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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1342 - (show annotations)
Mon Jul 1 12:28:12 2013 UTC (6 years, 3 months ago) by ph10
File size: 248099 byte(s)
Error occurred while calculating annotation data.
Remove SUPPORT_GCOV from config.h, because it is not used and is misleading.
(Coverage support does not need any code changes.)
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.34 xx-xxxx-201x
5 --------------------------
6
7 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
8 executable memory. Patch inspired by Carsten Klein.
9
10 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
11 this macro is never tested and has no effect, because the work to support
12 coverage involves only compiling and linking options and special targets in
13 the Makefile. The comment in config.h implied that defining the macro would
14 enable coverage support, which is totally false. There was also support for
15 setting this macro in the CMake files (my fault, I just copied it from
16 configure). SUPPORT_GCOV has now been removed.
17
18
19 Version 8.33 28-May-2013
20 --------------------------
21
22 1. Added 'U' to some constants that are compared to unsigned integers, to
23 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
24 variables that are added to signed variables, to ensure the result is
25 signed and can be negated.
26
27 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
28 CMake config files.
29
30 3. Revise the creation of config.h.generic so that all boolean macros are
31 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
32 overriding via -D on the command line possible.
33
34 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
35 to unsigned int is reported to make a quite noticeable speed difference in
36 a specific Windows environment. Testing on Linux did also appear to show
37 some benefit (and it is clearly not harmful). Also fixed the definition of
38 Xop which should be unsigned.
39
40 5. Related to (4), changing the definition of the intermediate variable cc
41 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
42 improvements.
43
44 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
45 unnecessary spaces.
46
47 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
48 and later.
49
50 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
51
52 9. Optimizing fast_forward_start_bits in JIT.
53
54 10. Adding support for callouts in JIT, and fixing some issues revealed
55 during this work. Namely:
56
57 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
58
59 (b) Minimum length was not checked before the matching is started.
60
61 11. The value of capture_last that is passed to callouts was incorrect in some
62 cases when there was a capture on one path that was subsequently abandoned
63 after a backtrack. Also, the capture_last value is now reset after a
64 recursion, since all captures are also reset in this case.
65
66 12. The interpreter no longer returns the "too many substrings" error in the
67 case when an overflowing capture is in a branch that is subsequently
68 abandoned after a backtrack.
69
70 13. In the pathological case when an offset vector of size 2 is used, pcretest
71 now prints out the matched string after a yield of 0 or 1.
72
73 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
74 Only supported by the JIT compiler at the moment.
75
76 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
77
78 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
79 offset of the starting point of the matching process, provided the offsets
80 vector is large enough.
81
82 17. The \A escape now records a lookbehind value of 1, though its execution
83 does not actually inspect the previous character. This is to ensure that,
84 in partial multi-segment matching, at least one character from the old
85 segment is retained when a new segment is processed. Otherwise, if there
86 are no lookbehinds in the pattern, \A might match incorrectly at the start
87 of a new segment.
88
89 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
90
91 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
92 gives some modest performance improvement in 8-bit mode.
93
94 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
95 be expressed in certain programming languages using Universal Character
96 Names.
97
98 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
99 which points out that "non characters" are not "characters that may not
100 appear in Unicode strings" but rather "characters that are reserved for
101 internal use and have only local meaning".
102
103 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
104 there was a conditional group that depended on an assertion, if the
105 assertion was false, the callout that immediately followed the alternation
106 in the condition was skipped when pcre_exec() was used for matching.
107
108 23. Allow an explicit callout to be inserted before an assertion that is the
109 condition for a conditional group, for compatibility with automatic
110 callouts, which always insert a callout at this point.
111
112 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
113 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
114
115 25. (*PRUNE) is now supported by the JIT compiler.
116
117 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
118
119 27. Fix the case where there are two or more SKIPs with arguments that may be
120 ignored.
121
122 28. (*SKIP) is now supported by the JIT compiler.
123
124 29. (*THEN) is now supported by the JIT compiler.
125
126 30. Update RunTest with additional test selector options.
127
128 31. The way PCRE handles backtracking verbs has been changed in two ways.
129
130 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
131 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
132 backtracking. In some cases this makes it more Perl-compatible, but Perl's
133 rather obscure rules do not always do the same thing.
134
135 (2) Previously, backtracking verbs were confined within assertions. This is
136 no longer the case for positive assertions, except for (*ACCEPT). Again,
137 this sometimes improves Perl compatibility, and sometimes does not.
138
139 32. A number of tests that were in test 2 because Perl did things differently
140 have been moved to test 1, because either Perl or PCRE has changed, and
141 these tests are now compatible.
142
143 32. Backtracking control verbs are now handled in the same way in JIT and
144 interpreter.
145
146 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
147 contained a forward subroutine reference caused a compile error.
148
149 34. Auto-detect and optimize limited repetitions in JIT.
150
151 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
152 blocking (*UTF) etc.
153
154 36. In the interpreter, maximizing pattern repetitions for characters and
155 character types now use tail recursion, which reduces stack usage.
156
157 37. The value of the max lookbehind was not correctly preserved if a compiled
158 and saved regex was reloaded on a host of different endianness.
159
160 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
161 of the compiled pattern block, expand the flags field from 16 to 32 bits
162 because it was almost full.
163
164 39. Try madvise first before posix_madvise.
165
166 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
167 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
168 original change locked out all matches of empty strings. This has been
169 changed so that one match of an empty string per line is recognized.
170 Subsequent searches on the same line (for colouring or for --only-matching,
171 for example) do not recognize empty strings.
172
173 41. Applied a user patch to fix a number of spelling mistakes in comments.
174
175 42. Data lines longer than 65536 caused pcretest to crash.
176
177 43. Clarified the data type for length and startoffset arguments for pcre_exec
178 and pcre_dfa_exec in the function-specific man pages, where they were
179 explicitly stated to be in bytes, never having been updated. I also added
180 some clarification to the pcreapi man page.
181
182 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
183 a segmentation fault.
184
185
186 Version 8.32 30-November-2012
187 -----------------------------
188
189 1. Improved JIT compiler optimizations for first character search and single
190 character iterators.
191
192 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
193 Patch by Daniel Richard G.
194
195 3. Single character iterator optimizations in the JIT compiler.
196
197 4. Improved JIT compiler optimizations for character ranges.
198
199 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
200 Reported by Giuseppe D'Angelo.
201
202 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
203 of a line, was being set incorrectly in cases where .* appeared inside
204 atomic brackets at the start of a pattern, or where there was a subsequent
205 *PRUNE or *SKIP.
206
207 7. Improved instruction cache flush for POWER/PowerPC.
208 Patch by Daniel Richard G.
209
210 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
211 grep:
212
213 (a) There is now no limit to the number of patterns to be matched.
214
215 (b) An error is given if a pattern is too long.
216
217 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
218 are now supported.
219
220 (d) --exclude-from and --include-from (multiple use) have been added.
221
222 (e) Exclusions and inclusions now apply to all files and directories, not
223 just to those obtained from scanning a directory recursively.
224
225 (f) Multiple uses of -f and --file-list are now supported.
226
227 (g) In a Windows environment, the default for -d has been changed from
228 "read" (the GNU grep default) to "skip", because otherwise the presence
229 of a directory in the file list provokes an error.
230
231 (h) The documentation has been revised and clarified in places.
232
233 9. Improve the matching speed of capturing brackets.
234
235 10. Changed the meaning of \X so that it now matches a Unicode extended
236 grapheme cluster.
237
238 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
239 out POSIX threads when JIT support is configured.
240
241 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
242
243 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
244 structure could go wrong in environments where size_t is not the same size
245 as int.
246
247 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
248 set.
249
250 15. The EBCDIC support had decayed; later updates to the code had included
251 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
252 general tidy up of EBCDIC-related issues, and the documentation was also
253 not quite right. There is now a test that can be run on ASCII systems to
254 check some of the EBCDIC-related things (but is it not a full test).
255
256 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
257 in a small tidy to the code.
258
259 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
260
261 18. If the --only-matching (-o) option in pcregrep is specified multiple
262 times, each one causes appropriate output. For example, -o1 -o2 outputs the
263 substrings matched by the 1st and 2nd capturing parentheses. A separating
264 string can be specified by --om-separator (default empty).
265
266 19. Improving the first n character searches.
267
268 20. Turn case lists for horizontal and vertical white space into macros so that
269 they are defined only once.
270
271 21. This set of changes together give more compatible Unicode case-folding
272 behaviour for characters that have more than one other case when UCP
273 support is available.
274
275 (a) The Unicode property table now has offsets into a new table of sets of
276 three or more characters that are case-equivalent. The MultiStage2.py
277 script that generates these tables (the pcre_ucd.c file) now scans
278 CaseFolding.txt instead of UnicodeData.txt for character case
279 information.
280
281 (b) The code for adding characters or ranges of characters to a character
282 class has been abstracted into a generalized function that also handles
283 case-independence. In UTF-mode with UCP support, this uses the new data
284 to handle characters with more than one other case.
285
286 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
287 whose other case is greater than 256 are now correctly matched
288 caselessly. Previously, the high codepoint matched the low one, but not
289 vice versa.
290
291 (d) The processing of \h, \H, \v, and \ in character classes now makes use
292 of the new class addition function, using character lists defined as
293 macros alongside the case definitions of 20 above.
294
295 (e) Caseless back references now work with characters that have more than
296 one other case.
297
298 (f) General caseless matching of characters with more than one other case
299 is supported.
300
301 22. Unicode character properties were updated from Unicode 6.2.0
302
303 23. Improved CMake support under Windows. Patch by Daniel Richard G.
304
305 24. Add support for 32-bit character strings, and UTF-32
306
307 25. Major JIT compiler update (code refactoring and bugfixing).
308 Experimental Sparc 32 support is added.
309
310 26. Applied a modified version of Daniel Richard G's patch to create
311 pcre.h.generic and config.h.generic by "make" instead of in the
312 PrepareRelease script.
313
314 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
315 pcre_compile.c when checking for a zero character.
316
317 28. Introducing a native interface for JIT. Through this interface, the compiled
318 machine code can be directly executed. The purpose of this interface is to
319 provide fast pattern matching, so several sanity checks are not performed.
320 However, feature tests are still performed. The new interface provides
321 1.4x speedup compared to the old one.
322
323 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
324 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
325 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
326
327 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
328 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
329 the "old" RFC 2279). Instead, it ended up passing a negative length to
330 pcre_exec().
331
332 31. Add support for GCC's visibility feature to hide internal functions.
333
334 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
335 "unknown -C option" after outputting 0 or 1.
336
337 33. There is now support for generating a code coverage report for the test
338 suite in environments where gcc is the compiler and lcov is installed. This
339 is mainly for the benefit of the developers.
340
341 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
342 unaddressable using valgrind annotations, allowing valgrind to detect
343 invalid memory accesses. This is mainly for the benefit of the developers.
344
345 25. (*UTF) can now be used to start a pattern in any of the three libraries.
346
347 26. Give configure error if --enable-cpp but no C++ compiler found.
348
349
350 Version 8.31 06-July-2012
351 -------------------------
352
353 1. Fixing a wrong JIT test case and some compiler warnings.
354
355 2. Removed a bashism from the RunTest script.
356
357 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
358 to unsigned type, result still unsigned" that was given by an MS compiler
359 on encountering the code "-sizeof(xxx)".
360
361 4. Partial matching support is added to the JIT compiler.
362
363 5. Fixed several bugs concerned with partial matching of items that consist
364 of more than one character:
365
366 (a) /^(..)\1/ did not partially match "aba" because checking references was
367 done on an "all or nothing" basis. This also applied to repeated
368 references.
369
370 (b) \R did not give a hard partial match if \r was found at the end of the
371 subject.
372
373 (c) \X did not give a hard partial match after matching one or more
374 characters at the end of the subject.
375
376 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
377 a partial match for the string "\r".
378
379 (e) When newline was set to CRLF, the metacharacter "." did not recognize
380 a partial match for a CR character at the end of the subject string.
381
382 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
383 running pcretest, the text "(JIT)" added to the output whenever JIT is
384 actually used to run the match.
385
386 7. Individual JIT compile options can be set in pcretest by following -s+[+]
387 or /S+[+] with a digit between 1 and 7.
388
389 8. OP_NOT now supports any UTF character not just single-byte ones.
390
391 9. (*MARK) control verb is now supported by the JIT compiler.
392
393 10. The command "./RunTest list" lists the available tests without actually
394 running any of them. (Because I keep forgetting what they all are.)
395
396 11. Add PCRE_INFO_MAXLOOKBEHIND.
397
398 12. Applied a (slightly modified) user-supplied patch that improves performance
399 when the heap is used for recursion (compiled with --disable-stack-for-
400 recursion). Instead of malloc and free for each heap frame each time a
401 logical recursion happens, frames are retained on a chain and re-used where
402 possible. This sometimes gives as much as 30% improvement.
403
404 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
405 call.
406
407 14. As documented, (*COMMIT) is now confined to within a positive assertion.
408
409 15. It is now possible to link pcretest with libedit as an alternative to
410 libreadline.
411
412 16. (*COMMIT) control verb is now supported by the JIT compiler.
413
414 17. The Unicode data tables have been updated to Unicode 6.1.0.
415
416 18. Added --file-list option to pcregrep.
417
418 19. Added binary file support to pcregrep, including the -a, --binary-files,
419 -I, and --text options.
420
421 20. The madvise function is renamed for posix_madvise for QNX compatibility
422 reasons. Fixed by Giuseppe D'Angelo.
423
424 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
425
426 22. Changed the option for creating symbolic links for 16-bit man pages from
427 -s to -sf so that re-installing does not cause issues.
428
429 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
430
431 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
432 matches in certain environments (the workspace was not being correctly
433 retained). Also added to pcre_dfa_exec() a simple plausibility check on
434 some of the workspace data at the beginning of a restart.
435
436 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
437 was not doing so when it should - probably a typo introduced by SVN 528
438 (change 8.10/14).
439
440 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
441 \w+ when the character tables indicated that \x{c4} was a word character.
442 There were several related cases, all because the tests for doing a table
443 lookup were testing for characters less than 127 instead of 255.
444
445 27. If a pattern contains capturing parentheses that are not used in a match,
446 their slots in the ovector are set to -1. For those that are higher than
447 any matched groups, this happens at the end of processing. In the case when
448 there were back references that the ovector was too small to contain
449 (causing temporary malloc'd memory to be used during matching), and the
450 highest capturing number was not used, memory off the end of the ovector
451 was incorrectly being set to -1. (It was using the size of the temporary
452 memory instead of the true size.)
453
454 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
455 ovector size, it uses memory at the end of the block that it has got.
456
457 29. Check for an overlong MARK name and give an error at compile time. The
458 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
459
460 30. JIT compiler update.
461
462 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
463 Rill for the patch.
464
465 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
466
467 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
468
469 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
470 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
471 was enabled.
472
473 35. Improve JIT code generation for greedy plus quantifier.
474
475 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
476 1 to "aa" instead of to an empty string. The bug affected repeated groups
477 that could potentially match an empty string.
478
479 37. Optimizing single character iterators in JIT.
480
481 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
482 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
483 codepoints that are too big for the mode are faulted, and in a UTF mode,
484 disallowed codepoints are also faulted.
485
486 39. If PCRE was compiled with UTF support, in three places in the DFA
487 matcher there was code that should only have been obeyed in UTF mode, but
488 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
489 processing when bytes with values greater than 127 were present. In 16-bit
490 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
491 both cases the values are those that cannot be the first data item in a UTF
492 character. The three items that might have provoked this were recursions,
493 possessively repeated groups, and atomic groups.
494
495 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
496 and pcregrep, because some OS require shared objects to be explicitly
497 passed to ld, causing the link step to fail if they are not.
498
499 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
500 mode, patterns that started with \h* or \R* might be incorrectly matched.
501
502
503 Version 8.30 04-February-2012
504 -----------------------------
505
506 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
507 name is defined in ctype.h.
508
509 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
510 only in quite long subpatterns.
511
512 3. Removed the function pcre_info(), which has been obsolete and deprecated
513 since it was replaced by pcre_fullinfo() in February 2000.
514
515 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
516 match a (*MARK), and the match failed at the start of the subject, a
517 reference to memory before the start of the subject could occur. This bug
518 was introduced by fix 17 of release 8.21.
519
520 5. A reference to an unset group with zero minimum repetition was giving
521 totally wrong answers (in non-JavaScript-compatibility mode). For example,
522 /(another)?(\1?)test/ matched against "hello world test". This bug was
523 introduced in release 8.13.
524
525 6. Add support for 16-bit character strings (a large amount of work involving
526 many changes and refactorings).
527
528 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
529 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
530 from a file.
531
532 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
533 rounding is not applied in this particular case).
534
535 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
536 if they appear, or are escaped, in patterns.
537
538 10. Get rid of a number of -Wunused-but-set-variable warnings.
539
540 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
541 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
542 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
543 also returns the mark "x". This bug applied to capturing parentheses,
544 non-capturing parentheses, and atomic parentheses. It also applied to some
545 assertions.
546
547 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
548 information out of configure.ac instead of relying on pcre.h.generic, which
549 is not stored in the repository.
