode property with a minimum greater than zero,
3230 (for example \pL{2,}), PCRE could look past the end of the subject if it
3231 reached it while seeking the minimum number of characters. This could
3232 happen only if some of the characters were more than one byte long, because
3233 there is a check for at least the minimum number of bytes.
3234
3235 21. Refactored the implementation of \p and \P so as to be more general, to
3236 allow for more different types of property in future. This has changed the
3237 compiled form incompatibly. Anybody with saved compiled patterns that use
3238 \p or \P will have to recompile them.
3239
3240 22. Added "Any" and "L&" to the supported property types.
3241
3242 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3243 but give a compile time error if the value is greater than 0xff.
3244
3245 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3246 accidentally not being installed or uninstalled.
3247
3248 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3249 made were to insert the current release number. This seemed silly, because
3250 it made things harder for people building PCRE on systems that don't run
3251 "configure". I have turned pcre.h into a distributed file, no longer built
3252 by "configure", with the version identification directly included. There is
3253 no longer a pcre.h.in file.
3254
3255 However, this change necessitated a change to the pcre-config script as
3256 well. It is built from pcre-config.in, and one of the substitutions was the
3257 release number. I have updated configure.ac so that ./configure now finds
3258 the release number by grepping pcre.h.
3259
3260 26. Added the ability to run the tests under valgrind.
3261
3262
3263 Version 6.4 05-Sep-05
3264 ---------------------
3265
3266 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3267 "--" to be printed when multiple files were scanned, even when none of the
3268 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3269 consider it to be a bug, and have restored the previous behaviour.
3270
3271 2. A couple of code tidies to get rid of compiler warnings.
3272
3273 3. The pcretest program used to cheat by referring to symbols in the library
3274 whose names begin with _pcre_. These are internal symbols that are not
3275 really supposed to be visible externally, and in some environments it is
3276 possible to suppress them. The cheating is now confined to including
3277 certain files from the library's source, which is a bit cleaner.
3278
3279 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3280 file's purpose clearer.
3281
3282 5. Reorganized pcre_ucp_findchar().
3283
3284
3285 Version 6.3 15-Aug-05
3286 ---------------------
3287
3288 1. The file libpcre.pc.in did not have general read permission in the tarball.
3289
3290 2. There were some problems when building without C++ support:
3291
3292 (a) If C++ support was not built, "make install" and "make test" still
3293 tried to test it.
3294
3295 (b) There were problems when the value of CXX was explicitly set. Some
3296 changes have been made to try to fix these, and ...
3297
3298 (c) --disable-cpp can now be used to explicitly disable C++ support.
3299
3300 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3301 backslash in a target when C++ was disabled. This confuses some
3302 versions of "make", apparently. Using an intermediate variable solves
3303 this. (Same for CPP_LOBJ.)
3304
3305 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3306 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3307 necessary on certain architectures.
3308
3309 4. Added a setting of -export-symbols-regex to the link command to remove
3310 those symbols that are exported in the C sense, but actually are local
3311 within the library, and not documented. Their names all begin with
3312 "_pcre_". This is not a perfect job, because (a) we have to except some
3313 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3314 available (and never for static libraries). I have made a note to try to
3315 find a way round (a) in the future.
3316
3317
3318 Version 6.2 01-Aug-05
3319 ---------------------
3320
3321 1. There was no test for integer overflow of quantifier values. A construction
3322 such as {1111111111111111} would give undefined results. What is worse, if
3323 a minimum quantifier for a parenthesized subpattern overflowed and became
3324 negative, the calculation of the memory size went wrong. This could have
3325 led to memory overwriting.
3326
3327 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3328
3329 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3330 operating environments where this matters.
3331
3332 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3333 PCRE options from within the C++ wrapper.
3334
3335 5. Named capturing subpatterns were not being correctly counted when a pattern
3336 was compiled. This caused two problems: (a) If there were more than 100
3337 such subpatterns, the calculation of the memory needed for the whole
3338 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3339 back references of the form \12, where the number was greater than 9, were
3340 not recognized as back references, even though there were sufficient
3341 previous subpatterns.
3342
3343 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3344 versions of gcc, e.g. 2.95.4.
3345
3346
3347 Version 6.1 21-Jun-05
3348 ---------------------
3349
3350 1. There was one reference to the variable "posix" in pcretest.c that was not
3351 surrounded by "#if !defined NOPOSIX".
3352
3353 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3354 the cross-check on the old pcre_info() function, for the benefit of the
3355 cut-down version of PCRE that is currently imported into Exim.
3356
3357 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3358 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3359 patterns that start (?i) but I don't think that matters. The use of (?i) is
3360 just an example; this all applies to the other options as well.
3361
3362 4. Since libtool seems to echo the compile commands it is issuing, the output
3363 from "make" can be reduced a bit by putting "@" in front of each libtool
3364 compile command.
3365
3366 5. Patch from the folks at Google for configure.in to be a bit more thorough
3367 in checking for a suitable C++ installation before trying to compile the
3368 C++ stuff. This should fix a reported problem when a compiler was present,
3369 but no suitable headers.
3370
3371 6. The man pages all had just "PCRE" as their title. I have changed them to
3372 be the relevant file name. I have also arranged that these names are
3373 retained in the file doc/pcre.txt, which is a concatenation in text format
3374 of all the man pages except the little individual ones for each function.
3375
3376 7. The NON-UNIX-USE file had not been updated for the different set of source
3377 files that come with release 6. I also added a few comments about the C++
3378 wrapper.
3379
3380
3381 Version 6.0 07-Jun-05
3382 ---------------------
3383
3384 1. Some minor internal re-organization to help with my DFA experiments.
3385
3386 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3387 didn't matter for the library itself when fully configured, but did matter
3388 when compiling without UCP support, or within Exim, where the ucp files are
3389 not imported.
3390
3391 3. Refactoring of the library code to split up the various functions into
3392 different source modules. The addition of the new DFA matching code (see
3393 below) to a single monolithic source would have made it really too
3394 unwieldy, quite apart from causing all the code to be include in a
3395 statically linked application, when only some functions are used. This is
3396 relevant even without the DFA addition now that patterns can be compiled in
3397 one application and matched in another.
3398
3399 The downside of splitting up is that there have to be some external
3400 functions and data tables that are used internally in different modules of
3401 the library but which are not part of the API. These have all had their
3402 names changed to start with "_pcre_" so that they are unlikely to clash
3403 with other external names.
3404
3405 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3406 a different (DFA) algorithm. Although it is slower than the original
3407 function, it does have some advantages for certain types of matching
3408 problem.
3409
3410 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3411 including restarting after a partial match.
3412
3413 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3414 defined when compiling for Windows was sent to me. I have put it into the
3415 code, though I have no means of testing or verifying it.
3416
3417 7. Added the pcre_refcount() auxiliary function.
3418
3419 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3420 match before or at the first newline in the subject string. In pcretest,
3421 the /f option on a pattern can be used to set this.
3422
3423 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3424 would behave wrongly. This has been present in PCRE since release 4.0.
3425
3426 10. A number of changes to the pcregrep command:
3427
3428 (a) Refactored how -x works; insert ^(...)$ instead of setting
3429 PCRE_ANCHORED and checking the length, in preparation for adding
3430 something similar for -w.
3431
3432 (b) Added the -w (match as a word) option.
3433
3434 (c) Refactored the way lines are read and buffered so as to have more
3435 than one at a time available.
3436
3437 (d) Implemented a pcregrep test script.
3438
3439 (e) Added the -M (multiline match) option. This allows patterns to match
3440 over several lines of the subject. The buffering ensures that at least
3441 8K, or the rest of the document (whichever is the shorter) is available
3442 for matching (and similarly the previous 8K for lookbehind assertions).
3443
3444 (f) Changed the --help output so that it now says
3445
3446 -w, --word-regex(p)
3447
3448 instead of two lines, one with "regex" and the other with "regexp"
3449 because that confused at least one person since the short forms are the
3450 same. (This required a bit of code, as the output is generated
3451 automatically from a table. It wasn't just a text change.)