550
551 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
552 -lreadline.
553
554 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
555
556 15. Applied Graycode's patch to put the top-level frame on the stack rather
557 than the heap when not using the stack for recursion. This gives a
558 performance improvement in many cases when recursion is not deep.
559
560 16. Experimental code added to "pcretest -C" to output the stack frame size.
561
562
563 Version 8.21 12-Dec-2011
564 ------------------------
565
566 1. Updating the JIT compiler.
567
568 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
569 are added as well.
570
571 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
572 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
573 calling _pcre_jit_exec. Some extra comments are added.
574
575 4. (*MARK) settings inside atomic groups that do not contain any capturing
576 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
577 was introduced by change 18 for 8.20.
578
579 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
580 ECMA-262 standard.
581
582 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
583 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
584 This bug was probably introduced by change 9 of 8.13.
585
586 7. While fixing 6 above, I noticed that a number of other items were being
587 incorrectly rejected as "not fixed length". This arose partly because newer
588 opcodes had not been added to the fixed-length checking code. I have (a)
589 corrected the bug and added tests for these items, and (b) arranged for an
590 error to occur if an unknown opcode is encountered while checking for fixed
591 length instead of just assuming "not fixed length". The items that were
592 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
593 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
594 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
595
596 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
597 being incorrectly compiled and would have given unpredicatble results.
598
599 9. A possessively repeated subpattern with minimum repeat count greater than
600 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
601 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
602 the first (A) could occur when it should not.
603
604 10. Add a cast and remove a redundant test from the code.
605
606 11. JIT should use pcre_malloc/pcre_free for allocation.
607
608 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
609 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
610 is anything other than /usr, -L is still shown).
611
612 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
613
614 14. Perl does not support \N without a following name in a [] class; PCRE now
615 also gives an error.
616
617 15. If a forward reference was repeated with an upper limit of around 2000,
618 it caused the error "internal error: overran compiling workspace". The
619 maximum number of forward references (including repeats) was limited by the
620 internal workspace, and dependent on the LINK_SIZE. The code has been
621 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
622 the default depends on LINK_SIZE. There is a new upper limit (for safety)
623 of around 200,000 forward references. While doing this, I also speeded up
624 the filling in of repeated forward references.
625
626 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
627 incorrectly expecting the subject to contain another "a" after the start.
628
629 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
630 in the match, the SKIP should be ignored. This was not happening; instead
631 the SKIP was being treated as NOMATCH. For patterns such as
632 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
633 tested.
634
635 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
636 now much more compatible with Perl, in particular in cases where the result
637 is a non-match for a non-anchored pattern. For example, if
638 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
639 "m", where previously it did not return a name. A side effect of this
640 change is that for partial matches, the last encountered mark name is
641 returned, as for non matches. A number of tests that were previously not
642 Perl-compatible have been moved into the Perl-compatible test files. The
643 refactoring has had the pleasing side effect of removing one argument from
644 the match() function, thus reducing its stack requirements.
645
646 19. If the /S+ option was used in pcretest to study a pattern using JIT,
647 subsequent uses of /S (without +) incorrectly behaved like /S+.
648
649 21. Retrieve executable code size support for the JIT compiler and fixing
650 some warnings.
651
652 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
653 not work when the shorter character appeared right at the end of the
654 subject string.
655
656 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
657 systems.
658
659 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
660 output it when the /M option is used in pcretest.
661
662 25. The CheckMan script was not being included in the distribution. Also, added
663 an explicit "perl" to run Perl scripts from the PrepareRelease script
664 because this is reportedly needed in Windows.
665
666 26. If study data was being save in a file and studying had not found a set of
667 "starts with" bytes for the pattern, the data written to the file (though
668 never used) was taken from uninitialized memory and so caused valgrind to
669 complain.
670
671 27. Updated RunTest.bat as provided by Sheri Pierce.
672
673 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
674
675 29. Computation of memory usage for the table of capturing group names was
676 giving an unnecessarily large value.
677
678
679 Version 8.20 21-Oct-2011
680 ------------------------
681
682 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
683 a POSIX class. After further experiments with Perl, which convinced me that
684 Perl has bugs and confusions, a closing square bracket is no longer allowed
685 in a POSIX name. This bug also affected patterns with classes that started
686 with full stops.
687
688 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
689 captured substring, but while checking the failing first alternative,
690 substring 1 is temporarily captured. If the output vector supplied to
691 pcre_exec() was not big enough for this capture, the yield of the function
692 was still zero ("insufficient space for captured substrings"). This cannot
693 be totally fixed without adding another stack variable, which seems a lot
694 of expense for a edge case. However, I have improved the situation in cases
695 such as /(a)(b)x|abc/ matched against "abc", where the return code
696 indicates that fewer than the maximum number of slots in the ovector have
697 been set.
698
699 3. Related to (2) above: when there are more back references in a pattern than
700 slots in the output vector, pcre_exec() uses temporary memory during
701 matching, and copies in the captures as far as possible afterwards. It was
702 using the entire output vector, but this conflicts with the specification
703 that only 2/3 is used for passing back captured substrings. Now it uses
704 only the first 2/3, for compatibility. This is, of course, another edge
705 case.
706
707 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
708 main code base, and can be used by building with --enable-jit. When this is
709 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
710 runtime --no-jit option is given.
711
712 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
713 ovector, the return from the function was zero, implying that there were
714 other matches that did not fit. The correct "exactly full" value is now
715 returned.
716
717 6. If a subpattern that was called recursively or as a subroutine contained
718 (*PRUNE) or any other control that caused it to give a non-standard return,
719 invalid errors such as "Error -26 (nested recursion at the same subject
720 position)" or even infinite loops could occur.
721
722 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
723 computing the minimum length on reaching *ACCEPT, and so ended up with the
724 wrong value of 1 rather than 0. Further investigation indicates that
725 computing a minimum subject length in the presence of *ACCEPT is difficult
726 (think back references, subroutine calls), and so I have changed the code
727 so that no minimum is registered for a pattern that contains *ACCEPT.
728
729 8. If (*THEN) was present in the first (true) branch of a conditional group,
730 it was not handled as intended. [But see 16 below.]
731
732 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
733 Sheri Pierce.
734
735 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
736 the first byte in a match must be "a".
737
738 11. Change 17 for 8.13 increased the recursion depth for patterns like
739 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
740 pattern contains any instances of (*THEN). If it does not, the old
741 optimizations are restored. It would be nice to do this on a per-group
742 basis, but at the moment that is not feasible.
743
744 12. In some environments, the output of pcretest -C is CRLF terminated. This
745 broke RunTest's code that checks for the link size. A single white space
746 character after the value is now allowed for.
747
748 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
749 For "fr", it uses the Windows-specific input and output files.
750
751 14. If (*THEN) appeared in a group that was called recursively or as a
752 subroutine, it did not work as intended. [But see next item.]
753
754 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
755 pattern fragments (but not containing any | characters). If A and B are
756 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
757 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
758 D. In other words, Perl considers parentheses that do not contain any |
759 characters to be part of a surrounding alternative, whereas PCRE was
760 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
761 differently. PCRE now behaves in the same way as Perl, except in the case
762 of subroutine/recursion calls such as (?1) which have in any case always
763 been different (but PCRE had them first :-).
764
765 16. Related to 15 above: Perl does not treat the | in a conditional group as
766 creating alternatives. Such a group is treated in the same way as an
767 ordinary group without any | characters when processing (*THEN). PCRE has
768 been changed to match Perl's behaviour.
769
770 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
771 RunGrepTest script failed.
772
773 18. Change 22 for version 13 caused atomic groups to use more stack. This is
774 inevitable for groups that contain captures, but it can lead to a lot of
775 stack use in large patterns. The old behaviour has been restored for atomic
776 groups that do not contain any capturing parentheses.
777
778 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
779 suppress the check for a minimum subject length at run time. (If it was
780 given to pcre_exec() or pcre_dfa_exec() it did work.)
781
782 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
783 fail to work when decoding hex characters in data strings in EBCDIC
784 environments.
785
786 21. It appears that in at least one Mac OS environment, the isxdigit() function
787 is implemented as a macro that evaluates to its argument more than once,
788 contravening the C 90 Standard (I haven't checked a later standard). There
789 was an instance in pcretest which caused it to go wrong when processing
790 \x{...} escapes in subject strings. The has been rewritten to avoid using
791 things like p++ in the argument of isxdigit().
792
793
794 Version 8.13 16-Aug-2011
795 ------------------------
796
797 1. The Unicode data tables have been updated to Unicode 6.0.0.
798
799 2. Two minor typos in pcre_internal.h have been fixed.
800
801 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
802 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
803 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
804
805 4. There were a number of related bugs in the code for matching backrefences
806 caselessly in UTF-8 mode when codes for the characters concerned were
807 different numbers of bytes. For example, U+023A and U+2C65 are an upper
808 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
809 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
810 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
811 2-byte code at the end of the subject (it thought there wasn't enough data
812 left).
813
814 5. Comprehensive information about what went wrong is now returned by
815 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
816 as the output vector has at least 2 elements. The offset of the start of
817 the failing character and a reason code are placed in the vector.
818
819 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
820 now returned is for the first byte of the failing character, instead of the
821 last byte inspected. This is an incompatible change, but I hope it is small
822 enough not to be a problem. It makes the returned offset consistent with
823 pcre_exec() and pcre_dfa_exec().
824
825 7. pcretest now gives a text phrase as well as the error number when
826 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
827 failure, the offset and reason code are output.
828
829 8. When \R was used with a maximizing quantifier it failed to skip backwards
830 over a \r\n pair if the subsequent match failed. Instead, it just skipped
831 back over a single character (\n). This seems wrong (because it treated the
832 two characters as a single entity when going forwards), conflicts with the
833 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
834 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
835 has been changed.
836
837 9. Some internal refactoring has changed the processing so that the handling
838 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
839 time (the PCRE_DOTALL option was changed this way some time ago: version
840 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
841 which was always a bit of a fudge. It also means that there is one less
842 argument for the match() function, which reduces its stack requirements
843 slightly. This change also fixes an incompatibility with Perl: the pattern
844 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
845
846 10. More internal refactoring has drastically reduced the number of recursive
847 calls to match() for possessively repeated groups such as (abc)++ when
848 using pcre_exec().
849
850 11. While implementing 10, a number of bugs in the handling of groups were
851 discovered and fixed:
852
853 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
854 (a|)*(?1) gave a compile-time internal error.
855 ((a|)+)+ did not notice that the outer group could match an empty string.
856 (^a|^)+ was not marked as anchored.
857 (.*a|.*)+ was not marked as matching at start or after a newline.
858
859 12. Yet more internal refactoring has removed another argument from the match()
860 function. Special calls to this function are now indicated by setting a
861 value in a variable in the "match data" data block.
862
863 13. Be more explicit in pcre_study() instead of relying on "default" for
864 opcodes that mean there is no starting character; this means that when new
865 ones are added and accidentally left out of pcre_study(), testing should
866 pick them up.
867
868 14. The -s option of pcretest has been documented for ages as being an old
869 synonym of -m (show memory usage). I have changed it to mean "force study
870 for every regex", that is, assume /S for every regex. This is similar to -i
871 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
872 using it. It makes it easier to run collections of tests with and without
873 study enabled, and thereby test pcre_study() more easily. All the standard
874 tests are now run with and without -s (but some patterns can be marked as
875 "never study" - see 20 below).
876
877 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
878 restoration of the capturing data to the outer values was not happening
879 correctly.
880
881 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
882 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
883 pattern had matched an empty string, and so incorrectly returned a no
884 match.
885
886 17. There was optimizing code for the last branch of non-capturing parentheses,
887 and also for the obeyed branch of a conditional subexpression, which used
888 tail recursion to cut down on stack usage. Unfortunately, now that there is
889 the possibility of (*THEN) occurring in these branches, tail recursion is
890 no longer possible because the return has to be checked for (*THEN). These
891 two optimizations have therefore been removed. [But see 8.20/11 above.]
892
893 18. If a pattern containing \R was studied, it was assumed that \R always
894 matched two bytes, thus causing the minimum subject length to be
895 incorrectly computed because \R can also match just one byte.
896
897 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
898 was incorrectly computed.
899
900 20. If /S is present twice on a test pattern in pcretest input, it now
901 *disables* studying, thereby overriding the use of -s on the command line
902 (see 14 above). This is necessary for one or two tests to keep the output
903 identical in both cases.
904
905 21. When (*ACCEPT) was used in an assertion that matched an empty string and
906 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
907
908 22. When an atomic group that contained a capturing parenthesis was
909 successfully matched, but the branch in which it appeared failed, the
910 capturing was not being forgotten if a higher numbered group was later
911 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
912 group 1 to "a", when in fact it should be unset. This applied to multi-
913 branched capturing and non-capturing groups, repeated or not, and also to
914 positive assertions (capturing in negative assertions does not happen
915 in PCRE) and also to nested atomic groups.
916
917 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
918 subject after a captured substring, to make it easier to tell which of a
919 number of identical substrings has been captured.
920
921 24. The way atomic groups are processed by pcre_exec() has been changed so that
922 if they are repeated, backtracking one repetition now resets captured
923 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
924 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
925 "aaa". Previously, it would have been "a". As part of this code
926 refactoring, the way recursive calls are handled has also been changed.
927
928 25. If an assertion condition captured any substrings, they were not passed
929 back unless some other capturing happened later. For example, if
930 (?(?=(a))a) was matched against "a", no capturing was returned.
931
932 26. When studying a pattern that contained subroutine calls or assertions,
933 the code for finding the minimum length of a possible match was handling
934 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
935 group 1 called group 2 while simultaneously a separate group 2 called group
936 1). A stack overflow occurred in this case. I have fixed this by limiting
937 the recursion depth to 10.
938
939 27. Updated RunTest.bat in the distribution to the version supplied by Tom
940 Fortmann. This supports explicit test numbers on the command line, and has
941 argument validation and error reporting.
942
943 28. An instance of \X with an unlimited repeat could fail if at any point the
944 first character it looked at was a mark character.
945
946 29. Some minor code refactoring concerning Unicode properties and scripts
947 should reduce the stack requirement of match() slightly.
948
949 30. Added the '=' option to pcretest to check the setting of unused capturing
950 slots at the end of the pattern, which are documented as being -1, but are
951 not included in the return count.
952
953 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
954 compiled something random. Now it gives a compile-time error (as does
955 Perl).
956
957 32. A *MARK encountered during the processing of a positive assertion is now
958 recorded and passed back (compatible with Perl).
959
960 33. If --only-matching or --colour was set on a pcregrep call whose pattern
961 had alternative anchored branches, the search for a second match in a line
962 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
963 matched the line "0102" twice. The same bug affected patterns that started
964 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
965 twice.
966
967 34. Previously, PCRE did not allow quantification of assertions. However, Perl
968 does, and because of capturing effects, quantifying parenthesized
969 assertions may at times be useful. Quantifiers are now allowed for
970 parenthesized assertions.
971
972 35. A minor code tidy in pcre_compile() when checking options for \R usage.
973
974 36. \g was being checked for fancy things in a character class, when it should
975 just be a literal "g".
976
977 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
978 appearance of a nested POSIX class supersedes an apparent external class.
979 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
980 unescaped square brackets may also appear as part of class names. For
981 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
982 more like Perl. (But see 8.20/1 above.)
983
984 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
985 was because it thought it was \N{name}, which is not supported).
986
987 39. Add minix to OS list not supporting the -S option in pcretest.
988
989 40. PCRE tries to detect cases of infinite recursion at compile time, but it
990 cannot analyze patterns in sufficient detail to catch mutual recursions
991 such as ((?1))((?2)). There is now a runtime test that gives an error if a
992 subgroup is called recursively as a subpattern for a second time at the
993 same position in the subject string. In previous releases this might have
994 been caught by the recursion limit, or it might have run out of stack.
995
996 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
997 happen only once. PCRE was, however incorrectly giving a compile time error
998 "recursive call could loop indefinitely" because it cannot analyze the
999 pattern in sufficient detail. The compile time test no longer happens when
1000 PCRE is compiling a conditional subpattern, but actual runaway loops are
1001 now caught at runtime (see 40 above).
1002
1003 42. It seems that Perl allows any characters other than a closing parenthesis
1004 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1005 has been changed to be the same.
1006
1007 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1008 as not to get warnings when autogen.sh is called. Also changed
1009 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1010
1011 44. To help people who use pcregrep to scan files containing exceedingly long
1012 lines, the following changes have been made:
1013
1014 (a) The default value of the buffer size parameter has been increased from
1015 8K to 20K. (The actual buffer used is three times this size.)
1016
1017 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1018 PCRE is built.
1019
1020 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1021 to be set at run time.
1022
1023 (d) Numerical values in pcregrep options can be followed by K or M, for
1024 example --buffer-size=50K.
1025
1026 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1027 given and the return code is set to 2.
1028
1029 45. Add a pointer to the latest mark to the callout data block.
1030
1031 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1032 partial match of an empty string instead of no match. This was specific to
1033 the use of ".".
1034
1035 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1036 complete match instead of a partial match. This bug was dependent on both
1037 the PCRE_UTF8 and PCRE_DOTALL options being set.
1038
1039 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1040 starting byte set, because \b was not being ignored.
1041
1042
1043 Version 8.12 15-Jan-2011
1044 ------------------------
1045
1046 1. Fixed some typos in the markup of the man pages, and wrote a script that
1047 checks for such things as part of the documentation building process.
1048
1049 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1050 --match-limit and --recursion-limit options (added for 8.11). In
1051 particular, this made one of the standard tests fail. (The integer value
1052 went into the wrong half of a long int.)
1053
1054 3. If the --colour option was given to pcregrep with -v (invert match), it
1055 did strange things, either producing crazy output, or crashing. It should,
1056 of course, ignore a request for colour when reporting lines that do not
1057 match.
1058
1059 4. Another pcregrep bug caused similar problems if --colour was specified with
1060 -M (multiline) and the pattern match finished with a line ending.
1061
1062 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1063 matched in multiline mode, the following line was shown as part of the
1064 match. This seems wrong, so I have changed it.
1065
1066 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1067 the check for further matches in the same line (so they could be coloured)
1068 to overrun the end of the current line. If another match was found, it was
1069 incorrectly shown (and then shown again when found in the next line).
1070
1071 7. If pcregrep was compiled under Windows, there was a reference to the
1072 function pcregrep_exit() before it was defined. I am assuming this was
1073 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1074 reported by a user. I've moved the definition above the reference.
1075
1076
1077 Version 8.11 10-Dec-2010
1078 ------------------------
1079
1080 1. (*THEN) was not working properly if there were untried alternatives prior
1081 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1082 backtracked to try for "b" instead of moving to the next alternative branch
1083 at the same level (in this case, to look for "c"). The Perl documentation
1084 is clear that when (*THEN) is backtracked onto, it goes to the "next
1085 alternative in the innermost enclosing group".
1086
1087 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1088 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1089 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1090 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1091 (*THEN).
1092
1093 3. If \s appeared in a character class, it removed the VT character from
1094 the class, even if it had been included by some previous item, for example
1095 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1096 of \s, but is part of the POSIX "space" class.)
1097
1098 4. A partial match never returns an empty string (because you can always
1099 match an empty string at the end of the subject); however the checking for
1100 an empty string was starting at the "start of match" point. This has been
1101 changed to the "earliest inspected character" point, because the returned
1102 data for a partial match starts at this character. This means that, for
1103 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1104 (previously it gave "no match").
1105
1106 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1107 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1108 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1109 has an implication that the given string is incomplete (because a partial
1110 match is preferred over a full match). For this reason, these items now
1111 give a partial match in this situation. [Aside: previously, the one case
1112 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1113 match rather than a full match, which was wrong by the old rules, but is
1114 now correct.]
1115
1116 6. There was a bug in the handling of #-introduced comments, recognized when
1117 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1118 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1119 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1120 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1121 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1122 places in pcre_compile().
1123
1124 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1125 comments when looking ahead for named forward references to subpatterns,
1126 the only newline sequence it recognized was NL. It now handles newlines
1127 according to the set newline convention.
1128
1129 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1130 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1131 cater for a lack of strerror(). These oversights have been fixed.
1132
1133 9. Added --match-limit and --recursion-limit to pcregrep.
1134
1135 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1136
1137 11. When the -o option was used, pcregrep was setting a return code of 1, even
1138 when matches were found, and --line-buffered was not being honoured.
1139
1140 12. Added an optional parentheses number to the -o and --only-matching options
1141 of pcregrep.
1142
1143 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1144 can match an empty string. The code to do it in pcretest and pcredemo
1145 needed fixing:
1146
1147 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1148 only one byte after an empty string match just before CRLF (this case
1149 just got forgotten; "any" and "anycrlf" were OK).
1150
1151 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1152 mode when an empty string match preceded an ASCII character followed by
1153 a non-ASCII character. (The code for advancing by one character rather
1154 than one byte was nonsense.)
1155
1156 (c) The pcredemo.c sample program did not have any code at all to handle
1157 the cases when CRLF is a valid newline sequence.
1158
1159 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1160 as a starting offset was within the subject string. There is now a new
1161 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1162 negative or greater than the length of the string. In order to test this,
1163 pcretest is extended to allow the setting of negative starting offsets.
1164
1165 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1166 starting offset points to the beginning of a UTF-8 character was
1167 unnecessarily clumsy. I tidied it up.
1168
1169 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1170 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1171
1172 17. Nobody had reported that the --include_dir option, which was added in
1173 release 7.7 should have been called --include-dir (hyphen, not underscore)
1174 for compatibility with GNU grep. I have changed it to --include-dir, but
1175 left --include_dir as an undocumented synonym, and the same for
1176 --exclude-dir, though that is not available in GNU grep, at least as of
1177 release 2.5.4.
1178
1179 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1180 characters from a string of bytes) have been redefined so as not to use
1181 loops, in order to improve performance in some environments. At the same
1182 time, I abstracted some of the common code into auxiliary macros to save
1183 repetition (this should not affect the compiled code).
1184
1185 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1186 compile-time error is now given if \c is not followed by an ASCII
1187 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1188 different, and any byte value is allowed.)
1189
1190 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1191 START_OPTIMIZE option, which is now allowed at compile time - but just
1192 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1193 to pcregrep and other applications that have no direct access to PCRE
1194 options. The new /Y option in pcretest sets this option when calling
1195 pcre_compile().
1196
1197 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1198 back references. Groups containing recursive back references were forced to
1199 be atomic by that change, but in the case of named groups, the amount of
1200 memory required was incorrectly computed, leading to "Failed: internal
1201 error: code overflow". This has been fixed.
1202
1203 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1204 pcretest.c, to avoid build problems in some Borland environments.
1205
1206
1207 Version 8.10 25-Jun-2010
1208 ------------------------
1209
1210 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1211 THEN.
1212
1213 2. (*ACCEPT) was not working when inside an atomic group.
1214
1215 3. Inside a character class, \B is treated as a literal by default, but
1216 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1217 causes the error). The code is unchanged, but I tidied the documentation.
1218
1219 4. Inside a character class, PCRE always treated \R and \X as literals,
1220 whereas Perl faults them if its -w option is set. I have changed PCRE so
1221 that it faults them when PCRE_EXTRA is set.
1222
1223 5. Added support for \N, which always matches any character other than
1224 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1225
1226 6. When compiling pcregrep with newer versions of gcc which may have
1227 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1228 declared with attribute warn_unused_result" were given. Just casting the
1229 result to (void) does not stop the warnings; a more elaborate fudge is
1230 needed. I've used a macro to implement this.
1231
1232 7. Minor change to pcretest.c to avoid a compiler warning.
1233
1234 8. Added four artifical Unicode properties to help with an option to make
1235 \s etc use properties (see next item). The new properties are: Xan
1236 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1237
1238 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1239 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1240 this option. Modified pcretest to add /W to test this facility. Added
1241 REG_UCP to make it available via the POSIX interface.
1242
1243 10. Added --line-buffered to pcregrep.
1244
1245 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1246 studied, and the match started with a letter with a code point greater than
1247 127 whose first byte was different to the first byte of the other case of
1248 the letter, the other case of this starting letter was not recognized
1249 (#976).
1250
1251 12. If a pattern that was studied started with a repeated Unicode property
1252 test, for example, \p{Nd}+, there was the theoretical possibility of
1253 setting up an incorrect bitmap of starting bytes, but fortunately it could
1254 not have actually happened in practice until change 8 above was made (it
1255 added property types that matched character-matching opcodes).
1256
1257 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1258 possible starting bytes for non-anchored patterns.
1259
1260 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1261 \R, and also a number of cases that involve Unicode properties, both
1262 explicit and implicit when PCRE_UCP is set.
1263
1264 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1265 input, it could crash or give wrong results if characters with values
1266 greater than 0xc0 were present in the subject string. (Detail: it assumed
1267 UTF-8 input when processing these items.)
1268
1269 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1270 size_t is 64-bit (#991).
1271
1272 17. Added a check for running out of memory when PCRE is compiled with
1273 --disable-stack-for-recursion (#990).
1274
1275 18. If the last data line in a file for pcretest does not have a newline on
1276 the end, a newline was missing in the output.
1277
1278 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1279 less than 128) in its various bitmaps. However, there is a facility for
1280 generating tables according to the current locale when PCRE is compiled. It
1281 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1282 characters, are recognized by isspace() and therefore were getting set in
1283 these tables, and indeed these tables seem to approximate to ISO 8859. This
1284 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1285 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1286 which of course cannot start UTF-8 characters. I have changed the code so
1287 that only real ASCII characters (less than 128) and the correct starting
1288 bytes for UTF-8 encodings are set for characters greater than 127 when in
1289 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1290 altogether.)
1291
1292 20. Added the /T option to pcretest so as to be able to run tests with non-
1293 standard character tables, thus making it possible to include the tests
1294 used for 19 above in the standard set of tests.
1295
1296 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1297 reference to a subpattern the other side of a comment that contains an
1298 opening parenthesis caused either an internal compiling error, or a
1299 reference to the wrong subpattern.
1300
1301
1302 Version 8.02 19-Mar-2010
1303 ------------------------
1304
1305 1. The Unicode data tables have been updated to Unicode 5.2.0.
1306
1307 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1308 configured.
1309
1310 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1311 original author of that file, following a query about its status.
1312
1313 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1314 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1315
1316 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1317 quantifier applied to a forward-referencing subroutine call, could compile
1318 incorrect code or give the error "internal error: previously-checked
1319 referenced subpattern not found".
1320
1321 6. Both MS Visual Studio and Symbian OS have problems with initializing
1322 variables to point to external functions. For these systems, therefore,
1323 pcre_malloc etc. are now initialized to local functions that call the
1324 relevant global functions.
1325
1326 7. There were two entries missing in the vectors called coptable and poptable
1327 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1328 I've fixed the data, and added a kludgy way of testing at compile time that
1329 the lengths are correct (equal to the number of opcodes).
1330
1331 8. Following on from 7, I added a similar kludge to check the length of the
1332 eint vector in pcreposix.c.
1333
1334 9. Error texts for pcre_compile() are held as one long string to avoid too
1335 much relocation at load time. To find a text, the string is searched,
1336 counting zeros. There was no check for running off the end of the string,
1337 which could happen if a new error number was added without updating the
1338 string.
1339
1340 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1341
1342 11. \K was not working if it appeared in an atomic group or in a group that
1343 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1344 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1345 the assertion is positive, but not if it is negative.
1346
1347 12. Change 11 fortuitously reduced the size of the stack frame used in the
1348 "match()" function of pcre_exec.c by one pointer. Forthcoming
1349 implementation of support for (*MARK) will need an extra pointer on the
1350 stack; I have reserved it now, so that the stack frame size does not
1351 decrease.
1352
1353 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1354 item in branch that calls a recursion is a subroutine call - as in the
1355 second branch in the above example - was incorrectly given the compile-
1356 time error "recursive call could loop indefinitely" because pcre_compile()
1357 was not correctly checking the subroutine for matching a non-empty string.
1358
1359 14. The checks for overrunning compiling workspace could trigger after an
1360 overrun had occurred. This is a "should never occur" error, but it can be
1361 triggered by pathological patterns such as hundreds of nested parentheses.
1362 The checks now trigger 100 bytes before the end of the workspace.
1363
1364 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1365
1366
1367 Version 8.01 19-Jan-2010
1368 ------------------------
1369
1370 1. If a pattern contained a conditional subpattern with only one branch (in
1371 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1372 computed the wrong minimum data length (which is of course zero for such
1373 subpatterns). This could cause incorrect "no match" results.
1374
1375 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1376 the pattern is reset in the first branch, pcre_compile() failed with
1377 "internal error: code overflow at offset...". This happened only when
1378 the reset was to the original external option setting. (An optimization
1379 abstracts leading options settings into an external setting, which was the
1380 cause of this.)
1381
1382 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1383 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1384 assertion pattern did not match (meaning that the assertion was true), it
1385 was incorrectly treated as false if the SKIP had been reached during the
1386 matching. This also applied to assertions used as conditions.
1387
1388 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1389 assertion subpattern, including such a pattern used as a condition,
1390 unpredictable results occurred, instead of the error return
1391 PCRE_ERROR_DFA_UITEM.
1392
1393 5. The C++ GlobalReplace function was not working like Perl for the special
1394 situation when an empty string is matched. It now does the fancy magic
1395 stuff that is necessary.
1396
1397 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1398 removed. (These were left over from very, very early versions of PCRE.)
1399
1400 7. Some cosmetic changes to the code to make life easier when compiling it
1401 as part of something else:
1402
1403 (a) Change DEBUG to PCRE_DEBUG.
1404
1405 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1406 called "current" as "current_branch", to prevent a collision with the
1407 Linux macro when compiled as a kernel module.
1408
1409 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1410 prevent a collision with the Linux macro when compiled as a kernel
1411 module.
1412
1413 8. In pcre_compile() there are some checks for integer overflows that used to
1414 cast potentially large values to (double). This has been changed to that
1415 when building, a check for int64_t is made, and if it is found, it is used
1416 instead, thus avoiding the use of floating point arithmetic. (There is no
1417 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1418 double.
1419
1420 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1421 2005 (difference between two addresses compared to an unsigned value).
1422
1423 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1424 custom one, because of the following reported problem in Windows:
1425
1426 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1427 under Win32.
1428 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1429 therefore missing the function definition.
1430 - The compiler thus generates a "C" signature for the test function.
1431 - The linker fails to find the "C" function.
1432 - PCRE fails to configure if asked to do so against libbz2.
1433
1434 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1435 messages were output:
1436
1437 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1438 rerunning libtoolize, to keep the correct libtool macros in-tree.
1439 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1440
1441 I have done both of these things.
1442
1443 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1444 most of the time, it *can* run out if it is given a pattern that contains a
1445 runaway infinite recursion. I updated the discussion in the pcrestack man
1446 page.
1447
1448 13. Now that we have gone to the x.xx style of version numbers, the minor
1449 version may start with zero. Using 08 or 09 is a bad idea because users
1450 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1451 interpreted as invalid octal numbers. I've updated the previous comment in
1452 configure.ac, and also added a check that gives an error if 08 or 09 are
1453 used.
1454
1455 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1456 causing partial matching to fail when the end of the subject matched \W
1457 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1458
1459 15. There were some discrepancies between the declarations in pcre_internal.h
1460 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1461 their definitions. The declarations used "const uschar *" and the
1462 definitions used USPTR. Even though USPTR is normally defined as "const
1463 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1464 reported that: "This difference in casting confuses some C++ compilers, for
1465 example, SunCC recognizes above declarations as different functions and
1466 generates broken code for hbpcre." I have changed the declarations to use
1467 USPTR.
1468
1469 16. GNU libtool is named differently on some systems. The autogen.sh script now
1470 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1471 (FreeBSD).
1472
1473 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1474 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1475 comment: "Figure out how to create a longlong from a string: strtoll and
1476 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1477 instance, but it only takes 2 args instead of 3!"
1478
1479 18. A subtle bug concerned with back references has been fixed by a change of
1480 specification, with a corresponding code fix. A pattern such as
1481 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1482 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1483 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1484 same bug. Such groups have to be quantified to be useful, or contained
1485 inside another quantified group. (If there's no repetition, the reference
1486 can never match.) The problem arises because, having left the group and
1487 moved on to the rest of the pattern, a later failure that backtracks into
1488 the group uses the captured value from the final iteration of the group
1489 rather than the correct earlier one. I have fixed this in PCRE by forcing
1490 any group that contains a reference to itself to be an atomic group; that
1491 is, there cannot be any backtracking into it once it has completed. This is
1492 similar to recursive and subroutine calls.
1493
1494
1495 Version 8.00 19-Oct-09
1496 ----------------------
1497
1498 1. The table for translating pcre_compile() error codes into POSIX error codes
1499 was out-of-date, and there was no check on the pcre_compile() error code
1500 being within the table. This could lead to an OK return being given in
1501 error.
1502
1503 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1504 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1505 in a Windows environment.
1506
1507 3. The pcregrep --count option prints the count for each file even when it is
1508 zero, as does GNU grep. However, pcregrep was also printing all files when
1509 --files-with-matches was added. Now, when both options are given, it prints
1510 counts only for those files that have at least one match. (GNU grep just
1511 prints the file name in this circumstance, but including the count seems
1512 more useful - otherwise, why use --count?) Also ensured that the
1513 combination -clh just lists non-zero counts, with no names.
1514
1515 4. The long form of the pcregrep -F option was incorrectly implemented as
1516 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1517 but it seems right to fix it, and I didn't think it was worth preserving
1518 the old behaviour.
1519
1520 5. The command line items --regex=pattern and --regexp=pattern were not
1521 recognized by pcregrep, which required --regex pattern or --regexp pattern
1522 (with a space rather than an '='). The man page documented the '=' forms,
1523 which are compatible with GNU grep; these now work.
1524
1525 6. No libpcreposix.pc file was created for pkg-config; there was just
1526 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1527
1528 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1529 when UCP support is not needed, by modifying the Python script that
1530 generates it from Unicode data files. This should not matter if the module
1531 is correctly used as a library, but I received one complaint about 50K of
1532 unwanted data. My guess is that the person linked everything into his
1533 program rather than using a library. Anyway, it does no harm.