3452
3453 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3454 option but starts with a hyphen. Could be a pattern or a path name
3455 starting with a hyphen, for instance.
3456
3457 (h) "-" can be given as a file name to represent stdin.
3458
3459 (i) When file names are being printed, "(standard input)" is used for
3460 the standard input, for compatibility with GNU grep. Previously
3461 "<stdin>" was used.
3462
3463 (j) The option --label=xxx can be used to supply a name to be used for
3464 stdin when file names are being printed. There is no short form.
3465
3466 (k) Re-factored the options decoding logic because we are going to add
3467 two more options that take data. Such options can now be given in four
3468 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3469
3470 (l) Added the -A, -B, and -C options for requesting that lines of context
3471 around matches be printed.
3472
3473 (m) Added the -L option to print the names of files that do not contain
3474 any matching lines, that is, the complement of -l.
3475
3476 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3477 continue to scan other files.
3478
3479 (o) The -s option was incorrectly implemented. For compatibility with other
3480 greps, it now suppresses the error message for a non-existent or non-
3481 accessible file (but not the return code). There is a new option called
3482 -q that suppresses the output of matching lines, which was what -s was
3483 previously doing.
3484
3485 (p) Added --include and --exclude options to specify files for inclusion
3486 and exclusion when recursing.
3487
3488 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3489 Hopefully, it now does.
3490
3491 12. Missing cast in pcre_study().
3492
3493 13. Added an "uninstall" target to the makefile.
3494
3495 14. Replaced "extern" in the function prototypes in Makefile.in with
3496 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3497 world, but is set differently for Windows.
3498
3499 15. Added a second compiling function called pcre_compile2(). The only
3500 difference is that it has an extra argument, which is a pointer to an
3501 integer error code. When there is a compile-time failure, this is set
3502 non-zero, in addition to the error test pointer being set to point to an
3503 error message. The new argument may be NULL if no error number is required
3504 (but then you may as well call pcre_compile(), which is now just a
3505 wrapper). This facility is provided because some applications need a
3506 numeric error indication, but it has also enabled me to tidy up the way
3507 compile-time errors are handled in the POSIX wrapper.
3508
3509 16. Added VPATH=.libs to the makefile; this should help when building with one
3510 prefix path and installing with another. (Or so I'm told by someone who
3511 knows more about this stuff than I do.)
3512
3513 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3514 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3515 match everything, including newlines. This is not POSIX-compatible, but
3516 somebody wanted the feature. From pcretest it can be activated by using
3517 both the P and the s flags.
3518
3519 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3520
3521 19. libpcre.pc was being incorrectly installed as executable.
3522
3523 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3524 it now also looks for '\r' so that it will work unmodified on Windows.
3525
3526 21. Added Google's contributed C++ wrapper to the distribution.
3527
3528 22. Added some untidy missing memory free() calls in pcretest, to keep
3529 Electric Fence happy when testing.
3530
3531
3532
3533 Version 5.0 13-Sep-04
3534 ---------------------
3535
3536 1. Internal change: literal characters are no longer packed up into items
3537 containing multiple characters in a single byte-string. Each character
3538 is now matched using a separate opcode. However, there may be more than one
3539 byte in the character in UTF-8 mode.
3540
3541 2. The pcre_callout_block structure has two new fields: pattern_position and
3542 next_item_length. These contain the offset in the pattern to the next match
3543 item, and its length, respectively.
3544
3545 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3546 insertion of callouts before each pattern item. Added the /C option to
3547 pcretest to make use of this.
3548
3549 4. On the advice of a Windows user, the lines
3550
3551 #if defined(_WIN32) || defined(WIN32)
3552 _setmode( _fileno( stdout ), 0x8000 );
3553 #endif /* defined(_WIN32) || defined(WIN32) */
3554
3555 have been added to the source of pcretest. This apparently does useful
3556 magic in relation to line terminators.
3557
3558 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3559 for the benefit of those environments where the "b" makes a difference.
3560
3561 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3562 to know about it. I have put a hack into configure.in that adds in code
3563 to set GCC=yes if CC=icc. This seems to end up at a point in the
3564 generated configure script that is early enough to affect the setting of
3565 compiler options, which is what is needed, but I have no means of testing
3566 whether it really works. (The user who reported this had patched the
3567 generated configure script, which of course I cannot do.)
3568
3569 LATER: After change 22 below (new libtool files), the configure script
3570 seems to know about icc (and also ecc). Therefore, I have commented out
3571 this hack in configure.in.
3572
3573 7. Added support for pkg-config (2 patches were sent in).
3574
3575 8. Negated POSIX character classes that used a combination of internal tables
3576 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3577 [[:^ascii]]. Typically, they would match almost any characters. The other
3578 POSIX classes were not broken in this way.
3579
3580 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3581 to find the match, as PCRE was deluded into thinking that the match had to
3582 start at the start point or following a newline. The same bug applied to
3583 patterns with negative forward assertions or any backward assertions
3584 preceding ".*" at the start, unless the pattern required a fixed first
3585 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3586
3587 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3588 starting at the last subject character, bytes beyond the end of the subject
3589 string were read.
3590
3591 11. Renamed the variable "class" as "classbits" to make life easier for C++
3592 users. (Previously there was a macro definition, but it apparently wasn't
3593 enough.)
3594
3595 12. Added the new field "tables" to the extra data so that tables can be passed
3596 in at exec time, or the internal tables can be re-selected. This allows
3597 a compiled regex to be saved and re-used at a later time by a different
3598 program that might have everything at different addresses.
3599
3600 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3601 -R library as well as a -L library.
3602
3603 14. The debugging options of pcretest (-d on the command line or D on a
3604 pattern) showed incorrect output for anything following an extended class
3605 that contained multibyte characters and which was followed by a quantifier.
3606
3607 15. Added optional support for general category Unicode character properties
3608 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3609 support. It adds about 90K to the size of the library. The meanings of the
3610 inbuilt class escapes such as \d and \s have NOT been changed.
3611
3612 16. Updated pcredemo.c to include calls to free() to release the memory for the
3613 compiled pattern.
3614
3615 17. The generated file chartables.c was being created in the source directory
3616 instead of in the building directory. This caused the build to fail if the
3617 source directory was different from the building directory, and was
3618 read-only.
3619
3620 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3621 file. No doubt somebody will tell me if they don't make sense... Also added
3622 Dan Mooney's comments about building on OpenVMS.
3623
3624 19. Added support for partial matching via the PCRE_PARTIAL option for
3625 pcre_exec() and the \P data escape in pcretest.
3626
3627 20. Extended pcretest with 3 new pattern features:
3628
3629 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3630 write the compiled pattern to the file whose name is "rest-of-line".
3631 This is a straight binary dump of the data, with the saved pointer to
3632 the character tables forced to be NULL. The study data, if any, is
3633 written too. After writing, pcretest reads a new pattern.
3634
3635 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3636 compiled pattern from the given file. There must not be any
3637 occurrences of "<" in the file name (pretty unlikely); if there are,
3638 pcretest will instead treat the initial "<" as a pattern delimiter.
3639 After reading in the pattern, pcretest goes on to read data lines as
3640 usual.
3641
3642 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3643 and 16-bit fields in a compiled pattern, to simulate a pattern that
3644 was compiled on a host of opposite endianness.
3645
3646 21. The pcre-exec() function can now cope with patterns that were compiled on
3647 hosts of opposite endianness, with this restriction:
3648
3649 As for any compiled expression that is saved and used later, the tables
3650 pointer field cannot be preserved; the extra_data field in the arguments
3651 to pcre_exec() should be used to pass in a tables address if a value
3652 other than the default internal tables were used at compile time.
3653
3654 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3655 now diagnosed as an error. Previously, most of the time, a negative number
3656 would have been treated as zero, but if in addition "ovector" was passed as
3657 NULL, a crash could occur.
3658
3659 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3660 new versions from the libtool 1.5 distribution (the last one is a copy of
3661 a file called libtool.m4). This seems to have fixed the need to patch
3662 "configure" to support Darwin 1.3 (which I used to do). However, I still
3663 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3664 workstation).