1534
1535 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1536 was a minimum greater than 1 for a wide character in a possessive
1537 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1538 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1539 character. Chaos in the form of incorrect output or a compiling loop could
1540 result.
1541
1542 9. The restrictions on what a pattern can contain when partial matching is
1543 requested for pcre_exec() have been removed. All patterns can now be
1544 partially matched by this function. In addition, if there are at least two
1545 slots in the offset vector, the offset of the earliest inspected character
1546 for the match and the offset of the end of the subject are set in them when
1547 PCRE_ERROR_PARTIAL is returned.
1548
1549 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1550 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1551 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1552 and may be more useful for multi-segment matching.
1553
1554 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1555 used to be given if ever the end of the subject was reached; now it is
1556 given only if matching could not proceed because another character was
1557 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1558 string "Z", which now yields "no match" instead of "partial match". In the
1559 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1560 final character ended with (*FAIL).
1561
1562 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1563 if the pattern had a "must contain" character that was already found in the
1564 earlier partial match, unless partial matching was again requested. For
1565 example, with the pattern /dog.(body)?/, the "must contain" character is
1566 "g". If the first part-match was for the string "dog", restarting with
1567 "sbody" failed. This bug has been fixed.
1568
1569 13. The string returned by pcre_dfa_exec() after a partial match has been
1570 changed so that it starts at the first inspected character rather than the
1571 first character of the match. This makes a difference only if the pattern
1572 starts with a lookbehind assertion or \b or \B (\K is not supported by
1573 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1574 matching functions compatible, and I think it's the right thing to do.
1575
1576 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1577 so that the demonstration program is easily available in environments where
1578 PCRE has not been installed from source.
1579
1580 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1581 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1582 library.
1583
1584 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1585 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1586 is not the first non-POSIX option to be added. Clearly some people find
1587 these options useful.
1588
1589 17. If a caller to the POSIX matching function regexec() passes a non-zero
1590 value for nmatch with a NULL value for pmatch, the value of
1591 nmatch is forced to zero.
1592
1593 18. RunGrepTest did not have a test for the availability of the -u option of
1594 the diff command, as RunTest does. It now checks in the same way as
1595 RunTest, and also checks for the -b option.
1596
1597 19. If an odd number of negated classes containing just a single character
1598 interposed, within parentheses, between a forward reference to a named
1599 subpattern and the definition of the subpattern, compilation crashed with
1600 an internal error, complaining that it could not find the referenced
1601 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1602 [The bug was that it was starting one character too far in when skipping
1603 over the character class, thus treating the ] as data rather than
1604 terminating the class. This meant it could skip too much.]
1605
1606 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1607 /g option in pcretest when the pattern contains \K, which makes it possible
1608 to have an empty string match not at the start, even when the pattern is
1609 anchored. Updated pcretest and pcredemo to use this option.
1610
1611 21. If the maximum number of capturing subpatterns in a recursion was greater
1612 than the maximum at the outer level, the higher number was returned, but
1613 with unset values at the outer level. The correct (outer level) value is
1614 now given.
1615
1616 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1617 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1618 make it do so. The string so far is captured, making this feature
1619 compatible with Perl.
1620
1621 23. The tests have been re-organized, adding tests 11 and 12, to make it
1622 possible to check the Perl 5.10 features against Perl 5.10.
1623
1624 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1625 pattern matches a fixed length string. PCRE did not allow this; now it
1626 does. Neither allows recursion.
1627
1628 25. I finally figured out how to implement a request to provide the minimum
1629 length of subject string that was needed in order to match a given pattern.
1630 (It was back references and recursion that I had previously got hung up
1631 on.) This code has now been added to pcre_study(); it finds a lower bound
1632 to the length of subject needed. It is not necessarily the greatest lower
1633 bound, but using it to avoid searching strings that are too short does give
1634 some useful speed-ups. The value is available to calling programs via
1635 pcre_fullinfo().
1636
1637 26. While implementing 25, I discovered to my embarrassment that pcretest had
1638 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1639 study optimizations had never been tested with that matching function.
1640 Oops. What is worse, even when it was passed study data, there was a bug in
1641 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1642 were also very few tests of studied patterns with pcre_dfa_exec().
1643
1644 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1645 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1646 on the other side of the coin, they are no longer allowed to have different
1647 names, because these cannot be distinguished in PCRE, and this has caused
1648 confusion. (This is a difference from Perl.)
1649
1650 28. When duplicate subpattern names are present (necessarily with different
1651 numbers, as required by 27 above), and a test is made by name in a
1652 conditional pattern, either for a subpattern having been matched, or for
1653 recursion in such a pattern, all the associated numbered subpatterns are
1654 tested, and the overall condition is true if the condition is true for any
1655 one of them. This is the way Perl works, and is also more like the way
1656 testing by number works.
1657
1658
1659 Version 7.9 11-Apr-09
1660 ---------------------
1661
1662 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1663 (pcretest), all targets were linked against these libraries. This included
1664 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1665 libraries. This caused unwanted dependencies to be created. This problem
1666 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1667 pcretest is linked with readline.
1668
1669 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1670 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1671 moved outside it again, because FALSE and TRUE are already defined in AIX,
1672 but BOOL is not.
1673
1674 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1675 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1676
1677 4. The pcregrep documentation said spaces were inserted as well as colons (or
1678 hyphens) following file names and line numbers when outputting matching
1679 lines. This is not true; no spaces are inserted. I have also clarified the
1680 wording for the --colour (or --color) option.
1681
1682 5. In pcregrep, when --colour was used with -o, the list of matching strings
1683 was not coloured; this is different to GNU grep, so I have changed it to be
1684 the same.
1685
1686 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1687 each matching line was coloured. Now it goes on to look for further matches
1688 of any of the test patterns, which is the same behaviour as GNU grep.
1689
1690 7. A pattern that could match an empty string could cause pcregrep to loop; it
1691 doesn't make sense to accept an empty string match in pcregrep, so I have
1692 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1693 seems to be how GNU grep behaves. [But see later change 40 for release
1694 8.33.]
1695
1696 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1697 start or after a newline", because the conditional assertion was not being
1698 correctly handled. The rule now is that both the assertion and what follows
1699 in the first alternative must satisfy the test.
1700
1701 9. If auto-callout was enabled in a pattern with a conditional group whose
1702 condition was an assertion, PCRE could crash during matching, both with
1703 pcre_exec() and pcre_dfa_exec().
1704
1705 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1706 used for matching.
1707
1708 11. Unicode property support in character classes was not working for
1709 characters (bytes) greater than 127 when not in UTF-8 mode.
1710
1711 12. Added the -M command line option to pcretest.
1712
1713 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1714
1715 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1716
1717 16. Added comments and documentation about mis-use of no_arg in the C++
1718 wrapper.
1719
1720 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1721 from Martin Jerabek that uses macro names for all relevant character and
1722 string constants.
1723
1724 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1725 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1726 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1727 these, but not everybody uses configure.
1728
1729 19. A conditional group that had only one branch was not being correctly
1730 recognized as an item that could match an empty string. This meant that an
1731 enclosing group might also not be so recognized, causing infinite looping
1732 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1733 with the subject "ab", where knowledge that the repeated group can match
1734 nothing is needed in order to break the loop.
1735
1736 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1737 exec(), but without supplying a callout function, matching went wrong.
1738
1739 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1740 leak if the size of the offset vector was greater than 30. When the vector
1741 is smaller, the saved offsets during recursion go onto a local stack
1742 vector, but for larger vectors malloc() is used. It was failing to free
1743 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1744 error, in fact).
1745
1746 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1747 heapframe that is used only when UTF-8 support is enabled. This caused no
1748 problem, but was untidy.
1749
1750 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1751 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1752 included within another project.
1753
1754 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1755 slightly modified by me:
1756
1757 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1758 not building pcregrep.
1759
1760 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1761 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1762
1763 25. Forward references, both numeric and by name, in patterns that made use of
1764 duplicate group numbers, could behave incorrectly or give incorrect errors,
1765 because when scanning forward to find the reference group, PCRE was not
1766 taking into account the duplicate group numbers. A pattern such as
1767 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1768
1769 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1770 the feature of a custom pointer more persuasive (as requested by a user).
1771
1772 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1773 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1774 pre-defined.
1775
1776 28. Added support for (*UTF8) at the start of a pattern.
1777
1778 29. Arrange for flags added by the "release type" setting in CMake to be shown
1779 in the configuration summary.
1780
1781
1782 Version 7.8 05-Sep-08
1783 ---------------------
1784
1785 1. Replaced UCP searching code with optimized version as implemented for Ad
1786 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1787 stage table and inline lookup instead of a function, giving speed ups of 2
1788 to 5 times on some simple patterns that I tested. Permission was given to
1789 distribute the MultiStage2.py script that generates the tables (it's not in
1790 the tarball, but is in the Subversion repository).
1791
1792 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1793 scripts.
1794
1795 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1796 a group with a zero qualifier. The result of the study could be incorrect,
1797 or the function might crash, depending on the pattern.
1798
1799 4. Caseless matching was not working for non-ASCII characters in back
1800 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1801 It now works when Unicode Property Support is available.
1802
1803 5. In pcretest, an escape such as \x{de} in the data was always generating
1804 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1805 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1806 truncation.
1807
1808 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1809
1810 7. Added two (int) casts to pcregrep when printing the difference of two
1811 pointers, in case they are 64-bit values.
1812
1813 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1814 test 2 if it fails.
1815
1816 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1817 and a #define of that name to empty if it is not externally set. This is to
1818 allow users of MSVC to set it if necessary.
1819
1820 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1821 the convenience functions in the pcre_get.c source file.
1822
1823 11. An option change at the start of a pattern that had top-level alternatives
1824 could cause overwriting and/or a crash. This command provoked a crash in
1825 some environments:
1826
1827 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1828
1829 This potential security problem was recorded as CVE-2008-2371.
1830
1831 12. For a pattern where the match had to start at the beginning or immediately
1832 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1833 pcre_dfa_exec() could read past the end of the passed subject if there was
1834 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1835 pcretest so that it places the subject at the end of its malloc-ed buffer.
1836
1837 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1838 exec() might read past the end of the data buffer in UTF-8 mode.
1839
1840 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1841 the data contained the byte 0x85 as part of a UTF-8 character within its
1842 first line. This applied both to normal and DFA matching.
1843
1844 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1845 /^[^d]*?$/8 failed to match "abc".
1846
1847 16. Added a missing copyright notice to pcrecpp_internal.h.
1848
1849 17. Make it more clear in the documentation that values returned from
1850 pcre_exec() in ovector are byte offsets, not character counts.
1851
1852 18. Tidied a few places to stop certain compilers from issuing warnings.
1853
1854 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1855 supplied by Stefan Weber. I made a further small update for 7.8 because
1856 there is a change of source arrangements: the pcre_searchfuncs.c module is
1857 replaced by pcre_ucd.c.
1858
1859
1860 Version 7.7 07-May-08
1861 ---------------------
1862
1863 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1864 a string to a long long, pretend we don't even have a long long." This is
1865 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1866
1867 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1868 pre-7.6 versions, which defined a global no_arg variable instead of putting
1869 it in the RE class. (See also #8 below.)
1870
1871 3. Remove a line of dead code, identified by coverity and reported by Nuno
1872 Lopes.
1873
1874 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1875
1876 (1) The include/exclude patterns were being applied to the whole pathnames
1877 of files, instead of just to the final components.
1878
1879 (2) If there was more than one level of directory, the subdirectories were
1880 skipped unless they satisfied the include/exclude conditions. This is
1881 inconsistent with GNU grep (and could even be seen as contrary to the
1882 pcregrep specification - which I improved to make it absolutely clear).
1883 The action now is always to scan all levels of directory, and just
1884 apply the include/exclude patterns to regular files.
1885
1886 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1887 --exclude_dir in the tests to avoid scanning .svn directories.
1888
1889 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1890 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1891 doesn't support NULs in patterns.
1892
1893 7. Added some missing "const"s to declarations of static tables in
1894 pcre_compile.c and pcre_dfa_exec.c.
1895
1896 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1897 caused by fix #2 above. (Subsequently also a second patch to fix the
1898 first patch. And a third patch - this was a messy problem.)
1899
1900 9. Applied Craig's patch to remove the use of push_back().
1901
1902 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1903 matching function regexec().
1904
1905 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1906 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1907 references. PCRE supports relative numbers with this syntax (I don't think
1908 Oniguruma does).
1909
1910 12. Previously, a group with a zero repeat such as (...){0} was completely
1911 omitted from the compiled regex. However, this means that if the group
1912 was called as a subroutine from elsewhere in the pattern, things went wrong
1913 (an internal error was given). Such groups are now left in the compiled
1914 pattern, with a new opcode that causes them to be skipped at execution
1915 time.
1916
1917 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1918 to the way PCRE behaves:
1919
1920 (a) A lone ] character is dis-allowed (Perl treats it as data).
1921
1922 (b) A back reference to an unmatched subpattern matches an empty string
1923 (Perl fails the current match path).
1924
1925 (c) A data ] in a character class must be notated as \] because if the
1926 first data character in a class is ], it defines an empty class. (In
1927 Perl it is not possible to have an empty class.) The empty class []
1928 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1929 The negative empty class [^] matches any one character, independently
1930 of the DOTALL setting.
1931
1932 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1933 non-existent subpattern following a character class starting with ']' and
1934 containing () gave an internal compiling error instead of "reference to
1935 non-existent subpattern". Fortunately, when the pattern did exist, the
1936 compiled code was correct. (When scanning forwards to check for the
1937 existence of the subpattern, it was treating the data ']' as terminating
1938 the class, so got the count wrong. When actually compiling, the reference
1939 was subsequently set up correctly.)
1940
1941 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1942 it was being rejected as not supported by pcre_dfa_exec(), even though
1943 other assertions are supported. I have made pcre_dfa_exec() support
1944 (*FAIL).
1945
1946 16. The implementation of 13c above involved the invention of a new opcode,
1947 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1948 cannot be changed at match time, I realized I could make a small
1949 improvement to matching performance by compiling OP_ALLANY instead of
1950 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1951 on the OP_ANY path.
1952
1953 17. Compiling pcretest on Windows with readline support failed without the
1954 following two fixes: (1) Make the unistd.h include conditional on
1955 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1956
1957 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1958 ncurses library to be included for pcretest when ReadLine support is
1959 requested, but also to allow for it to be overridden. This patch came from
1960 Daniel Bergström.
1961
1962 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1963 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1964 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1965 spotting this.
1966
1967
1968 Version 7.6 28-Jan-08
1969 ---------------------
1970
1971 1. A character class containing a very large number of characters with
1972 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1973 overflow.
1974
1975 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1976 HAVE_LONG_LONG is not defined.
1977
1978 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1979 bring them up to date and include new features. This patch includes:
1980
1981 - Fixed PH's badly added libz and libbz2 support.
1982 - Fixed a problem with static linking.
1983 - Added pcredemo. [But later removed - see 7 below.]
1984 - Fixed dftables problem and added an option.
1985 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1986 HAVE_LONG_LONG.
1987 - Added readline support for pcretest.
1988 - Added an listing of the option settings after cmake has run.
1989
1990 4. A user submitted a patch to Makefile that makes it easy to create
1991 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1992 Makefile.am that cause it to include this special target, without
1993 affecting anything else. Note that the same mingw target plus all
1994 the other distribution libraries and programs are now supported
1995 when configuring with CMake (see 6 below) instead of with
1996 Configure/Make.
1997
1998 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1999 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2000 exported in the Windows port". It has not yet been confirmed that the patch
2001 solves the problem, but it does no harm.
2002
2003 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2004 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2005 with CMake, and also correct the comment about stack recursion.
2006
2007 7. Remove the automatic building of pcredemo from the ./configure system and
2008 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2009 of a program that users should build themselves after PCRE is installed, so
2010 building it automatically is not really right. What is more, it gave
2011 trouble in some build environments.
2012
2013 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2014
2015
2016 Version 7.5 10-Jan-08
2017 ---------------------
2018
2019 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2020 values in parens when parsing an RE using the C++ wrapper."
2021
2022 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2023 Characters greater than 255 were excluded from the class instead of being
2024 included.
2025
2026 3. The same bug as (2) above applied to negated POSIX classes such as
2027 [:^space:].
2028
2029 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2030 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2031 I have changed it.
2032
2033 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2034 first named subpattern) and a construct such as (?&a) would reference the
2035 first named subpattern whose name started with "a" (in other words, the
2036 length check was missing). Both these problems are fixed. "Subpattern name
2037 expected" is now given for (?&) (a zero-length name), and this patch also
2038 makes it give the same error for \k'' (previously it complained that that
2039 was a reference to a non-existent subpattern).
2040
2041 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2042 this is right because (?- can be followed by option settings as well as by
2043 digits. I have, however, made the messages clearer.
2044
2045 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2046 than the number used in the conditional) now cause a compile-time error.