3665
3666 24. Changed the PCRE licence to be the more standard "BSD" licence.
3667
3668
3669 Version 4.5 01-Dec-03
3670 ---------------------
3671
3672 1. There has been some re-arrangement of the code for the match() function so
3673 that it can be compiled in a version that does not call itself recursively.
3674 Instead, it keeps those local variables that need separate instances for
3675 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3676 needs to "recurse". Keeping track of where control must go is done by means
3677 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3678 hide most of the details from the main code, and operates only if
3679 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3680 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3681 operating.
3682
3683 To make it easier for callers to provide specially tailored get/free
3684 functions for this usage, two new functions, pcre_stack_malloc, and
3685 pcre_stack_free, are used. They are always called in strict stacking order,
3686 and the size of block requested is always the same.
3687
3688 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3689 PCRE has been compiled to use the stack or the heap for recursion. The
3690 -C option of pcretest uses this to show which version is compiled.
3691
3692 A new data escape \S, is added to pcretest; it causes the amounts of store
3693 obtained and freed by both kinds of malloc/free at match time to be added
3694 to the output.
3695
3696 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3697 what's available on my current Linux desktop machine.
3698
3699 3. When matching a UTF-8 string, the test for a valid string at the start has
3700 been extended. If start_offset is not zero, PCRE now checks that it points
3701 to a byte that is the start of a UTF-8 character. If not, it returns
3702 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3703 this is necessary because there may be backward assertions in the pattern.
3704 When matching the same subject several times, it may save resources to use
3705 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3706
3707 4. The code for checking the validity of UTF-8 strings has been tightened so
3708 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3709 containing "overlong sequences".
3710
3711 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3712 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3713 should have been "&", but it just so happened that all the cases this let
3714 through by mistake were picked up later in the function.
3715
3716 6. I had used a variable called "isblank" - this is a C99 function, causing
3717 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3718
3719 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3720 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3721 so the version is shown; (c) stop "make test" echoing "./RunTest".
3722
3723 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3724
3725 9. The prototype for memmove() for systems that don't have it was using
3726 size_t, but the inclusion of the header that defines size_t was later. I've
3727 moved the #includes for the C headers earlier to avoid this.
3728
3729 10. Added some adjustments to the code to make it easier to compiler on certain
3730 special systems:
3731
3732 (a) Some "const" qualifiers were missing.
3733 (b) Added the macro EXPORT before all exported functions; by default this
3734 is defined to be empty.
3735 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3736 that it reads its output file name as an argument instead of writing
3737 to the standard output and assuming this can be redirected.
3738
3739 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3740 class containing characters with values greater than 255, PCRE compilation
3741 went into a loop.
3742
3743 12. A recursive reference to a subpattern that was within another subpattern
3744 that had a minimum quantifier of zero caused PCRE to crash. For example,
3745 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3746 recursion. If the recursively-called subpattern itself had a zero repeat,
3747 that was OK.
3748
3749 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3750 buffer into which it was copied (for escape processing) was still set at
3751 1024, so long lines caused crashes.
3752
3753 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3754 "internal error: code overflow...". This applied to any character class
3755 that was followed by a possessive quantifier.
3756
3757 15. Modified the Makefile to add libpcre.la as a prerequisite for
3758 libpcreposix.la because I was told this is needed for a parallel build to
3759 work.
3760
3761 16. If a pattern that contained .* following optional items at the start was
3762 studied, the wrong optimizing data was generated, leading to matching
3763 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3764 matching string must start with a or b or c. The correct conclusion for
3765 this pattern is that a match can start with any character.
3766
3767
3768 Version 4.4 13-Aug-03
3769 ---------------------
3770
3771 1. In UTF-8 mode, a character class containing characters with values between
3772 127 and 255 was not handled correctly if the compiled pattern was studied.
3773 In fixing this, I have also improved the studying algorithm for such
3774 classes (slightly).
3775
3776 2. Three internal functions had redundant arguments passed to them. Removal
3777 might give a very teeny performance improvement.
3778
3779 3. Documentation bug: the value of the capture_top field in a callout is *one
3780 more than* the number of the hightest numbered captured substring.
3781
3782 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3783 in incorrectly linking with a previously installed version. They now link
3784 explicitly with libpcre.la.
3785
3786 5. configure.in no longer needs to recognize Cygwin specially.
3787
3788 6. A problem in pcre.in for Windows platforms is fixed.
3789
3790 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3791 pcretest, it used to include the size of the study block as part of its
3792 output. Unfortunately, the structure contains a field that has a different
3793 size on different hardware architectures. This meant that the tests that
3794 showed this size failed. As the block is currently always of a fixed size,
3795 this information isn't actually particularly useful in pcretest output, so
3796 I have just removed it.
3797
3798 8. Three pre-processor statements accidentally did not start in column 1.
3799 Sadly, there are *still* compilers around that complain, even though
3800 standard C has not required this for well over a decade. Sigh.
3801
3802 9. In pcretest, the code for checking callouts passed small integers in the
3803 callout_data field, which is a void * field. However, some picky compilers
3804 complained about the casts involved for this on 64-bit systems. Now
3805 pcretest passes the address of the small integer instead, which should get
3806 rid of the warnings.
3807
3808 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3809 both compile and run time, and gives an error if an invalid UTF-8 sequence
3810 is found. There is a option for disabling this check in cases where the
3811 string is known to be correct and/or the maximum performance is wanted.
3812
3813 11. In response to a bug report, I changed one line in Makefile.in from
3814
3815 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3816 to
3817 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3818
3819 to look similar to other lines, but I have no way of telling whether this
3820 is the right thing to do, as I do not use Windows. No doubt I'll get told
3821 if it's wrong...
3822
3823
3824 Version 4.3 21-May-03
3825 ---------------------
3826
3827 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3828 Makefile.
3829
3830 2. Some refactoring to improve the quality of the code:
3831
3832 (i) The utf8_table... variables are now declared "const".
3833
3834 (ii) The code for \cx, which used the "case flipping" table to upper case
3835 lower case letters, now just substracts 32. This is ASCII-specific,
3836 but the whole concept of \cx is ASCII-specific, so it seems
3837 reasonable.
3838
3839 (iii) PCRE was using its character types table to recognize decimal and
3840 hexadecimal digits in the pattern. This is silly, because it handles
3841 only 0-9, a-f, and A-F, but the character types table is locale-
3842 specific, which means strange things might happen. A private
3843 table is now used for this - though it costs 256 bytes, a table is
3844 much faster than multiple explicit tests. Of course, the standard
3845 character types table is still used for matching digits in subject
3846 strings against \d.
3847
3848 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3849 ending in _t are). So I've renamed it as ESC_tee.
3850
3851 3. The first argument for regexec() in the POSIX wrapper should have been
3852 defined as "const".
3853
3854 4. Changed pcretest to use malloc() for its buffers so that they can be
3855 Electric Fenced for debugging.
3856
3857 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3858 to read one or more bytes before the start of the subject string. Often this
3859 had no effect on PCRE's behaviour, but in some circumstances it could
3860 provoke a segmentation fault.
3861
3862 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3863 to try to read one or more bytes before the start of the subject string.
3864
3865 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3866 UTF-8 support could misbehave in various ways if the subject string
3867 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3868 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3869 back over UTF-8 characters.)
3870
3871
3872 Version 4.2 14-Apr-03
3873 ---------------------
3874
3875 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3876
3877 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3878 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3879 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3880 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3881 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3882 and BUILD_EXEEXT
3883 Note: automatic setting of the BUILD variables is not yet working
3884 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3885 compile-time but not at link-time
3886 [LINK]: use for linking executables only
3887 make different versions for Windows and non-Windows
3888 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3889 libraries
3890 [LINK_FOR_BUILD]: new variable
3891 [OBJEXT]: use throughout
3892 [EXEEXT]: use throughout
3893 <winshared>: new target
3894 <wininstall>: new target
3895 <dftables.o>: use native compiler
3896 <dftables>: use native linker
3897 <install>: handle Windows platform correctly
3898 <clean>: ditto
3899 <check>: ditto
3900 copy DLL to top builddir before testing
3901
3902 As part of these changes, -no-undefined was removed again. This was reported
3903 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3904 in any case.
3905
3906 3. Some tidies to get rid of compiler warnings:
3907
3908 . In the match_data structure, match_limit was an unsigned long int, whereas
3909 match_call_count was an int. I've made them both unsigned long ints.
3910
3911 . In pcretest the fact that a const uschar * doesn't automatically cast to
3912 a void * provoked a warning.
3913
3914 . Turning on some more compiler warnings threw up some "shadow" variables
3915 and a few more missing casts.
3916
3917 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3918 option, a class that contained a single character with a value between 128
3919 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3920
3921 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3922 option, a class that contained several characters, but with at least one
3923 whose value was between 128 and 255 caused PCRE to crash.
3924
3925
3926 Version 4.1 12-Mar-03
3927 ---------------------
3928
3929 1. Compiling with gcc -pedantic found a couple of places where casts were
3930 needed, and a string in dftables.c that was longer than standard compilers are
3931 required to support.
3932
3933 2. Compiling with Sun's compiler found a few more places where the code could
3934 be tidied up in order to avoid warnings.
3935
3936 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3937 first of these names is deprecated in the latest Autoconf in favour of the name
3938 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3939 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3940 analogy I have changed it to CFLAGS_FOR_BUILD.
3941
3942 4. Added -no-undefined to the linking command in the Makefile, because this is
3943 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3944 linking step for the pcreposix library.
3945
3946 5. PCRE was failing to diagnose the case of two named groups with the same
3947 name.
3948
3949 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3950 literal character that is needed in the subject for a match, and scans along to
3951 ensure that it is present before embarking on the full matching process. This
3952 saves time in cases of nested unlimited repeats that are never going to match.
3953 Problem: the scan can take a lot of time if the subject is very long (e.g.
3954 megabytes), thus penalizing straightforward matches. It is now done only if the
3955 amount of subject to be scanned is less than 1000 bytes.
3956
3957 7. A lesser problem with the same optimization is that it was recording the
3958 first character of an anchored pattern as "needed", thus provoking a search
3959 right along the subject, even when the first match of the pattern was going to
3960 fail. The "needed" character is now not set for anchored patterns, unless it
3961 follows something in the pattern that is of non-fixed length. Thus, it still
3962 fulfils its original purpose of finding quick non-matches in cases of nested
3963 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3964
3965
3966 Version 4.0 17-Feb-03
3967 ---------------------
3968
3969 1. If a comment in an extended regex that started immediately after a meta-item
3970 extended to the end of string, PCRE compiled incorrect data. This could lead to
3971 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3972
3973 2. Moved to autoconf 2.53 and libtool 1.4.2.
3974
3975 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3976 the special perltest8 script is no longer needed - all the tests can be run
3977 from a single perltest script.
3978
3979 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3980 by \s. It has now been removed in PCRE. This means it isn't recognized as
3981 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3982 class [:space:] *does* include VT, thereby creating a mess.
3983
3984 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3985 space and tab.
3986
3987 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3988 its new features into the main test script, reducing the number of scripts.
3989
3990 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3991 were backward compatible, and made the (?i) apply to the whole pattern, as if
3992 /i were given. Now it behaves more logically, and applies the option setting
3993 only to what follows. PCRE has been changed to follow suit. However, if it
3994 finds options settings right at the start of the pattern, it extracts them into
3995 the global options, as before. Thus, they show up in the info data.
3996
3997 8. Added support for the \Q...\E escape sequence. Characters in between are
3998 treated as literals. This is slightly different from Perl in that $ and @ are
3999 also handled as literals inside the quotes. In Perl, they will cause variable
4000 interpolation. Note the following examples:
4001
4002 Pattern PCRE matches Perl matches
4003
4004 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4005 \Qabc\$xyz\E abc\$xyz abc\$xyz
4006 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4007
4008 For compatibility with Perl, \Q...\E sequences are recognized inside character
4009 classes as well as outside them.
4010
4011 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4012 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4013 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4014 signed/unsigned warnings.
4015
4016 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4017 option for pcretest, so I've replaced it by a simple function that does just
4018 that job.
4019
4020 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4021 "pcregrep -".
4022
4023 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4024 Java package. This provides some syntactic sugar for simple cases of what my
4025 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4026 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4027 item, you can use this simplified notation. Note that only makes sense with
4028 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4029 greediness, whatever the setting of the PCRE_UNGREEDY option.
4030
4031 13. A change of greediness default within a pattern was not taking effect at
4032 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4033 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4034 was abstracted outside.
4035
4036 14. PCRE now supports the \G assertion. It is true when the current matching
4037 position is at the start point of the match. This differs from \A when the
4038 starting offset is non-zero. Used with the /g option of pcretest (or similar
4039 code), it works in the same way as it does for Perl's /g option. If all
4040 alternatives of a regex begin with \G, the expression is anchored to the start
4041 match position, and the "anchored" flag is set in the compiled expression.
4042
4043 15. Some bugs concerning the handling of certain option changes within patterns
4044 have been fixed. These applied to options other than (?ims). For example,
4045 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4046 been the other way round. Some of this was related to change 7 above.
4047
4048 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4049 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4050 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4051 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4052
4053 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4054 mode. Unlike ".", it always matches newline, whatever the setting of
4055 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4056 assertions. Perl allows it, but it doesn't (in general) work because it can't
4057 calculate the length of the lookbehind. At least, that's the case for Perl
4058 5.8.0 - I've been told they are going to document that it doesn't work in
4059 future.
4060
4061 18. Added an error diagnosis for escapes that PCRE does not support: these are
4062 \L, \l, \N, \P, \p, \U, \u, and \X.
4063
4064 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4065 reading past the end of the pattern in cases such as /[abcd/.
4066
4067 20. PCRE was getting more memory than necessary for patterns with classes that
4068 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4069
4070 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4071 compiling PCRE for use with Virtual Pascal.
4072
4073 22. Small fix to the Makefile to make it work properly if the build is done
4074 outside the source tree.
4075
4076 23. Added a new extension: a condition to go with recursion. If a conditional
4077 subpattern starts with (?(R) the "true" branch is used if recursion has
4078 happened, whereas the "false" branch is used only at the top level.
4079
4080 24. When there was a very long string of literal characters (over 255 bytes
4081 without UTF support, over 250 bytes with UTF support), the computation of how
4082 much memory was required could be incorrect, leading to segfaults or other
4083 strange effects.
4084
4085 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4086 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4087 there was a subsequent back reference to those brackets. This meant that, for
4088 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4089 possible to check for precisely this case. All we can do is abandon the
4090 optimization if .* occurs inside capturing brackets when there are any back
4091 references whatsoever. (See below for a better fix that came later.)
4092
4093 26. The handling of the optimization for finding the first character of a
4094 non-anchored pattern, and for finding a character that is required later in the
4095 match were failing in some cases. This didn't break the matching; it just
4096 failed to optimize when it could. The way this is done has been re-implemented.
4097
4098 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4099
4100 28. Added a new feature that provides some of the functionality that Perl
4101 provides with (?{...}). The facility is termed a "callout". The way it is done
4102 in PCRE is for the caller to provide an optional function, by setting
4103 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4104 global variable. By default it is unset, which disables all calling out. To get
4105 the function called, the regex must include (?C) at appropriate points. This
4106 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4107 This provides a means of identifying different callout points. When PCRE
4108 reaches such a point in the regex, if pcre_callout has been set, the external
4109 function is called. It is provided with data in a structure called
4110 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4111 matching continues; if it returns a non-zero value, the match at the current
4112 point fails. However, backtracking will occur if possible. [This was changed
4113 later and other features added - see item 49 below.]