2047 This is actually not compatible with Perl, which accepts such patterns, but
2048 treats the conditional as always being FALSE (as PCRE used to), but it
2049 seems to me that giving a diagnostic is better.
2050
2051 8. Change "alphameric" to the more common word "alphanumeric" in comments
2052 and messages.
2053
2054 9. Fix two occurrences of "backslash" in comments that should have been
2055 "backspace".
2056
2057 10. Remove two redundant lines of code that can never be obeyed (their function
2058 was moved elsewhere).
2059
2060 11. The program that makes PCRE's Unicode character property table had a bug
2061 which caused it to generate incorrect table entries for sequences of
2062 characters that have the same character type, but are in different scripts.
2063 It amalgamated them into a single range, with the script of the first of
2064 them. In other words, some characters were in the wrong script. There were
2065 thirteen such cases, affecting characters in the following ranges:
2066
2067 U+002b0 - U+002c1
2068 U+0060c - U+0060d
2069 U+0061e - U+00612
2070 U+0064b - U+0065e
2071 U+0074d - U+0076d
2072 U+01800 - U+01805
2073 U+01d00 - U+01d77
2074 U+01d9b - U+01dbf
2075 U+0200b - U+0200f
2076 U+030fc - U+030fe
2077 U+03260 - U+0327f
2078 U+0fb46 - U+0fbb1
2079 U+10450 - U+1049d
2080
2081 12. The -o option (show only the matching part of a line) for pcregrep was not
2082 compatible with GNU grep in that, if there was more than one match in a
2083 line, it showed only the first of them. It now behaves in the same way as
2084 GNU grep.
2085
2086 13. If the -o and -v options were combined for pcregrep, it printed a blank
2087 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2088 does the same. The return code can be used to tell if there were any
2089 non-matching lines.
2090
2091 14. Added --file-offsets and --line-offsets to pcregrep.
2092
2093 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2094 infinitely looping recursion. The bug was that positive lookaheads were not
2095 being skipped when checking for a possible empty match (negative lookaheads
2096 and both kinds of lookbehind were skipped).
2097
2098 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2099 inclusion of <windows.h> to before rather than after the definition of
2100 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2101
2102 17. Specifying a possessive quantifier with a specific limit for a Unicode
2103 character property caused pcre_compile() to compile bad code, which led at
2104 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2105 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2106 caused the error; without that there was no problem.
2107
2108 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2109
2110 19. Added --enable-pcretest-libreadline.
2111
2112 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2113 RE::GlobalReplace(). As a result, the number of replacements returned was
2114 double what it should be. I removed one of the increments, but Craig sent a
2115 later patch that removed the other one (the right fix) and added unit tests
2116 that check the return values (which was not done before).
2117
2118 21. Several CMake things:
2119
2120 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2121 the names libpcre and libpcreposix, not just pcre and pcreposix.
2122
2123 (2) The above change means that pcretest and pcregrep are now correctly
2124 linked with the newly-built libraries, not previously installed ones.
2125
2126 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2127
2128 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2129 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2130 UTF-8 newline character). The key issue is that the pattern starts .*;
2131 this means that the match must be either at the beginning, or after a
2132 newline. The bug was in the code for advancing after a failed match and
2133 checking that the new position followed a newline. It was not taking
2134 account of UTF-8 characters correctly.
2135
2136 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2137 character classes. PCRE was not treating the sequence [:...:] as a
2138 character class unless the ... were all letters. Perl, however, seems to
2139 allow any characters between [: and :], though of course it rejects as
2140 unknown any "names" that contain non-letters, because all the known class
2141 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2142 for example, whereas PCRE did not - it did not recognize a POSIX character
2143 class. This seemed a bit dangerous, so the code has been changed to be
2144 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2145 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2146 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2147 Perl does, and where it didn't before.
2148
2149 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2150 Windows environments %n is disabled by default.
2151
2152
2153 Version 7.4 21-Sep-07
2154 ---------------------
2155
2156 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2157 means that a class such as [\s] counted as "explicit reference to CR or
2158 LF". That isn't really right - the whole point of the change was to try to
2159 help when there was an actual mention of one of the two characters. So now
2160 the change happens only if \r or \n (or a literal CR or LF) character is
2161 encountered.
2162
2163 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2164 of both had grown to the point where there were only 3 bits left.
2165 Fortunately, there was spare space in the data structure, and so I have
2166 moved the internal flags into a new 16-bit field to free up more option
2167 bits.
2168
2169 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2170 but did not set the internal JCHANGED flag - either of these is enough to
2171 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2172 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2173 start sets both bits.
2174
2175 4. Added options (at build time, compile time, exec time) to change \R from
2176 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2177
2178 5. doc/pcresyntax.html was missing from the distribution.
2179
2180 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2181 compatibility, even though it is no longer used.
2182
2183 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2184 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2185 windows.h file is present (where different names are used). [This was
2186 reversed later after testing - see 16 below.]
2187
2188 8. Changed all #include <config.h> to #include "config.h". There were also
2189 some further <pcre.h> cases that I changed to "pcre.h".
2190
2191 9. When pcregrep was used with the --colour option, it missed the line ending
2192 sequence off the lines that it output.
2193
2194 10. It was pointed out to me that arrays of string pointers cause lots of
2195 relocations when a shared library is dynamically loaded. A technique of
2196 using a single long string with a table of offsets can drastically reduce
2197 these. I have refactored PCRE in four places to do this. The result is
2198 dramatic:
2199
2200 Originally: 290
2201 After changing UCP table: 187
2202 After changing error message table: 43
2203 After changing table of "verbs" 36
2204 After changing table of Posix names 22
2205
2206 Thanks to the folks working on Gregex for glib for this insight.
2207
2208 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2209 unicode-properties was also set.
2210
2211 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2212
2213 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2214 checked only for CRLF.
2215
2216 14. Added casts to pcretest.c to avoid compiler warnings.
2217
2218 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2219
2220 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2221 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2222 entirely. This removes changes made in 7 above.
2223
2224 17. The CMake files have been updated, and there is now more information about
2225 building with CMake in the NON-UNIX-USE document.
2226
2227
2228 Version 7.3 28-Aug-07
2229 ---------------------
2230
2231 1. In the rejigging of the build system that eventually resulted in 7.1, the
2232 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2233 brackets there is not right, since it causes compilers to look for an
2234 installed pcre.h, not the version that is in the source that is being
2235 compiled (which of course may be different). I have changed it back to:
2236
2237 #include "pcre.h"
2238
2239 I have a vague recollection that the change was concerned with compiling in
2240 different directories, but in the new build system, that is taken care of
2241 by the VPATH setting the Makefile.
2242
2243 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2244 when the subject happened to end in the byte 0x85 (e.g. if the last
2245 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2246 characters but of course it shouldn't be taken as a newline when it is part
2247 of another character. The bug was that, for an unlimited repeat of . in
2248 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2249 characters when looking for a newline.
2250
2251 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2252
2253 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2254 in debug output.
2255
2256 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2257 long printing in the pcrecpp unittest when running under MinGW.
2258
2259 6. ESC_K was left out of the EBCDIC table.
2260
2261 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2262 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2263 limit also applies to "virtual nesting" when a pattern is recursive, and in
2264 this case 1000 isn't so big. I have been able to remove this limit at the
2265 expense of backing off one optimization in certain circumstances. Normally,
2266 when pcre_exec() would call its internal match() function recursively and
2267 immediately return the result unconditionally, it uses a "tail recursion"
2268 feature to save stack. However, when a subpattern that can match an empty
2269 string has an unlimited repetition quantifier, it no longer makes this
2270 optimization. That gives it a stack frame in which to save the data for
2271 checking that an empty string has been matched. Previously this was taken
2272 from the 1000-entry workspace that had been reserved. So now there is no
2273 explicit limit, but more stack is used.
2274
2275 8. Applied Daniel's patches to solve problems with the import/export magic
2276 syntax that is required for Windows, and which was going wrong for the
2277 pcreposix and pcrecpp parts of the library. These were overlooked when this
2278 problem was solved for the main library.
2279
2280 9. There were some crude static tests to avoid integer overflow when computing
2281 the size of patterns that contain repeated groups with explicit upper
2282 limits. As the maximum quantifier is 65535, the maximum group length was
2283 set at 30,000 so that the product of these two numbers did not overflow a
2284 32-bit integer. However, it turns out that people want to use groups that
2285 are longer than 30,000 bytes (though not repeat them that many times).
2286 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2287 made it possible to implement the integer overflow checks in a much more
2288 dynamic way, which I have now done. The artificial limitation on group
2289 length has been removed - we now have only the limit on the total length of
2290 the compiled pattern, which depends on the LINK_SIZE setting.
2291
2292 10. Fixed a bug in the documentation for get/copy named substring when
2293 duplicate names are permitted. If none of the named substrings are set, the
2294 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2295 empty string.
2296
2297 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2298 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2299 because the ] is interpreted as the first data character and the
2300 terminating ] is not found. PCRE has been made compatible with Perl in this
2301 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2302 cause memory overwriting.
2303
2304 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2305 string has been matched (to stop an infinite loop). It was not recognizing
2306 a conditional subpattern that could match an empty string if that
2307 subpattern was within another subpattern. For example, it looped when
2308 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2309 condition was not nested. This bug has been fixed.
2310
2311 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2312 past the start of the subject in the presence of bytes with the top bit
2313 set, for example "\x8aBCD".
2314
2315 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2316 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2317
2318 14. Optimized (?!) to (*FAIL).
2319
2320 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2321 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2322 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2323 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2324 does: it's just the validity check that is more restrictive.
2325
2326 16. Inserted checks for integer overflows during escape sequence (backslash)
2327 processing, and also fixed erroneous offset values for syntax errors during
2328 backslash processing.
2329
2330 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2331 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2332
2333 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2334 caused an overrun.
2335
2336 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2337 something other than just ASCII characters) inside a group that had an
2338 unlimited repeat caused a loop at compile time (while checking to see
2339 whether the group could match an empty string).
2340
2341 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2342 [\P{Any}] did so. (Error in the code for printing property names.)
2343
2344 21. An orphan \E inside a character class could cause a crash.
2345
2346 22. A repeated capturing bracket such as (A)? could cause a wild memory
2347 reference during compilation.
2348
2349 23. There are several functions in pcre_compile() that scan along a compiled
2350 expression for various reasons (e.g. to see if it's fixed length for look
2351 behind). There were bugs in these functions when a repeated \p or \P was
2352 present in the pattern. These operators have additional parameters compared
2353 with \d, etc, and these were not being taken into account when moving along
2354 the compiled data. Specifically:
2355
2356 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2357 length.
2358
2359 (b) An item such as \pL+ within a repeated group could cause crashes or
2360 loops.
2361
2362 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2363 "reference to non-existent subpattern" error.
2364
2365 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2366
2367 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2368 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2369
2370 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2371
2372 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2373 character were causing crashes (broken optimization).
2374
2375 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2376 \p or \P) caused a compile-time loop.
2377
2378 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2379 break. For example, the unstudied pattern [\r\n]A does not match the string
2380 "\r\nA" because change 7.0/46 below moves the current point on by two
2381 characters after failing to match at the start. However, the pattern \nA
2382 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2383 the same is true. There doesn't seem any very clean way out of this, but
2384 what I have chosen to do makes the common cases work: PCRE now takes note
2385 of whether there can be an explicit match for \r or \n anywhere in the
2386 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2387 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2388 pattern has explicit CR or LF references.
2389
2390 29. Added (*CR) etc for changing newline setting at start of pattern.
2391
2392
2393 Version 7.2 19-Jun-07
2394 ---------------------
2395
2396 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2397 which is apparently normally available under Windows.
2398
2399 2. Re-jig the pcregrep tests with different newline settings in an attempt
2400 to make them independent of the local environment's newline setting.
2401
2402 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2403
2404 4. Some of the "internals" tests were previously cut out when the link size
2405 was not 2, because the output contained actual offsets. The recent new
2406 "Z" feature of pcretest means that these can be cut out, making the tests
2407 usable with all link sizes.
2408
2409 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2410 stack recursion. This gives a massive performance boost under BSD, but just
2411 a small improvement under Linux. However, it saves one field in the frame
2412 in all cases.
2413
2414 6. Added more features from the forthcoming Perl 5.10:
2415
2416 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2417 recursion call. It refers to the nth most recently opened parentheses.
2418
2419 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2420 to be opened parentheses.
2421
2422 (c) Conditions that refer to capturing parentheses can be specified
2423 relatively, for example, (?(-2)... or (?(+3)...
2424
2425 (d) \K resets the start of the current match so that everything before
2426 is not part of it.
2427
2428 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2429
2430 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2431 reference syntax.
2432
2433 (g) (?| introduces a group in which the numbering of parentheses in each
2434 alternative starts with the same number.
2435
2436 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2437
2438 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2439 PCRE_INFO_JCHANGED.
2440
2441 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2442 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2443 for detecting groups that can match an empty string.
2444
2445 9. A pattern with a very large number of alternatives (more than several
2446 hundred) was running out of internal workspace during the pre-compile
2447 phase, where pcre_compile() figures out how much memory will be needed. A
2448 bit of new cunning has reduced the workspace needed for groups with
2449 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2450 workspace instead of running out of the 4096 that are available.
2451
2452 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2453
2454 11. Applied patch from Google to remove an optimization that didn't quite work.
2455 The report of the bug said:
2456
2457 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2458 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2459 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2460
2461 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2462 it matched the wrong number of bytes.
2463
2464
2465 Version 7.1 24-Apr-07
2466 ---------------------
2467
2468 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2469 that is more "standard", making use of automake and other Autotools. There
2470 is some re-arrangement of the files and adjustment of comments consequent
2471 on this.
2472
2473 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2474 for recursive directory scanning broke on some systems because the files
2475 are not scanned in any specific order and on different systems the order
2476 was different. A call to "sort" has been inserted into RunGrepTest for the
2477 approprate test as a short-term fix. In the longer term there may be an
2478 alternative.
2479
2480 3. I had an email from Eric Raymond about problems translating some of PCRE's
2481 man pages to HTML (despite the fact that I distribute HTML pages, some
2482 people do their own conversions for various reasons). The problems
2483 concerned the use of low-level troff macros .br and .in. I have therefore
2484 removed all such uses from the man pages (some were redundant, some could
2485 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2486 HTML has been updated to handle .nf/.fi and to complain if it encounters
2487 .br or .in.
2488
2489 4. Updated comments in configure.ac that get placed in config.h.in and also
2490 arranged for config.h to be included in the distribution, with the name
2491 config.h.generic, for the benefit of those who have to compile without
2492 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2493
2494 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2495 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2496 makevp.bat for the current PCRE, using the additional files
2497 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2498
2499 6. A Windows user reported a minor discrepancy with test 2, which turned out
2500 to be caused by a trailing space on an input line that had got lost in his
2501 copy. The trailing space was an accident, so I've just removed it.
2502
2503 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2504 that is needed.
2505
2506 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2507 as "const" (a) because they are and (b) because it helps the PHP
2508 maintainers who have recently made a script to detect big data structures
2509 in the php code that should be moved to the .rodata section. I remembered
2510 to update Builducptable as well, so it won't revert if ucptable.h is ever
2511 re-created.
2512
2513 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2514 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2515 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2516 support is not required. This saves 1.5-2K of code, which is important in
2517 some applications.
2518
2519 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2520 so as not to refer to the tables, even though these functions will never be
2521 called when UTF-8 support is disabled. Otherwise there are problems with a
2522 shared library.
2523
2524 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2525
2526 (a) It was defining its arguments as char * instead of void *.
2527
2528 (b) It was assuming that all moves were upwards in memory; this was true
2529 a long time ago when I wrote it, but is no longer the case.
2530
2531 The emulated memove() is provided for those environments that have neither
2532 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2533 is clearly not the case, as these two bugs were recently reported.
2534
2535 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2536 and Detrail to create the HTML documentation, the .txt form of the man
2537 pages, and it removes trailing spaces from listed files. It also creates
2538 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2539 case, it wraps all the #defines with #ifndefs. This script should be run
2540 before "make dist".
2541
2542 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2543 with Unicode property support.
2544
2545 (a) For a maximizing quantifier, if the two different cases of the
2546 character were of different lengths in their UTF-8 codings (there are
2547 some cases like this - I found 11), and the matching function had to
2548 back up over a mixture of the two cases, it incorrectly assumed they
2549 were both the same length.
2550
2551 (b) When PCRE was configured to use the heap rather than the stack for
2552 recursion during matching, it was not correctly preserving the data for
2553 the other case of a UTF-8 character when checking ahead for a match
2554 while processing a minimizing repeat. If the check also involved
2555 matching a wide character, but failed, corruption could cause an
2556 erroneous result when trying to check for a repeat of the original
2557 character.
2558
2559 13. Some tidying changes to the testing mechanism:
2560
2561 (a) The RunTest script now detects the internal link size and whether there
2562 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2563 values substituted by "configure". (The RunGrepTest script already did
2564 this for UTF-8.) The configure.ac script no longer substitutes the
2565 relevant variables.