4114
4115 29. pcretest is upgraded to test the callout functionality. It provides a
4116 callout function that displays information. By default, it shows the start of
4117 the match and the current position in the text. There are some new data escapes
4118 to vary what happens:
4119
4120 \C+ in addition, show current contents of captured substrings
4121 \C- do not supply a callout function
4122 \C!n return 1 when callout number n is reached
4123 \C!n!m return 1 when callout number n is reached for the mth time
4124
4125 30. If pcregrep was called with the -l option and just a single file name, it
4126 output "<stdin>" if a match was found, instead of the file name.
4127
4128 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4129 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4130 pcre_exec(). This saves a malloc/free per call. The default value of
4131 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4132 when configuring.
4133
4134 32. The default maximum size of a compiled pattern is 64K. There have been a
4135 few cases of people hitting this limit. The code now uses macros to handle the
4136 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4137 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4138 configuring. Tests 2 and 5 work only with 2-byte links because they output
4139 debugging information about compiled patterns.
4140
4141 33. Internal code re-arrangements:
4142
4143 (a) Moved the debugging function for printing out a compiled regex into
4144 its own source file (printint.c) and used #include to pull it into
4145 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4146 separate copies.
4147
4148 (b) Defined the list of op-code names for debugging as a macro in
4149 internal.h so that it is next to the definition of the opcodes.
4150
4151 (c) Defined a table of op-code lengths for simpler skipping along compiled
4152 code. This is again a macro in internal.h so that it is next to the
4153 definition of the opcodes.
4154
4155 34. Added support for recursive calls to individual subpatterns, along the
4156 lines of Robin Houston's patch (but implemented somewhat differently).
4157
4158 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4159 allow it to read and process whole directories in Win32. This code was
4160 contributed by Lionel Fourquaux; it has not been tested by me.
4161
4162 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4163 used to name a group. Names consist of alphanumerics and underscores, and must
4164 be unique. Back references use the syntax (?P=name) and recursive calls use
4165 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4166 numbers. The function pcre_fullinfo() can be used after compilation to extract
4167 a name/number map. There are three relevant calls:
4168
4169 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4170 PCRE_INFO_NAMECOUNT yields the number of entries
4171 PCRE_INFO_NAMETABLE yields a pointer to the map.
4172
4173 The map is a vector of fixed-size entries. The size of each entry depends on
4174 the length of the longest name used. The first two bytes of each entry are the
4175 group number, most significant byte first. There follows the corresponding
4176 name, zero terminated. The names are in alphabetical order.
4177
4178 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4179 case instead of 255. Making it the same both with and without UTF-8 support
4180 means that the same test output works with both.
4181
4182 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4183 calling malloc() with a zero argument.
4184
4185 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4186 optimization. I've improved things by keeping a bitmap of backreferences with
4187 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4188 fact referenced, the optimization can be applied. It is unlikely that a
4189 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4190 the match to follow \n) will appear inside brackets with a number greater than
4191 31, but if it does, any back reference > 31 suppresses the optimization.
4192
4193 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4194 of disabling numbered capturing parentheses. Any opening parenthesis that is
4195 not followed by ? behaves as if it were followed by ?: but named parentheses
4196 can still be used for capturing (and they will acquire numbers in the usual
4197 way).
4198
4199 41. Redesigned the return codes from the match() function into yes/no/error so
4200 that errors can be passed back from deep inside the nested calls. A malloc
4201 failure while inside a recursive subpattern call now causes the
4202 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4203
4204 42. It is now possible to set a limit on the number of times the match()
4205 function is called in a call to pcre_exec(). This facility makes it possible to
4206 limit the amount of recursion and backtracking, though not in a directly
4207 obvious way, because the match() function is used in a number of different
4208 circumstances. The count starts from zero for each position in the subject
4209 string (for non-anchored patterns). The default limit is, for compatibility, a
4210 large number, namely 10 000 000. You can change this in two ways:
4211
4212 (a) When configuring PCRE before making, you can use --with-match-limit=n
4213 to set a default value for the compiled library.
4214
4215 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4216 a different value is set. See 45 below.
4217
4218 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4219
4220 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4221 of things that can be changed at compile time. The first argument specifies
4222 what is wanted and the second points to where the information is to be placed.
4223 The current list of available information is:
4224
4225 PCRE_CONFIG_UTF8
4226
4227 The output is an integer that is set to one if UTF-8 support is available;
4228 otherwise it is set to zero.
4229
4230 PCRE_CONFIG_NEWLINE
4231
4232 The output is an integer that it set to the value of the code that is used for
4233 newline. It is either LF (10) or CR (13).
4234
4235 PCRE_CONFIG_LINK_SIZE
4236
4237 The output is an integer that contains the number of bytes used for internal
4238 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4239
4240 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4241
4242 The output is an integer that contains the threshold above which the POSIX
4243 interface uses malloc() for output vectors. See item 31 above.
4244
4245 PCRE_CONFIG_MATCH_LIMIT
4246
4247 The output is an unsigned integer that contains the default limit of the number
4248 of match() calls in a pcre_exec() execution. See 42 above.
4249
4250 44. pcretest has been upgraded by the addition of the -C option. This causes it
4251 to extract all the available output from the new pcre_config() function, and to
4252 output it. The program then exits immediately.
4253
4254 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4255 order to support additional features. One way would have been to define
4256 pcre_exec2() (for example) with extra arguments, but this would not have been
4257 extensible, and would also have required all calls to the original function to
4258 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4259 is used for passing in "extra" data from pcre_study().
4260
4261 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4262 contains the following fields:
4263
4264 flags a bitmap indicating which of the following fields are set
4265 study_data opaque data from pcre_study()
4266 match_limit a way of specifying a limit on match() calls for a specific
4267 call to pcre_exec()
4268 callout_data data for callouts (see 49 below)
4269
4270 The flag bits are also defined in pcre.h, and are
4271
4272 PCRE_EXTRA_STUDY_DATA
4273 PCRE_EXTRA_MATCH_LIMIT
4274 PCRE_EXTRA_CALLOUT_DATA
4275
4276 The pcre_study() function now returns one of these new pcre_extra blocks, with
4277 the actual study data pointed to by the study_data field, and the
4278 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4279 before. That is, this change is entirely upwards-compatible and requires no
4280 change to existing code.
4281
4282 If you want to pass in additional data to pcre_exec(), you can either place it
4283 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4284 block.
4285
4286 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4287 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4288 times with different match limits, until it finds the minimum value needed for
4289 pcre_exec() to complete. The value is then output. This can be instructive; for
4290 most simple matches the number is quite small, but for pathological cases it
4291 gets very large very quickly.
4292
4293 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4294 returns the size of the data block pointed to by the study_data field in a
4295 pcre_extra block, that is, the value that was passed as the argument to
4296 pcre_malloc() when PCRE was getting memory in which to place the information
4297 created by pcre_study(). The fourth argument should point to a size_t variable.
4298 pcretest has been extended so that this information is shown after a successful
4299 pcre_study() call when information about the compiled regex is being displayed.
4300
4301 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4302 because what follows is always an absolute path. (Later: it turns out that this
4303 is more than cosmetic for MinGW, because it doesn't like empty path
4304 components.)
4305
4306 49. Some changes have been made to the callout feature (see 28 above):
4307
4308 (i) A callout function now has three choices for what it returns:
4309
4310 0 => success, carry on matching
4311 > 0 => failure at this point, but backtrack if possible
4312 < 0 => serious error, return this value from pcre_exec()
4313
4314 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4315 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4316 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4317 use by callout functions. It will never be used by PCRE itself.
4318
4319 (ii) The pcre_extra structure (see 45 above) has a void * field called
4320 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4321 pcre_callout_block structure has a field of the same name. The contents of
4322 the field passed in the pcre_extra structure are passed to the callout
4323 function in the corresponding field in the callout block. This makes it
4324 easier to use the same callout-containing regex from multiple threads. For
4325 testing, the pcretest program has a new data escape
4326
4327 \C*n pass the number n (may be negative) as callout_data
4328
4329 If the callout function in pcretest receives a non-zero value as
4330 callout_data, it returns that value.
4331
4332 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4333 there were some redundant $(CFLAGS) in commands that are now specified as
4334 $(LINK), which already includes $(CFLAGS).
4335
4336 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4337 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4338 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4339 one-byte characters throughout. Note that case-insensitive matching applies
4340 only to characters whose values are less than 256. PCRE doesn't support the
4341 notion of cases for higher-valued characters.
4342
4343 (i) A character class whose characters are all within 0-255 is handled as
4344 a bit map, and the map is inverted for negative classes. Previously, a
4345 character > 255 always failed to match such a class; however it should
4346 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4347
4348 (ii) A negated character class with a single character < 255 is coded as
4349 "not this character" (OP_NOT). This wasn't working properly when the test
4350 character was multibyte, either singly or repeated.
4351
4352 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4353 mode, for example: \x{100}{2,3}.
4354
4355 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4356 singly or repeated) now correctly test multibyte characters. However,
4357 PCRE doesn't recognize any characters with values greater than 255 as
4358 digits, spaces, or word characters. Such characters always match \D, \S,
4359 and \W, and never match \d, \s, or \w.
4360
4361 (v) Classes may now contain characters and character ranges with values
4362 greater than 255. For example: [ab\x{100}-\x{400}].
4363
4364 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4365 PCRE in UTF-8 mode.
4366
4367 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4368 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4369 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4370 value.)
4371
4372 53. The single man page has become too large. I have therefore split it up into
4373 a number of separate man pages. These also give rise to individual HTML pages;
4374 these are now put in a separate directory, and there is an index.html page that
4375 lists them all. Some hyperlinking between the pages has been installed.
4376
4377 54. Added convenience functions for handling named capturing parentheses.
4378
4379 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4380 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4381 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4382 are faulted.
4383
4384 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4385 calling configure. These values are used when compiling the dftables.c program
4386 which is run to generate the source of the default character tables. They
4387 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4388 you will need to set these values.
4389
4390 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4391
4392
4393 Version 3.9 02-Jan-02
4394 ---------------------
4395
4396 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4397
4398 2. If --disable-static was given, the building process failed when trying to
4399 build pcretest and pcregrep. (For some reason it was using libtool to compile
4400 them, which is not right, as they aren't part of the library.)
4401
4402
4403 Version 3.8 18-Dec-01
4404 ---------------------
4405
4406 1. The experimental UTF-8 code was completely screwed up. It was packing the
4407 bytes in the wrong order. How dumb can you get?
4408
4409
4410 Version 3.7 29-Oct-01
4411 ---------------------
4412
4413 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4414 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4415 this didn't happen under Solaris 8, where I normally test things.
4416
4417 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4418 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4419 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4420 this fix an hour or so after the initial 3.7 release.)
4421
4422
4423 Version 3.6 23-Oct-01
4424 ---------------------
4425
4426 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4427 offsets passed as NULL with zero offset count.
4428
4429 2. The config.guess and config.sub files had not been updated when I moved to
4430 the latest autoconf.
4431
4432
4433 Version 3.5 15-Aug-01
4434 ---------------------
4435
4436 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4437 had been forgotten.
4438
4439 2. By using declared but undefined structures, we can avoid using "void"
4440 definitions in pcre.h while keeping the internal definitions of the structures
4441 private.
4442
4443 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4444 user point of view, this means that both static and shared libraries are built
4445 by default, but this can be individually controlled. More of the work of
4446 handling this static/shared cases is now inside libtool instead of PCRE's make
4447 file.
4448
4449 4. The pcretest utility is now installed along with pcregrep because it is
4450 useful for users (to test regexs) and by doing this, it automatically gets
4451 relinked by libtool. The documentation has been turned into a man page, so
4452 there are now .1, .txt, and .html versions in /doc.
4453
4454 5. Upgrades to pcregrep:
4455 (i) Added long-form option names like gnu grep.
4456 (ii) Added --help to list all options with an explanatory phrase.
4457 (iii) Added -r, --recursive to recurse into sub-directories.
4458 (iv) Added -f, --file to read patterns from a file.
4459
4460 6. pcre_exec() was referring to its "code" argument before testing that
4461 argument for NULL (and giving an error if it was NULL).
4462
4463 7. Upgraded Makefile.in to allow for compiling in a different directory from
4464 the source directory.
4465
4466 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4467 options bits, the pointer it was passed was to an int instead of to an unsigned
4468 long int. This mattered only on 64-bit systems.
4469
4470 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4471 generated) instead of pcre.in, which it its source. Also made the same change
4472 in several of the .c files.
4473
4474 10. A new release of gcc defines printf() as a macro, which broke pcretest
4475 because it had an ifdef in the middle of a string argument for printf(). Fixed
4476 by using separate calls to printf().
4477
4478 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4479 script, to force use of CR or LF instead of \n in the source. On non-Unix
4480 systems, the value can be set in config.h.
4481
4482 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4483 absolute limit. Changed the text of the error message to make this clear, and
4484 likewise updated the man page.
4485
4486 13. The limit of 99 on the number of capturing subpatterns has been removed.
4487 The new limit is 65535, which I hope will not be a "real" limit.
4488
4489
4490 Version 3.4 22-Aug-00
4491 ---------------------
4492
4493 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4494
4495 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4496
4497
4498 Version 3.3 01-Aug-00
4499 ---------------------
4500
4501 1. If an octal character was given, but the value was greater than \377, it
4502 was not getting masked to the least significant bits, as documented. This could
4503 lead to crashes in some systems.
4504
4505 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4506 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4507
4508 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4509 These just pass their arguments on to (pcre_free)(), but they are provided
4510 because some uses of PCRE bind it to non-C systems that can call its functions,
4511 but cannot call free() or pcre_free() directly.
4512
4513 4. Add "make test" as a synonym for "make check". Corrected some comments in
4514 the Makefile.
4515
4516 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4517 Makefile.
4518
4519 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4520 command called pgrep for grepping around the active processes.
4521
4522 7. Added the beginnings of support for UTF-8 character strings.
4523
4524 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4525 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4526 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4527 out for the ar command.)
4528
4529
4530 Version 3.2 12-May-00
4531 ---------------------
4532
4533 This is purely a bug fixing release.
4534
4535 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4536 of ZA. This was just one example of several cases that could provoke this bug,
4537 which was introduced by change 9 of version 2.00. The code for breaking
4538 infinite loops after an iteration that matches an empty string was't working
4539 correctly.
4540
4541 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4542 when matched against abbab (for example). After matching an empty string, it
4543 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4544 caused it to match further down the string than it should.
4545
4546 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4547 was there because it doesn't seem to be needed, and it causes trouble on some
4548 systems, as it is not a Standard C header. It has been removed.
4549
4550 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4551 were reported on the Macintosh. The changes were from
4552
4553 while ((c = *(++ptr)) != 0 && c != '\n');
4554 to
4555 while ((c = *(++ptr)) != 0 && c != '\n') ;
4556
4557 Totally extraordinary, but if that's what it takes...
4558
4559 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4560 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4561 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4562 assumes the way PCRE uses memmove() (always moving upwards).
4563
4564 6. PCRE is being used in one environment where strchr() is not available. There
4565 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4566 faster code anyway.
4567
4568
4569 Version 3.1 09-Feb-00
4570 ---------------------
4571
4572 The only change in this release is the fixing of some bugs in Makefile.in for
4573 the "install" target:
4574
4575 (1) It was failing to install pcreposix.h.
4576
4577 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4578
4579
4580 Version 3.0 01-Feb-00
4581 ---------------------
4582
4583 1. Add support for the /+ modifier to perltest (to output $` like it does in
4584 pcretest).
4585
4586 2. Add support for the /g modifier to perltest.
4587
4588 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4589 matches null strings.
4590
4591 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4592 pattern. Perl treats empty patterns specially - it reuses the most recent
4593 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4594 effect.
4595
4596 5. The POSIX interface was broken in that it was just handing over the POSIX
4597 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4598 required a bigger vector, with some working space on the end. This means that
4599 the POSIX wrapper now has to get and free some memory, and copy the results.