2566
2567 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2568 with length and offset values. This means that the output is different
2569 for different internal link sizes. Test 2 is skipped for link sizes
2570 other than 2 because of this, bypassing the problem. Unfortunately,
2571 there was also a test in test 3 (the locale tests) that used /B and
2572 failed for link sizes other than 2. Rather than cut the whole test out,
2573 I have added a new /Z option to pcretest that replaces the length and
2574 offset values with spaces. This is now used to make test 3 independent
2575 of link size. (Test 2 will be tidied up later.)
2576
2577 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2578 segmentation fault instead of returning the appropriate error message.
2579
2580 15. In multiline mode when the newline sequence was set to "any", the pattern
2581 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2582 This doesn't seem right; it now treats the CRLF combination as the line
2583 ending, and so does not match in that case. It's only a pattern such as ^$
2584 that would hit this one: something like ^ABC$ would have failed after \r
2585 and then tried again after \r\n.
2586
2587 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2588 in an attempt to make files that differ only in their line terminators
2589 compare equal. This works on Linux.
2590
2591 17. Under certain error circumstances pcregrep might try to free random memory
2592 as it exited. This is now fixed, thanks to valgrind.
2593
2594 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2595 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2596 was because its rules for how to advance for /g after matching an empty
2597 string at the end of a line did not allow for this case. They now check for
2598 it specially.
2599
2600 20. pcretest is supposed to handle patterns and data of any length, by
2601 extending its buffers when necessary. It was getting this wrong when the
2602 buffer for a data line had to be extended.
2603
2604 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2605 CRLF as a newline sequence.
2606
2607 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2608 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2609 I have nevertheless tidied it up.
2610
2611 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2612
2613 24. Added a man page for pcre-config.
2614
2615
2616 Version 7.0 19-Dec-06
2617 ---------------------
2618
2619 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2620 moving to gcc 4.1.1.
2621
2622 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2623 sys/time.h, which is documented as needed for this function. It doesn't
2624 seem to matter on Linux, but it showed up on some releases of OS X.
2625
2626 3. It seems that there are systems where bytes whose values are greater than
2627 127 match isprint() in the "C" locale. The "C" locale should be the
2628 default when a C program starts up. In most systems, only ASCII printing
2629 characters match isprint(). This difference caused the output from pcretest
2630 to vary, making some of the tests fail. I have changed pcretest so that:
2631
2632 (a) When it is outputting text in the compiled version of a pattern, bytes
2633 other than 32-126 are always shown as hex escapes.
2634
2635 (b) When it is outputting text that is a matched part of a subject string,
2636 it does the same, unless a different locale has been set for the match
2637 (using the /L modifier). In this case, it uses isprint() to decide.
2638
2639 4. Fixed a major bug that caused incorrect computation of the amount of memory
2640 required for a compiled pattern when options that changed within the
2641 pattern affected the logic of the preliminary scan that determines the
2642 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2643 that the computed length was too small. The symptoms of this bug were
2644 either the PCRE error "internal error: code overflow" from pcre_compile(),
2645 or a glibc crash with a message such as "pcretest: free(): invalid next
2646 size (fast)". Examples of patterns that provoked this bug (shown in
2647 pcretest format) are:
2648
2649 /(?-x: )/x
2650 /(?x)(?-x: \s*#\s*)/
2651 /((?i)[\x{c0}])/8
2652 /(?i:[\x{c0}])/8
2653
2654 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2655 is now done differently.
2656
2657 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2658 wrapper classes; (b) implement a new function in the C++ scanner that is
2659 more efficient than the old way of doing things because it avoids levels of
2660 recursion in the regex matching; (c) add a paragraph to the documentation
2661 for the FullMatch() function.
2662
2663 6. The escape sequence \n was being treated as whatever was defined as
2664 "newline". Not only was this contrary to the documentation, which states
2665 that \n is character 10 (hex 0A), but it also went horribly wrong when
2666 "newline" was defined as CRLF. This has been fixed.
2667
2668 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2669 was being set to -1 for the "end of line" case (supposedly a value that no
2670 character can have). Though this value is never used (the check for end of
2671 line is "zero bytes in current character"), it caused compiler complaints.
2672 I've changed it to 0xffffffff.
2673
2674 8. In pcre_version.c, the version string was being built by a sequence of
2675 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2676 string (as it is for production releases) called a macro with an empty
2677 argument. The C standard says the result of this is undefined. The gcc
2678 compiler treats it as an empty string (which was what was wanted) but it is
2679 reported that Visual C gives an error. The source has been hacked around to
2680 avoid this problem.
2681
2682 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2683 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2684 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2685 of them did).
2686
2687 10. Originally, pcretest opened its input and output without "b"; then I was
2688 told that "b" was needed in some environments, so it was added for release
2689 5.0 to both the input and output. (It makes no difference on Unix-like
2690 systems.) Later I was told that it is wrong for the input on Windows. I've
2691 now abstracted the modes into two macros, to make it easier to fiddle with
2692 them, and removed "b" from the input mode under Windows.
2693
2694 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2695
2696 12. Added -help and --help to pcretest as an official way of being reminded
2697 of the options.
2698
2699 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2700 and pcrecpp.cc because they annoy compilers at high warning levels.
2701
2702 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2703
2704 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2705 scripts are not C programs :-) and which was not noticed because it works
2706 on Linux.
2707
2708 16. pcretest is supposed to handle any length of pattern and data line (as one
2709 line or as a continued sequence of lines) by extending its input buffer if
2710 necessary. This feature was broken for very long pattern lines, leading to
2711 a string of junk being passed to pcre_compile() if the pattern was longer
2712 than about 50K.
2713
2714 17. I have done a major re-factoring of the way pcre_compile() computes the
2715 amount of memory needed for a compiled pattern. Previously, there was code
2716 that made a preliminary scan of the pattern in order to do this. That was
2717 OK when PCRE was new, but as the facilities have expanded, it has become
2718 harder and harder to keep it in step with the real compile phase, and there
2719 have been a number of bugs (see for example, 4 above). I have now found a
2720 cunning way of running the real compile function in a "fake" mode that
2721 enables it to compute how much memory it would need, while actually only
2722 ever using a few hundred bytes of working memory and without too many
2723 tests of the mode. This should make future maintenance and development
2724 easier. A side effect of this work is that the limit of 200 on the nesting
2725 depth of parentheses has been removed (though this was never a serious
2726 limitation, I suspect). However, there is a downside: pcre_compile() now
2727 runs more slowly than before (30% or more, depending on the pattern). I
2728 hope this isn't a big issue. There is no effect on runtime performance.
2729
2730 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2731 newline (only possible for the last line of a file) and it was a
2732 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2733
2734 19. Added additional timing features to pcretest. (1) The -tm option now times
2735 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2736 separate command line item, by a number that specifies the number of
2737 repeats to use when timing. The default is 50000; this gives better
2738 precision, but takes uncomfortably long for very large patterns.
2739
2740 20. Extended pcre_study() to be more clever in cases where a branch of a
2741 subpattern has no definite first character. For example, (a*|b*)[cd] would
2742 previously give no result from pcre_study(). Now it recognizes that the
2743 first character must be a, b, c, or d.
2744
2745 21. There was an incorrect error "recursive call could loop indefinitely" if
2746 a subpattern (or the entire pattern) that was being tested for matching an
2747 empty string contained only one non-empty item after a nested subpattern.
2748 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2749 incorrectly, because the \d was being skipped in the check.
2750
2751 22. The pcretest program now has a new pattern option /B and a command line
2752 option -b, which is equivalent to adding /B to every pattern. This causes
2753 it to show the compiled bytecode, without the additional information that
2754 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2755 is the same as /B/I).
2756
2757 23. A new optimization is now able automatically to treat some sequences such
2758 as a*b as a*+b. More specifically, if something simple (such as a character
2759 or a simple class like \d) has an unlimited quantifier, and is followed by
2760 something that cannot possibly match the quantified thing, the quantifier
2761 is automatically "possessified".
2762
2763 24. A recursive reference to a subpattern whose number was greater than 39
2764 went wrong under certain circumstances in UTF-8 mode. This bug could also
2765 have affected the operation of pcre_study().
2766
2767 25. Realized that a little bit of performance could be had by replacing
2768 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2769
2770 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2771
2772 27. Possessive quantifiers such as a++ were previously implemented by turning
2773 them into atomic groups such as ($>a+). Now they have their own opcodes,
2774 which improves performance. This includes the automatically created ones
2775 from 23 above.
2776
2777 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2778 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2779 the first matched character to be a colon. This applied both to named and
2780 numbered groups.
2781
2782 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2783
2784 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2785 building PCRE on Windows easier, so I have included it in the distribution.
2786
2787 31. There is now a check in pcretest against a ridiculously large number being
2788 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2789 loop, the loop is abandoned.
2790
2791 32. Forward references to subpatterns in conditions such as (?(2)...) where
2792 subpattern 2 is defined later cause pcre_compile() to search forwards in
2793 the pattern for the relevant set of parentheses. This search went wrong
2794 when there were unescaped parentheses in a character class, parentheses
2795 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2796
2797 33. "Subroutine" calls and backreferences were previously restricted to
2798 referencing subpatterns earlier in the regex. This restriction has now
2799 been removed.
2800
2801 34. Added a number of extra features that are going to be in Perl 5.10. On the
2802 whole, these are just syntactic alternatives for features that PCRE had
2803 previously implemented using the Python syntax or my own invention. The
2804 other formats are all retained for compatibility.
2805
2806 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2807 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2808 also .NET compatible.
2809
2810 (b) A recursion or subroutine call to a named group can now be defined as
2811 (?&name) as well as (?P>name).
2812
2813 (c) A backreference to a named group can now be defined as \k<name> or
2814 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2815 5.10, are also .NET compatible.
2816
2817 (d) A conditional reference to a named group can now use the syntax
2818 (?(<name>) or (?('name') as well as (?(name).
2819
2820 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2821 groups (named and numbered) that are never evaluated inline, but can be
2822 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2823 is always false. There may be only one alternative in such a group.
2824
2825 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2826 as the simple (?(R). The condition is true only if the most recent
2827 recursion is that of the given number or name. It does not search out
2828 through the entire recursion stack.
2829
2830 (g) The escape \gN or \g{N} has been added, where N is a positive or
2831 negative number, specifying an absolute or relative reference.
2832
2833 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2834 some "unreachable code" warnings.
2835
2836 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2837 things, this adds five new scripts.
2838
2839 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2840 There were also incompatibilities regarding the handling of \Q..\E inside
2841 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2842 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2843
2844 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2845 matches an empty string, and forcibly breaks the loop. There were bugs in
2846 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2847 against aaaa the result was just "a" rather than "aaaa", for example. Two
2848 separate and independent bugs (that affected different cases) have been
2849 fixed.
2850
2851 39. Refactored the code to abolish the use of different opcodes for small
2852 capturing bracket numbers. This is a tidy that I avoided doing when I
2853 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2854 The new approach is not only tidier, it makes it possible to reduce the
2855 memory needed to fix the previous bug (38).
2856
2857 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2858 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2859 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2860 mode.
2861
2862 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2863 report.
2864
2865 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2866 copy construction and assignment in the C++ wrapper.
2867
2868 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2869 couple of bugs that could have given wrong results in the "--newline=crlf"
2870 case.
2871
2872 44. Added a number of casts and did some reorganization of signed/unsigned int
2873 variables following suggestions from Dair Grant. Also renamed the variable
2874 "this" as "item" because it is a C++ keyword.
2875
2876 45. Arranged for dftables to add
2877
2878 #include "pcre_internal.h"
2879
2880 to pcre_chartables.c because without it, gcc 4.x may remove the array
2881 definition from the final binary if PCRE is built into a static library and
2882 dead code stripping is activated.
2883
2884 46. For an unanchored pattern, if a match attempt fails at the start of a
2885 newline sequence, and the newline setting is CRLF or ANY, and the next two
2886 characters are CRLF, advance by two characters instead of one.
2887
2888
2889 Version 6.7 04-Jul-06
2890 ---------------------
2891
2892 1. In order to handle tests when input lines are enormously long, pcretest has
2893 been re-factored so that it automatically extends its buffers when
2894 necessary. The code is crude, but this _is_ just a test program. The
2895 default size has been increased from 32K to 50K.
2896
2897 2. The code in pcre_study() was using the value of the re argument before
2898 testing it for NULL. (Of course, in any sensible call of the function, it
2899 won't be NULL.)
2900
2901 3. The memmove() emulation function in pcre_internal.h, which is used on
2902 systems that lack both memmove() and bcopy() - that is, hardly ever -
2903 was missing a "static" storage class specifier.
2904
2905 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2906 containing an extended class (one that cannot be represented by a bitmap
2907 because it contains high-valued characters or Unicode property items, e.g.
2908 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2909 pattern, but PCRE should not loop if you do not (it no longer does).
2910 [Detail: two cases were found: (a) a repeated subpattern containing an
2911 extended class; (b) a recursive reference to a subpattern that followed a
2912 previous extended class. It wasn't skipping over the extended class
2913 correctly when UTF-8 mode was not set.]
2914
2915 5. A negated single-character class was not being recognized as fixed-length
2916 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2917 compile error "lookbehind assertion is not fixed length".
2918
2919 6. The RunPerlTest auxiliary script was showing an unexpected difference
2920 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2921 write a Perl script that can interpret lines of an input file either as
2922 byte characters or as UTF-8, which is what "perltest" was being required to
2923 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2924 can't do is switch easily at run time between having the "use utf8;" pragma
2925 or not. In the end, I fudged it by using the RunPerlTest script to insert
2926 "use utf8;" explicitly for the UTF-8 tests.
2927
2928 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2929 the end of the subject string, contrary to the documentation and to what
2930 Perl does. This was true of both matching functions. Now it matches only at
2931 the start of the subject and immediately after *internal* newlines.
2932
2933 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2934 a pointer to an int instead of a pointer to an unsigned long int. This
2935 caused problems on 64-bit systems.
2936
2937 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2938 instance of the 'standard' template library not being so standard".
2939
2940 10. There was no check on the number of named subpatterns nor the maximum
2941 length of a subpattern name. The product of these values is used to compute
2942 the size of the memory block for a compiled pattern. By supplying a very
2943 long subpattern name and a large number of named subpatterns, the size
2944 computation could be caused to overflow. This is now prevented by limiting
2945 the length of names to 32 characters, and the number of named subpatterns
2946 to 10,000.
2947
2948 11. Subpatterns that are repeated with specific counts have to be replicated in
2949 the compiled pattern. The size of memory for this was computed from the
2950 length of the subpattern and the repeat count. The latter is limited to
2951 65535, but there was no limit on the former, meaning that integer overflow
2952 could in principle occur. The compiled length of a repeated subpattern is
2953 now limited to 30,000 bytes in order to prevent this.
2954
2955 12. Added the optional facility to have named substrings with the same name.
2956
2957 13. Added the ability to use a named substring as a condition, using the
2958 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2959 are numbers (not recommended). Forward references are permitted.
2960
2961 14. Added forward references in named backreferences (if you see what I mean).
2962
2963 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2964 pattern could run off the end of the subject. For example, the pattern
2965 "(?s)(.{1,5})"8 did this with the subject "ab".
2966
2967 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2968 PCRE_CASELESS was set when matching characters that were quantified with ?
2969 or *.
2970
2971 17. A character class other than a single negated character that had a minimum
2972 but no maximum quantifier - for example [ab]{6,} - was not handled
2973 correctly by pce_dfa_exec(). It would match only one character.
2974
2975 18. A valid (though odd) pattern that looked like a POSIX character
2976 class but used an invalid character after [ (for example [[,abc,]]) caused
2977 pcre_compile() to give the error "Failed: internal error: code overflow" or
2978 in some cases to crash with a glibc free() error. This could even happen if
2979 the pattern terminated after [[ but there just happened to be a sequence of
2980 letters, a binary zero, and a closing ] in the memory that followed.
2981
2982 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2983 over the years. Originally (before any Unicode support), just the bottom 8
2984 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2985 output from "man perlunicode" includes this:
2986
2987 The regular expression compiler produces polymorphic opcodes. That
2988 is, the pattern adapts to the data and automatically switches to
2989 the Unicode character scheme when presented with Unicode data--or
2990 instead uses a traditional byte scheme when presented with byte
2991 data.
2992
2993 Sadly, a wide octal escape does not cause a switch, and in a string with
2994 no other multibyte characters, these octal escapes are treated as before.
2995 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2996 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2997 Unicode string.
2998
2999 I have not perpetrated such confusion in PCRE. Up till now, it took just
3000 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3001 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3002 translate to the appropriate multibyte character.
3003
3004 29. Applied some refactoring to reduce the number of warnings from Microsoft
3005 and Borland compilers. This has included removing the fudge introduced
3006 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3007 a warning about an unused variable.
3008
3009 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3010 characters since release 4.0, because Perl (from release 5.004) does not.
3011 [Or at least, is documented not to: some releases seem to be in conflict
3012 with the documentation.] However, when a pattern was studied with
3013 pcre_study() and all its branches started with \s, PCRE still included VT
3014 as a possible starting character. Of course, this did no harm; it just
3015 caused an unnecessary match attempt.