4600
4601 6. Added some simple autoconf support, placing the test data and the
4602 documentation in separate directories, re-organizing some of the
4603 information files, and making it build pcre-config (a GNU standard). Also added
4604 libtool support for building PCRE as a shared library, which is now the
4605 default.
4606
4607 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4608 09 are not valid octal constants. Single digits will be used for minor values
4609 less than 10.
4610
4611 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4612 existing programs that set these in the POSIX interface can use PCRE without
4613 modification.
4614
4615 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4616 return all that pcre_info() returns, plus additional data. The pcre_info()
4617 function is retained for compatibility, but is considered to be obsolete.
4618
4619 10. Added experimental recursion feature (?R) to handle one common case that
4620 Perl 5.6 will be able to do with (?p{...}).
4621
4622 11. Added support for POSIX character classes like [:alpha:], which Perl is
4623 adopting.
4624
4625
4626 Version 2.08 31-Aug-99
4627 ----------------------
4628
4629 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4630 trying to match at the startoffset position, but instead was moving forward to
4631 the next newline as if a previous match had failed.
4632
4633 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4634 and could get into a loop if a null string was matched other than at the start
4635 of the subject.
4636
4637 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4638 be distinguished at compile time, and for completeness also added PCRE_DATE.
4639
4640 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4641 in GnuWin32 environments.
4642
4643
4644 Version 2.07 29-Jul-99
4645 ----------------------
4646
4647 1. The documentation is now supplied in plain text form and HTML as well as in
4648 the form of man page sources.
4649
4650 2. C++ compilers don't like assigning (void *) values to other pointer types.
4651 In particular this affects malloc(). Although there is no problem in Standard
4652 C, I've put in casts to keep C++ compilers happy.
4653
4654 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4655 should be (const char *).
4656
4657 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4658 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4659 However, I haven't made this a standard facility. The documentation doesn't
4660 mention it, and the Makefile doesn't support it.
4661
4662 5. The Makefile now contains an "install" target, with editable destinations at
4663 the top of the file. The pcretest program is not installed.
4664
4665 6. pgrep -V now gives the PCRE version number and date.
4666
4667 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4668 causing the entire string to be ignored, instead of just the last character.
4669
4670 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4671 non-matching string, it can take a very, very long time, even for strings of
4672 quite modest length, because of the nested recursion. PCRE now does better in
4673 some of these cases. It does this by remembering the last required literal
4674 character in the pattern, and pre-searching the subject to ensure it is present
4675 before running the real match. In other words, it applies a heuristic to detect
4676 some types of certain failure quickly, and in the above example, if presented
4677 with a string that has no trailing " it gives "no match" very quickly.
4678
4679 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4680 other alternatives are tried instead.
4681
4682
4683 Version 2.06 09-Jun-99
4684 ----------------------
4685
4686 1. Change pcretest's output for amount of store used to show just the code
4687 space, because the remainder (the data block) varies in size between 32-bit and
4688 64-bit systems.
4689
4690 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4691 start matching at. This allows lookbehinds to work when searching for multiple
4692 occurrences in a string.
4693
4694 3. Added additional options to pcretest for testing multiple occurrences:
4695
4696 /+ outputs the rest of the string that follows a match
4697 /g loops for multiple occurrences, using the new startoffset argument
4698 /G loops for multiple occurrences by passing an incremented pointer
4699
4700 4. PCRE wasn't doing the "first character" optimization for patterns starting
4701 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4702 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4703 the letter 'x'. On long subject strings, this gives a significant speed-up.
4704
4705
4706 Version 2.05 21-Apr-99
4707 ----------------------
4708
4709 1. Changed the type of magic_number from int to long int so that it works
4710 properly on 16-bit systems.
4711
4712 2. Fixed a bug which caused patterns starting with .* not to work correctly
4713 when the subject string contained newline characters. PCRE was assuming
4714 anchoring for such patterns in all cases, which is not correct because .* will
4715 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4716 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4717 must be retried after every newline in the subject.
4718
4719
4720 Version 2.04 18-Feb-99
4721 ----------------------
4722
4723 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4724 computation of the store needed to hold the pattern was incorrect (too large).
4725 If such patterns were nested a few deep, this could multiply and become a real
4726 problem.
4727
4728 2. Added /M option to pcretest to show the memory requirement of a specific
4729 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4730
4731 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4732 compiled in such a way that the backtracking after subsequent failure was
4733 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4734 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4735
4736
4737 Version 2.03 02-Feb-99
4738 ----------------------
4739
4740 1. Fixed typo and small mistake in man page.
4741
4742 2. Added 4th condition (GPL supersedes if conflict) and created separate
4743 LICENCE file containing the conditions.
4744
4745 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4746 Perl, that is the internal \ allows the delimiter to be included in the
4747 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4748 the final delimiter, add \ to the end of the pattern (to test the error).
4749
4750 4. Added the convenience functions for extracting substrings after a successful
4751 match. Updated pcretest to make it able to test these functions.
4752
4753
4754 Version 2.02 14-Jan-99
4755 ----------------------
4756
4757 1. Initialized the working variables associated with each extraction so that
4758 their saving and restoring doesn't refer to uninitialized store.
4759
4760 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4761 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4762 fix the problem.
4763
4764 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4765 calls, and wasn't printing the correct value for compiling calls. Increased the
4766 default value of LOOPREPEAT, and the number of significant figures in the
4767 times.
4768
4769 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4770
4771 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4772 a building problem on Windows NT with a FAT file system.
4773
4774
4775 Version 2.01 21-Oct-98
4776 ----------------------
4777
4778 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4779 to character tables built by pcre_maketables() in the current locale. If NULL
4780 is passed, the default tables are used.
4781
4782
4783 Version 2.00 24-Sep-98
4784 ----------------------
4785
4786 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4787 it any more.
4788
4789 2. Allow quantification of (?>) groups, and make it work correctly.
4790
4791 3. The first character computation wasn't working for (?>) groups.
4792
4793 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4794 end of the subject) and add 5.005's \z, which really does match only at the
4795 very end of the subject.
4796
4797 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4798
4799 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4800 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4801 localized options. All options to pcre_study() were also removed.
4802
4803 7. Add other new features from 5.005:
4804
4805 $(?<= positive lookbehind
4806 $(?<! negative lookbehind
4807 (?imsx-imsx) added the unsetting capability
4808 such a setting is global if at outer level; local otherwise
4809 (?imsx-imsx:) non-capturing groups with option setting
4810 (?(cond)re|re) conditional pattern matching
4811
4812 A backreference to itself in a repeated group matches the previous
4813 captured string.
4814
4815 8. General tidying up of studying (both automatic and via "study")
4816 consequential on the addition of new assertions.
4817
4818 9. As in 5.005, unlimited repeated groups that could match an empty substring
4819 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4820 runtime if any iteration does actually match an empty substring.
4821
4822 10. Include the RunTest script in the distribution.
4823
4824 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4825 discrepancies, some of which were old and were also with respect to 5.004. They
4826 have now been fixed.
4827
4828
4829 Version 1.09 28-Apr-98
4830 ----------------------
4831
4832 1. A negated single character class followed by a quantifier with a minimum
4833 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4834 program crashes, or just wrong answers. This did not apply to negated classes
4835 containing more than one character, or to minima other than one.
4836
4837
4838 Version 1.08 27-Mar-98
4839 ----------------------
4840
4841 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4842
4843 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4844 latter must appear before anything that relies on it in the pattern.
4845
4846
4847 Version 1.07 16-Feb-98
4848 ----------------------
4849
4850 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4851 repeat of a potentially empty string).
4852
4853
4854 Version 1.06 23-Jan-98
4855 ----------------------
4856
4857 1. Added Markus Oberhumer's little patches for C++.
4858
4859 2. Literal strings longer than 255 characters were broken.
4860
4861
4862 Version 1.05 23-Dec-97
4863 ----------------------
4864
4865 1. Negated character classes containing more than one character were failing if
4866 PCRE_CASELESS was set at run time.
4867
4868
4869 Version 1.04 19-Dec-97
4870 ----------------------
4871
4872 1. Corrected the man page, where some "const" qualifiers had been omitted.