3016
3017 22. Removed a now-redundant internal flag bit that recorded the fact that case
3018 dependency changed within the pattern. This was once needed for "required
3019 byte" processing, but is no longer used. This recovers a now-scarce options
3020 bit. Also moved the least significant internal flag bit to the most-
3021 significant bit of the word, which was not previously used (hangover from
3022 the days when it was an int rather than a uint) to free up another bit for
3023 the future.
3024
3025 23. Added support for CRLF line endings as well as CR and LF. As well as the
3026 default being selectable at build time, it can now be changed at runtime
3027 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3028 specify that it is scanning data with non-default line endings.
3029
3030 24. Changed the definition of CXXLINK to make it agree with the definition of
3031 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3032
3033 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3034 recursions. This makes a big different to stack usage for some patterns.
3035
3036 26. If a subpattern containing a named recursion or subroutine reference such
3037 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3038 the space required for the compiled pattern went wrong and gave too small a
3039 value. Depending on the environment, this could lead to "Failed: internal
3040 error: code overflow at offset 49" or "glibc detected double free or
3041 corruption" errors.
3042
3043 27. Applied patches from Google (a) to support the new newline modes and (b) to
3044 advance over multibyte UTF-8 characters in GlobalReplace.
3045
3046 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3047 difference for some implementation of PCRE in some Windows version.
3048
3049 29. Added some extra testing facilities to pcretest:
3050
3051 \q<number> in a data line sets the "match limit" value
3052 \Q<number> in a data line sets the "match recursion limt" value
3053 -S <number> sets the stack size, where <number> is in megabytes
3054
3055 The -S option isn't available for Windows.
3056
3057
3058 Version 6.6 06-Feb-06
3059 ---------------------
3060
3061 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3062 in pcreposix.h. I have copied the definition from pcre.h.
3063
3064 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3065 because pcre.h is no longer a built file.
3066
3067 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3068 not normally included in the compiled code.
3069
3070
3071 Version 6.5 01-Feb-06
3072 ---------------------
3073
3074 1. When using the partial match feature with pcre_dfa_exec(), it was not
3075 anchoring the second and subsequent partial matches at the new starting
3076 point. This could lead to incorrect results. For example, with the pattern
3077 /1234/, partially matching against "123" and then "a4" gave a match.
3078
3079 2. Changes to pcregrep:
3080
3081 (a) All non-match returns from pcre_exec() were being treated as failures
3082 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3083 error message is output. Some extra information is given for the
3084 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3085 probably the only errors that are likely to be caused by users (by
3086 specifying a regex that has nested indefinite repeats, for instance).
3087 If there are more than 20 of these errors, pcregrep is abandoned.
3088
3089 (b) A binary zero was treated as data while matching, but terminated the
3090 output line if it was written out. This has been fixed: binary zeroes
3091 are now no different to any other data bytes.
3092
3093 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3094 used to set a locale for matching. The --locale=xxxx long option has
3095 been added (no short equivalent) to specify a locale explicitly on the
3096 pcregrep command, overriding the environment variables.
3097
3098 (d) When -B was used with -n, some line numbers in the output were one less
3099 than they should have been.
3100
3101 (e) Added the -o (--only-matching) option.
3102
3103 (f) If -A or -C was used with -c (count only), some lines of context were
3104 accidentally printed for the final match.
3105
3106 (g) Added the -H (--with-filename) option.
3107
3108 (h) The combination of options -rh failed to suppress file names for files
3109 that were found from directory arguments.
3110
3111 (i) Added the -D (--devices) and -d (--directories) options.
3112
3113 (j) Added the -F (--fixed-strings) option.
3114
3115 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3116
3117 (l) Added the --colo(u)r option.
3118
3119 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3120 is not present by default.
3121
3122 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3123 items such as (?R) or (?1), when the recursion could match a number of
3124 alternatives. If it matched one of the alternatives, but subsequently,
3125 outside the recursion, there was a failure, the code tried to back up into
3126 the recursion. However, because of the way PCRE is implemented, this is not
3127 possible, and the result was an incorrect result from the match.
3128
3129 In order to prevent this happening, the specification of recursion has
3130 been changed so that all such subpatterns are automatically treated as
3131 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3132
3133 4. I had overlooked the fact that, in some locales, there are characters for
3134 which isalpha() is true but neither isupper() nor islower() are true. In
3135 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3136 and ordfeminine) are like this. This affected the treatment of \w and \W
3137 when they appeared in character classes, but not when they appeared outside
3138 a character class. The bit map for "word" characters is now created
3139 separately from the results of isalnum() instead of just taking it from the
3140 upper, lower, and digit maps. (Plus the underscore character, of course.)
3141
3142 5. The above bug also affected the handling of POSIX character classes such as
3143 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3144 permanent tables. Instead, the bit maps for such a class were previously
3145 created as the appropriate unions of the upper, lower, and digit bitmaps.
3146 Now they are created by subtraction from the [[:word:]] class, which has
3147 its own bitmap.
3148
3149 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3150 It is created by subtracting the vertical space characters (\x09, \x0a,
3151 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3152 subtraction was done in the overall bitmap for a character class, meaning
3153 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3154 be recognized. This bug has been fixed.
3155
3156 7. Patches from the folks at Google:
3157
3158 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3159 real life, but is still worth protecting against".
3160
3161 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3162 regular expressions".
3163
3164 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3165 have it.
3166
3167 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3168 "configure" and the latter not, in order to fix a problem somebody had
3169 with compiling the Arg class on HP-UX.
3170
3171 (e) Improve the error-handling of the C++ wrapper a little bit.
3172
3173 (f) New tests for checking recursion limiting.
3174
3175 8. The pcre_memmove() function, which is used only if the environment does not
3176 have a standard memmove() function (and is therefore rarely compiled),
3177 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3178 returning a result (though PCRE never actually uses the result).
3179
3180 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3181 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3182 returned instead of calling malloc() with an overflowing number that would
3183 most likely cause subsequent chaos.
3184
3185 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3186
3187 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3188 with this option is matched, the nmatch and pmatch options of regexec() are
3189 ignored.
3190
3191 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3192 provided in case anyone wants to the the POSIX interface with UTF-8
3193 strings.
3194
3195 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3196 C++ linking (needed for some HP-UX environments).
3197
3198 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3199 (unused parameter) and in the pcre_printint() function (omitted "default"
3200 switch label when the default is to do nothing).
3201
3202 15. Added some code to make it possible, when PCRE is compiled as a C++
3203 library, to replace subject pointers for pcre_exec() with a smart pointer
3204 class, thus making it possible to process discontinuous strings.
3205
3206 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3207 much the same function. They were added by different people who were trying
3208 to make PCRE easy to compile on non-Unix systems. It has been suggested
3209 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3210 for compiling on Windows systems. I have therefore replaced it with
3211 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3212 defaults to "extern" for C or "extern C" for C++, which works fine on
3213 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3214 SCOPE with something explicit in config.h. In addition:
3215
3216 (a) pcreposix.h still had just "extern" instead of either of these macros;
3217 I have replaced it with PCRE_DATA_SCOPE.
3218
3219 (b) Functions such as _pcre_xclass(), which are internal to the library,
3220 but external in the C sense, all had PCRE_EXPORT in their definitions.
3221 This is apparently wrong for the Windows case, so I have removed it.
3222 (It makes no difference on Unix-like systems.)
3223
3224 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3225 of recursive calls to match(). This is different to MATCH_LIMIT because
3226 that limits the total number of calls to match(), not all of which increase
3227 the depth of recursion. Limiting the recursion depth limits the amount of
3228 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3229 when PCRE is compiled, and changed at run time. A patch from Google adds
3230 this functionality to the C++ interface.
3231
3232 18. Changes to the handling of Unicode character properties:
3233
3234 (a) Updated the table to Unicode 4.1.0.
3235
3236 (b) Recognize characters that are not in the table as "Cn" (undefined).
3237
3238 (c) I revised the way the table is implemented to a much improved format
3239 which includes recognition of ranges. It now supports the ranges that
3240 are defined in UnicodeData.txt, and it also amalgamates other
3241 characters into ranges. This has reduced the number of entries in the
3242 table from around 16,000 to around 3,000, thus reducing its size
3243 considerably. I realized I did not need to use a tree structure after
3244 all - a binary chop search is just as efficient. Having reduced the
3245 number of entries, I extended their size from 6 bytes to 8 bytes to
3246 allow for more data.
3247
3248 (d) Added support for Unicode script names via properties such as \p{Han}.
3249
3250 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3251 matching that character.
3252
3253 20. When matching a repeated Unicode property with a minimum greater than zero,
3254 (for example \pL{2,}), PCRE could look past the end of the subject if it
3255 reached it while seeking the minimum number of characters. This could
3256 happen only if some of the characters were more than one byte long, because
3257 there is a check for at least the minimum number of bytes.
3258
3259 21. Refactored the implementation of \p and \P so as to be more general, to
3260 allow for more different types of property in future. This has changed the
3261 compiled form incompatibly. Anybody with saved compiled patterns that use
3262 \p or \P will have to recompile them.
3263
3264 22. Added "Any" and "L&" to the supported property types.
3265
3266 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3267 but give a compile time error if the value is greater than 0xff.
3268
3269 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3270 accidentally not being installed or uninstalled.
3271
3272 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3273 made were to insert the current release number. This seemed silly, because
3274 it made things harder for people building PCRE on systems that don't run
3275 "configure". I have turned pcre.h into a distributed file, no longer built
3276 by "configure", with the version identification directly included. There is
3277 no longer a pcre.h.in file.
3278
3279 However, this change necessitated a change to the pcre-config script as
3280 well. It is built from pcre-config.in, and one of the substitutions was the
3281 release number. I have updated configure.ac so that ./configure now finds
3282 the release number by grepping pcre.h.
3283
3284 26. Added the ability to run the tests under valgrind.
3285
3286
3287 Version 6.4 05-Sep-05
3288 ---------------------
3289
3290 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3291 "--" to be printed when multiple files were scanned, even when none of the
3292 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3293 consider it to be a bug, and have restored the previous behaviour.
3294
3295 2. A couple of code tidies to get rid of compiler warnings.
3296
3297 3. The pcretest program used to cheat by referring to symbols in the library
3298 whose names begin with _pcre_. These are internal symbols that are not
3299 really supposed to be visible externally, and in some environments it is
3300 possible to suppress them. The cheating is now confined to including
3301 certain files from the library's source, which is a bit cleaner.
3302
3303 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3304 file's purpose clearer.
3305
3306 5. Reorganized pcre_ucp_findchar().
3307
3308
3309 Version 6.3 15-Aug-05
3310 ---------------------
3311
3312 1. The file libpcre.pc.in did not have general read permission in the tarball.
3313
3314 2. There were some problems when building without C++ support:
3315
3316 (a) If C++ support was not built, "make install" and "make test" still
3317 tried to test it.
3318
3319 (b) There were problems when the value of CXX was explicitly set. Some
3320 changes have been made to try to fix these, and ...
3321
3322 (c) --disable-cpp can now be used to explicitly disable C++ support.
3323
3324 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3325 backslash in a target when C++ was disabled. This confuses some
3326 versions of "make", apparently. Using an intermediate variable solves
3327 this. (Same for CPP_LOBJ.)
3328
3329 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3330 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3331 necessary on certain architectures.
3332
3333 4. Added a setting of -export-symbols-regex to the link command to remove
3334 those symbols that are exported in the C sense, but actually are local
3335 within the library, and not documented. Their names all begin with
3336 "_pcre_". This is not a perfect job, because (a) we have to except some
3337 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3338 available (and never for static libraries). I have made a note to try to
3339 find a way round (a) in the future.
3340
3341
3342 Version 6.2 01-Aug-05
3343 ---------------------
3344
3345 1. There was no test for integer overflow of quantifier values. A construction
3346 such as {1111111111111111} would give undefined results. What is worse, if
3347 a minimum quantifier for a parenthesized subpattern overflowed and became
3348 negative, the calculation of the memory size went wrong. This could have
3349 led to memory overwriting.
3350
3351 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3352
3353 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3354 operating environments where this matters.
3355
3356 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3357 PCRE options from within the C++ wrapper.
3358
3359 5. Named capturing subpatterns were not being correctly counted when a pattern
3360 was compiled. This caused two problems: (a) If there were more than 100
3361 such subpatterns, the calculation of the memory needed for the whole
3362 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3363 back references of the form \12, where the number was greater than 9, were
3364 not recognized as back references, even though there were sufficient
3365 previous subpatterns.
3366
3367 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3368 versions of gcc, e.g. 2.95.4.
3369
3370
3371 Version 6.1 21-Jun-05
3372 ---------------------
3373
3374 1. There was one reference to the variable "posix" in pcretest.c that was not
3375 surrounded by "#if !defined NOPOSIX".
3376
3377 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3378 the cross-check on the old pcre_info() function, for the benefit of the
3379 cut-down version of PCRE that is currently imported into Exim.
3380
3381 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3382 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3383 patterns that start (?i) but I don't think that matters. The use of (?i) is
3384 just an example; this all applies to the other options as well.
3385
3386 4. Since libtool seems to echo the compile commands it is issuing, the output
3387 from "make" can be reduced a bit by putting "@" in front of each libtool
3388 compile command.
3389
3390 5. Patch from the folks at Google for configure.in to be a bit more thorough
3391 in checking for a suitable C++ installation before trying to compile the
3392 C++ stuff. This should fix a reported problem when a compiler was present,
3393 but no suitable headers.
3394
3395 6. The man pages all had just "PCRE" as their title. I have changed them to
3396 be the relevant file name. I have also arranged that these names are
3397 retained in the file doc/pcre.txt, which is a concatenation in text format
3398 of all the man pages except the little individual ones for each function.
3399
3400 7. The NON-UNIX-USE file had not been updated for the different set of source
3401 files that come with release 6. I also added a few comments about the C++
3402 wrapper.
3403
3404
3405 Version 6.0 07-Jun-05
3406 ---------------------
3407
3408 1. Some minor internal re-organization to help with my DFA experiments.
3409
3410 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3411 didn't matter for the library itself when fully configured, but did matter
3412 when compiling without UCP support, or within Exim, where the ucp files are
3413 not imported.
3414
3415 3. Refactoring of the library code to split up the various functions into
3416 different source modules. The addition of the new DFA matching code (see
3417 below) to a single monolithic source would have made it really too
3418 unwieldy, quite apart from causing all the code to be include in a
3419 statically linked application, when only some functions are used. This is
3420 relevant even without the DFA addition now that patterns can be compiled in
3421 one application and matched in another.
3422
3423 The downside of splitting up is that there have to be some external
3424 functions and data tables that are used internally in different modules of
3425 the library but which are not part of the API. These have all had their
3426 names changed to start with "_pcre_" so that they are unlikely to clash
3427 with other external names.
3428
3429 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3430 a different (DFA) algorithm. Although it is slower than the original
3431 function, it does have some advantages for certain types of matching
3432 problem.
3433
3434 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3435 including restarting after a partial match.
3436
3437 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3438 defined when compiling for Windows was sent to me. I have put it into the
3439 code, though I have no means of testing or verifying it.
3440
3441 7. Added the pcre_refcount() auxiliary function.
3442
3443 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3444 match before or at the first newline in the subject string. In pcretest,
3445 the /f option on a pattern can be used to set this.
3446
3447 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3448 would behave wrongly. This has been present in PCRE since release 4.0.
3449
3450 10. A number of changes to the pcregrep command:
3451
3452 (a) Refactored how -x works; insert ^(...)$ instead of setting
3453 PCRE_ANCHORED and checking the length, in preparation for adding
3454 something similar for -w.
3455
3456 (b) Added the -w (match as a word) option.
3457
3458 (c) Refactored the way lines are read and buffered so as to have more
3459 than one at a time available.
3460
3461 (d) Implemented a pcregrep test script.
3462
3463 (e) Added the -M (multiline match) option. This allows patterns to match
3464 over several lines of the subject. The buffering ensures that at least
3465 8K, or the rest of the document (whichever is the shorter) is available
3466 for matching (and similarly the previous 8K for lookbehind assertions).
3467
3468 (f) Changed the --help output so that it now says
3469
3470 -w, --word-regex(p)
3471
3472 instead of two lines, one with "regex" and the other with "regexp"
3473 because that confused at least one person since the short forms are the
3474 same. (This required a bit of code, as the output is generated
3475 automatically from a table. It wasn't just a text change.)
3476
3477 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3478 option but starts with a hyphen. Could be a pattern or a path name
3479 starting with a hyphen, for instance.
3480
3481 (h) "-" can be given as a file name to represent stdin.
3482
3483 (i) When file names are being printed, "(standard input)" is used for
3484 the standard input, for compatibility with GNU grep. Previously
3485 "<stdin>" was used.
3486
3487 (j) The option --label=xxx can be used to supply a name to be used for
3488 stdin when file names are being printed. There is no short form.
3489
3490 (k) Re-factored the options decoding logic because we are going to add
3491 two more options that take data. Such options can now be given in four
3492 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3493
3494 (l) Added the -A, -B, and -C options for requesting that lines of context
3495 around matches be printed.
3496
3497 (m) Added the -L option to print the names of files that do not contain
3498 any matching lines, that is, the complement of -l.
3499
3500 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3501 continue to scan other files.