4873
4874 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4875 input syntax.
4876
4877 3. Fixed memory leak which occurred when a regex with back references was
4878 matched with an offsets vector that wasn't big enough. The temporary memory
4879 that is used in this case wasn't being freed if the match failed.
4880
4881 4. Tidied pcretest to ensure it frees memory that it gets.
4882
4883 5. Temporary memory was being obtained in the case where the passed offsets
4884 vector was exactly big enough.
4885
4886 6. Corrected definition of offsetof() from change 5 below.
4887
4888 7. I had screwed up change 6 below and broken the rules for the use of
4889 setjmp(). Now fixed.
4890
4891
4892 Version 1.03 18-Dec-97
4893 ----------------------
4894
4895 1. A erroneous regex with a missing opening parenthesis was correctly
4896 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4897 on some systems.
4898
4899 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4900 it was reported that one broken compiler failed on the former because "code" is
4901 also an independent variable.
4902
4903 3. The erroneous regex a[]b caused an array overrun reference.
4904
4905 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4906 fail on data ending with that character. (It was going on too far, and checking
4907 the next character, typically a binary zero.) This was specific to the
4908 optimized code for single-character negative classes.
4909
4910 5. Added a contributed patch from the TIN world which does the following:
4911
4912 + Add an undef for memmove, in case the the system defines a macro for it.
4913
4914 + Add a definition of offsetof(), in case there isn't one. (I don't know
4915 the reason behind this - offsetof() is part of the ANSI standard - but
4916 it does no harm).
4917
4918 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4919 most of the places where whitespace preceded '#'. I have given up and
4920 allowed the remaining 2 cases to be at the margin.
4921
4922 + Rename some variables in pcre to eliminate shadowing. This seems very
4923 pedantic, but does no harm, of course.
4924
4925 6. Moved the call to setjmp() into its own function, to get rid of warnings
4926 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4927
4928 7. Constructs such as \d{8,} were compiling into the equivalent of
4929 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4930 outcome, but in this particular case used more store than had been allocated,
4931 which caused the bug to be discovered because it threw up an internal error.
4932
4933 8. The debugging code in both pcre and pcretest for outputting the compiled
4934 form of a regex was going wrong in the case of back references followed by
4935 curly-bracketed repeats.
4936
4937
4938 Version 1.02 12-Dec-97
4939 ----------------------
4940
4941 1. Typos in pcre.3 and comments in the source fixed.
4942
4943 2. Applied a contributed patch to get rid of places where it used to remove
4944 'const' from variables, and fixed some signed/unsigned and uninitialized
4945 variable warnings.
4946
4947 3. Added the "runtest" target to Makefile.
4948
4949 4. Set default compiler flag to -O2 rather than just -O.
4950
4951
4952 Version 1.01 19-Nov-97
4953 ----------------------
4954
4955 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4956 like /([ab]*)*/, that is, for classes with more than one character in them.
4957
4958 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4959 as /((?>a*))*/ (a PCRE_EXTRA facility).
4960
4961
4962 Version 1.00 18-Nov-97
4963 ----------------------
4964
4965 1. Added compile-time macros to support systems such as SunOS4 which don't have
4966 memmove() or strerror() but have other things that can be used instead.
4967
4968 2. Arranged that "make clean" removes the executables.
4969
4970
4971 Version 0.99 27-Oct-97
4972 ----------------------
4973
4974 1. Fixed bug in code for optimizing classes with only one character. It was
4975 initializing a 32-byte map regardless, which could cause it to run off the end
4976 of the memory it had got.
4977
4978 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4979
4980
4981 Version 0.98 22-Oct-97
4982 ----------------------
4983
4984 1. Fixed bug in code for handling temporary memory usage when there are more
4985 back references than supplied space in the ovector. This could cause segfaults.
4986
4987
4988 Version 0.97 21-Oct-97
4989 ----------------------
4990
4991 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4992
4993 2. Optimized negated single characters not to use a bit map.
4994
4995 3. Brought error texts together as macro definitions; clarified some of them;
4996 fixed one that was wrong - it said "range out of order" when it meant "invalid
4997 escape sequence".
4998
4999 4. Changed some char * arguments to const char *.
5000
5001 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5002
5003 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5004 pcretest.
5005
5006
5007 Version 0.96 16-Oct-97
5008 ----------------------
5009
5010 1. Added a simple "pgrep" utility to the distribution.
5011
5012 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5013 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5014 where "ddd" means "one or more decimal digits".
5015
5016 3. Fixed serious bug. If a pattern had a back reference, but the call to
5017 pcre_exec() didn't supply a large enough ovector to record the related
5018 identifying subpattern, the match always failed. PCRE now remembers the number
5019 of the largest back reference, and gets some temporary memory in which to save
5020 the offsets during matching if necessary, in order to ensure that
5021 backreferences always work.
5022
5023 4. Increased the compatibility with Perl in a number of ways:
5024
5025 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5026 to request this handling. The option can be set at compile or exec time.
5027
5028 (b) $ matches before a terminating newline by default; an option
5029 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5030 mode). The option can be set at compile or exec time.
5031
5032 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5033 the same as Perl's. If the decimal number it represents is less than 10
5034 or there aren't that many previous left capturing parentheses, an octal
5035 escape is read. Inside a character class, it's always an octal escape,
5036 even if it is a single digit.
5037
5038 (d) An escaped but undefined alphabetic character is taken as a literal,
5039 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5040 escapes.
5041
5042 (e) {0} is now permitted. (The previous item is removed from the compiled
5043 pattern).
5044
5045 5. Changed all the names of code files so that the basic parts are no longer
5046 than 10 characters, and abolished the teeny "globals.c" file.
5047
5048 6. Changed the handling of character classes; they are now done with a 32-byte
5049 bit map always.
5050
5051 7. Added the -d and /D options to pcretest to make it possible to look at the
5052 internals of compilation without having to recompile pcre.
5053
5054
5055 Version 0.95 23-Sep-97
5056 ----------------------
5057
5058 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5059 \x20 at the start of a run of normal characters. These were being treated as
5060 real characters, instead of the source characters being re-checked.
5061
5062
5063 Version 0.94 18-Sep-97
5064 ----------------------
5065
5066 1. The functions are now thread-safe, with the caveat that the global variables
5067 containing pointers to malloc() and free() or alternative functions are the
5068 same for all threads.
5069
5070 2. Get pcre_study() to generate a bitmap of initial characters for non-
5071 anchored patterns when this is possible, and use it if passed to pcre_exec().
5072
5073
5074 Version 0.93 15-Sep-97
5075 ----------------------
5076
5077 1. /(b)|(:+)/ was computing an incorrect first character.
5078
5079 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5080 but not actually doing anything yet.
5081
5082 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5083 as Perl does (e.g. [-az] or [az-]).
5084
5085 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5086 all possible positions.
5087
5088 5. Split up into different modules to avoid including unneeded functions in a
5089 compiled binary. However, compile and exec are still in one module. The "study"
5090 function is split off.
5091
5092 6. The character tables are now in a separate module whose source is generated
5093 by an auxiliary program - but can then be edited by hand if required. There are
5094 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5095 toupper() in the code.
5096
5097 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5098 make them global. Abolish the function for setting them, as the caller can now
5099 set them directly.
5100
5101
5102 Version 0.92 11-Sep-97
5103 ----------------------
5104
5105 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5106 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5107
5108 2. Caseless matching was not working in character classes if the characters in
5109 the pattern were in upper case.
5110
5111 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5112
5113 4. Make PCRE_ANCHORED public and accept as a compile option.
5114
5115 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5116 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5117 pass them.
5118
5119 6. Give an error if bad option bits passed at compile or run time.
5120
5121 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5122 pcretest to cause it to pass that flag.
5123
5124 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5125 options, and the first character, if set.
5126
5127 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5128
5129
5130 Version 0.91 10-Sep-97
5131 ----------------------
5132
5133 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5134 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5135
5136 2. PCRE was looping on encountering an indefinitely repeated back reference to
5137 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5138 Perl does - treats the match as successful.
5139
5140 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5