3502
3503 (o) The -s option was incorrectly implemented. For compatibility with other
3504 greps, it now suppresses the error message for a non-existent or non-
3505 accessible file (but not the return code). There is a new option called
3506 -q that suppresses the output of matching lines, which was what -s was
3507 previously doing.
3508
3509 (p) Added --include and --exclude options to specify files for inclusion
3510 and exclusion when recursing.
3511
3512 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3513 Hopefully, it now does.
3514
3515 12. Missing cast in pcre_study().
3516
3517 13. Added an "uninstall" target to the makefile.
3518
3519 14. Replaced "extern" in the function prototypes in Makefile.in with
3520 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3521 world, but is set differently for Windows.
3522
3523 15. Added a second compiling function called pcre_compile2(). The only
3524 difference is that it has an extra argument, which is a pointer to an
3525 integer error code. When there is a compile-time failure, this is set
3526 non-zero, in addition to the error test pointer being set to point to an
3527 error message. The new argument may be NULL if no error number is required
3528 (but then you may as well call pcre_compile(), which is now just a
3529 wrapper). This facility is provided because some applications need a
3530 numeric error indication, but it has also enabled me to tidy up the way
3531 compile-time errors are handled in the POSIX wrapper.
3532
3533 16. Added VPATH=.libs to the makefile; this should help when building with one
3534 prefix path and installing with another. (Or so I'm told by someone who
3535 knows more about this stuff than I do.)
3536
3537 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3538 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3539 match everything, including newlines. This is not POSIX-compatible, but
3540 somebody wanted the feature. From pcretest it can be activated by using
3541 both the P and the s flags.
3542
3543 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3544
3545 19. libpcre.pc was being incorrectly installed as executable.
3546
3547 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3548 it now also looks for '\r' so that it will work unmodified on Windows.
3549
3550 21. Added Google's contributed C++ wrapper to the distribution.
3551
3552 22. Added some untidy missing memory free() calls in pcretest, to keep
3553 Electric Fence happy when testing.
3554
3555
3556
3557 Version 5.0 13-Sep-04
3558 ---------------------
3559
3560 1. Internal change: literal characters are no longer packed up into items
3561 containing multiple characters in a single byte-string. Each character
3562 is now matched using a separate opcode. However, there may be more than one
3563 byte in the character in UTF-8 mode.
3564
3565 2. The pcre_callout_block structure has two new fields: pattern_position and
3566 next_item_length. These contain the offset in the pattern to the next match
3567 item, and its length, respectively.
3568
3569 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3570 insertion of callouts before each pattern item. Added the /C option to
3571 pcretest to make use of this.
3572
3573 4. On the advice of a Windows user, the lines
3574
3575 #if defined(_WIN32) || defined(WIN32)
3576 _setmode( _fileno( stdout ), 0x8000 );
3577 #endif /* defined(_WIN32) || defined(WIN32) */
3578
3579 have been added to the source of pcretest. This apparently does useful
3580 magic in relation to line terminators.
3581
3582 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3583 for the benefit of those environments where the "b" makes a difference.
3584
3585 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3586 to know about it. I have put a hack into configure.in that adds in code
3587 to set GCC=yes if CC=icc. This seems to end up at a point in the
3588 generated configure script that is early enough to affect the setting of
3589 compiler options, which is what is needed, but I have no means of testing
3590 whether it really works. (The user who reported this had patched the
3591 generated configure script, which of course I cannot do.)
3592
3593 LATER: After change 22 below (new libtool files), the configure script
3594 seems to know about icc (and also ecc). Therefore, I have commented out
3595 this hack in configure.in.
3596
3597 7. Added support for pkg-config (2 patches were sent in).
3598
3599 8. Negated POSIX character classes that used a combination of internal tables
3600 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3601 [[:^ascii]]. Typically, they would match almost any characters. The other
3602 POSIX classes were not broken in this way.
3603
3604 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3605 to find the match, as PCRE was deluded into thinking that the match had to
3606 start at the start point or following a newline. The same bug applied to
3607 patterns with negative forward assertions or any backward assertions
3608 preceding ".*" at the start, unless the pattern required a fixed first
3609 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3610
3611 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3612 starting at the last subject character, bytes beyond the end of the subject
3613 string were read.
3614
3615 11. Renamed the variable "class" as "classbits" to make life easier for C++
3616 users. (Previously there was a macro definition, but it apparently wasn't
3617 enough.)
3618
3619 12. Added the new field "tables" to the extra data so that tables can be passed
3620 in at exec time, or the internal tables can be re-selected. This allows
3621 a compiled regex to be saved and re-used at a later time by a different
3622 program that might have everything at different addresses.
3623
3624 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3625 -R library as well as a -L library.
3626
3627 14. The debugging options of pcretest (-d on the command line or D on a
3628 pattern) showed incorrect output for anything following an extended class
3629 that contained multibyte characters and which was followed by a quantifier.
3630
3631 15. Added optional support for general category Unicode character properties
3632 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3633 support. It adds about 90K to the size of the library. The meanings of the
3634 inbuilt class escapes such as \d and \s have NOT been changed.
3635
3636 16. Updated pcredemo.c to include calls to free() to release the memory for the
3637 compiled pattern.
3638
3639 17. The generated file chartables.c was being created in the source directory
3640 instead of in the building directory. This caused the build to fail if the
3641 source directory was different from the building directory, and was
3642 read-only.
3643
3644 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3645 file. No doubt somebody will tell me if they don't make sense... Also added
3646 Dan Mooney's comments about building on OpenVMS.
3647
3648 19. Added support for partial matching via the PCRE_PARTIAL option for
3649 pcre_exec() and the \P data escape in pcretest.
3650
3651 20. Extended pcretest with 3 new pattern features:
3652
3653 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3654 write the compiled pattern to the file whose name is "rest-of-line".
3655 This is a straight binary dump of the data, with the saved pointer to
3656 the character tables forced to be NULL. The study data, if any, is
3657 written too. After writing, pcretest reads a new pattern.
3658
3659 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3660 compiled pattern from the given file. There must not be any
3661 occurrences of "<" in the file name (pretty unlikely); if there are,
3662 pcretest will instead treat the initial "<" as a pattern delimiter.
3663 After reading in the pattern, pcretest goes on to read data lines as
3664 usual.
3665
3666 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3667 and 16-bit fields in a compiled pattern, to simulate a pattern that
3668 was compiled on a host of opposite endianness.
3669
3670 21. The pcre-exec() function can now cope with patterns that were compiled on
3671 hosts of opposite endianness, with this restriction:
3672
3673 As for any compiled expression that is saved and used later, the tables
3674 pointer field cannot be preserved; the extra_data field in the arguments
3675 to pcre_exec() should be used to pass in a tables address if a value
3676 other than the default internal tables were used at compile time.
3677
3678 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3679 now diagnosed as an error. Previously, most of the time, a negative number
3680 would have been treated as zero, but if in addition "ovector" was passed as
3681 NULL, a crash could occur.
3682
3683 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3684 new versions from the libtool 1.5 distribution (the last one is a copy of
3685 a file called libtool.m4). This seems to have fixed the need to patch
3686 "configure" to support Darwin 1.3 (which I used to do). However, I still
3687 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3688 workstation).
3689
3690 24. Changed the PCRE licence to be the more standard "BSD" licence.
3691
3692
3693 Version 4.5 01-Dec-03
3694 ---------------------
3695
3696 1. There has been some re-arrangement of the code for the match() function so
3697 that it can be compiled in a version that does not call itself recursively.
3698 Instead, it keeps those local variables that need separate instances for
3699 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3700 needs to "recurse". Keeping track of where control must go is done by means
3701 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3702 hide most of the details from the main code, and operates only if
3703 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3704 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3705 operating.
3706
3707 To make it easier for callers to provide specially tailored get/free
3708 functions for this usage, two new functions, pcre_stack_malloc, and
3709 pcre_stack_free, are used. They are always called in strict stacking order,
3710 and the size of block requested is always the same.
3711
3712 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3713 PCRE has been compiled to use the stack or the heap for recursion. The
3714 -C option of pcretest uses this to show which version is compiled.
3715
3716 A new data escape \S, is added to pcretest; it causes the amounts of store
3717 obtained and freed by both kinds of malloc/free at match time to be added
3718 to the output.
3719
3720 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3721 what's available on my current Linux desktop machine.
3722
3723 3. When matching a UTF-8 string, the test for a valid string at the start has
3724 been extended. If start_offset is not zero, PCRE now checks that it points
3725 to a byte that is the start of a UTF-8 character. If not, it returns
3726 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3727 this is necessary because there may be backward assertions in the pattern.
3728 When matching the same subject several times, it may save resources to use
3729 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3730
3731 4. The code for checking the validity of UTF-8 strings has been tightened so
3732 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3733 containing "overlong sequences".
3734
3735 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3736 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3737 should have been "&", but it just so happened that all the cases this let
3738 through by mistake were picked up later in the function.
3739
3740 6. I had used a variable called "isblank" - this is a C99 function, causing
3741 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3742
3743 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3744 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3745 so the version is shown; (c) stop "make test" echoing "./RunTest".
3746
3747 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3748
3749 9. The prototype for memmove() for systems that don't have it was using
3750 size_t, but the inclusion of the header that defines size_t was later. I've
3751 moved the #includes for the C headers earlier to avoid this.
3752
3753 10. Added some adjustments to the code to make it easier to compiler on certain
3754 special systems:
3755
3756 (a) Some "const" qualifiers were missing.
3757 (b) Added the macro EXPORT before all exported functions; by default this
3758 is defined to be empty.
3759 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3760 that it reads its output file name as an argument instead of writing
3761 to the standard output and assuming this can be redirected.
3762
3763 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3764 class containing characters with values greater than 255, PCRE compilation
3765 went into a loop.
3766
3767 12. A recursive reference to a subpattern that was within another subpattern
3768 that had a minimum quantifier of zero caused PCRE to crash. For example,
3769 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3770 recursion. If the recursively-called subpattern itself had a zero repeat,
3771 that was OK.
3772
3773 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3774 buffer into which it was copied (for escape processing) was still set at
3775 1024, so long lines caused crashes.
3776
3777 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3778 "internal error: code overflow...". This applied to any character class
3779 that was followed by a possessive quantifier.
3780
3781 15. Modified the Makefile to add libpcre.la as a prerequisite for
3782 libpcreposix.la because I was told this is needed for a parallel build to
3783 work.
3784
3785 16. If a pattern that contained .* following optional items at the start was
3786 studied, the wrong optimizing data was generated, leading to matching
3787 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3788 matching string must start with a or b or c. The correct conclusion for
3789 this pattern is that a match can start with any character.
3790
3791
3792 Version 4.4 13-Aug-03
3793 ---------------------
3794
3795 1. In UTF-8 mode, a character class containing characters with values between
3796 127 and 255 was not handled correctly if the compiled pattern was studied.
3797 In fixing this, I have also improved the studying algorithm for such
3798 classes (slightly).
3799
3800 2. Three internal functions had redundant arguments passed to them. Removal
3801 might give a very teeny performance improvement.
3802
3803 3. Documentation bug: the value of the capture_top field in a callout is *one
3804 more than* the number of the hightest numbered captured substring.
3805
3806 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3807 in incorrectly linking with a previously installed version. They now link
3808 explicitly with libpcre.la.
3809
3810 5. configure.in no longer needs to recognize Cygwin specially.
3811
3812 6. A problem in pcre.in for Windows platforms is fixed.
3813
3814 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3815 pcretest, it used to include the size of the study block as part of its
3816 output. Unfortunately, the structure contains a field that has a different
3817 size on different hardware architectures. This meant that the tests that
3818 showed this size failed. As the block is currently always of a fixed size,
3819 this information isn't actually particularly useful in pcretest output, so
3820 I have just removed it.
3821
3822 8. Three pre-processor statements accidentally did not start in column 1.
3823 Sadly, there are *still* compilers around that complain, even though
3824 standard C has not required this for well over a decade. Sigh.
3825
3826 9. In pcretest, the code for checking callouts passed small integers in the
3827 callout_data field, which is a void * field. However, some picky compilers
3828 complained about the casts involved for this on 64-bit systems. Now
3829 pcretest passes the address of the small integer instead, which should get
3830 rid of the warnings.
3831
3832 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3833 both compile and run time, and gives an error if an invalid UTF-8 sequence
3834 is found. There is a option for disabling this check in cases where the
3835 string is known to be correct and/or the maximum performance is wanted.
3836
3837 11. In response to a bug report, I changed one line in Makefile.in from
3838
3839 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3840 to
3841 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3842
3843 to look similar to other lines, but I have no way of telling whether this
3844 is the right thing to do, as I do not use Windows. No doubt I'll get told
3845 if it's wrong...
3846
3847
3848 Version 4.3 21-May-03
3849 ---------------------
3850
3851 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3852 Makefile.
3853
3854 2. Some refactoring to improve the quality of the code:
3855
3856 (i) The utf8_table... variables are now declared "const".
3857
3858 (ii) The code for \cx, which used the "case flipping" table to upper case
3859 lower case letters, now just substracts 32. This is ASCII-specific,
3860 but the whole concept of \cx is ASCII-specific, so it seems
3861 reasonable.
3862
3863 (iii) PCRE was using its character types table to recognize decimal and
3864 hexadecimal digits in the pattern. This is silly, because it handles
3865 only 0-9, a-f, and A-F, but the character types table is locale-
3866 specific, which means strange things might happen. A private
3867 table is now used for this - though it costs 256 bytes, a table is
3868 much faster than multiple explicit tests. Of course, the standard
3869 character types table is still used for matching digits in subject
3870 strings against \d.
3871
3872 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3873 ending in _t are). So I've renamed it as ESC_tee.
3874
3875 3. The first argument for regexec() in the POSIX wrapper should have been
3876 defined as "const".
3877
3878 4. Changed pcretest to use malloc() for its buffers so that they can be
3879 Electric Fenced for debugging.
3880
3881 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3882 to read one or more bytes before the start of the subject string. Often this
3883 had no effect on PCRE's behaviour, but in some circumstances it could
3884 provoke a segmentation fault.
3885
3886 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3887 to try to read one or more bytes before the start of the subject string.
3888
3889 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3890 UTF-8 support could misbehave in various ways if the subject string
3891 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3892 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3893 back over UTF-8 characters.)
3894
3895
3896 Version 4.2 14-Apr-03
3897 ---------------------
3898
3899 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3900
3901 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3902 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3903 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3904 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3905 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3906 and BUILD_EXEEXT
3907 Note: automatic setting of the BUILD variables is not yet working
3908 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3909 compile-time but not at link-time
3910 [LINK]: use for linking executables only
3911 make different versions for Windows and non-Windows
3912 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3913 libraries
3914 [LINK_FOR_BUILD]: new variable
3915 [OBJEXT]: use throughout
3916 [EXEEXT]: use throughout
3917 <winshared>: new target
3918 <wininstall>: new target
3919 <dftables.o>: use native compiler
3920 <dftables>: use native linker
3921 <install>: handle Windows platform correctly
3922 <clean>: ditto
3923 <check>: ditto
3924 copy DLL to top builddir before testing
3925
3926 As part of these changes, -no-undefined was removed again. This was reported
3927 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3928 in any case.
3929
3930 3. Some tidies to get rid of compiler warnings:
3931
3932 . In the match_data structure, match_limit was an unsigned long int, whereas
3933 match_call_count was an int. I've made them both unsigned long ints.
3934
3935 . In pcretest the fact that a const uschar * doesn't automatically cast to
3936 a void * provoked a warning.
3937
3938 . Turning on some more compiler warnings threw up some "shadow" variables
3939 and a few more missing casts.
3940
3941 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3942 option, a class that contained a single character with a value between 128
3943 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3944
3945 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3946 option, a class that contained several characters, but with at least one
3947 whose value was between 128 and 255 caused PCRE to crash.
3948
3949
3950 Version 4.1 12-Mar-03
3951 ---------------------
3952
3953 1. Compiling with gcc -pedantic found a couple of places where casts were
3954 needed, and a string in dftables.c that was longer than standard compilers are
3955 required to support.
3956
3957 2. Compiling with Sun's compiler found a few more places where the code could
3958 be tidied up in order to avoid warnings.
3959
3960 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3961 first of these names is deprecated in the latest Autoconf in favour of the name
3962 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3963 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3964 analogy I have changed it to CFLAGS_FOR_BUILD.
3965
3966 4. Added -no-undefined to the linking command in the Makefile, because this is
3967 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3968 linking step for the pcreposix library.
3969
3970 5. PCRE was failing to diagnose the case of two named groups with the same
3971 name.
3972
3973 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3974 literal character that is needed in the subject for a match, and scans along to
3975 ensure that it is present before embarking on the full matching process. This
3976 saves time in cases of nested unlimited repeats that are never going to match.
3977 Problem: the scan can take a lot of time if the subject is very long (e.g.
3978 megabytes), thus penalizing straightforward matches. It is now done only if the
3979 amount of subject to be scanned is less than 1000 bytes.
3980
3981 7. A lesser problem with the same optimization is that it was recording the
3982 first character of an anchored pattern as "needed", thus provoking a search
3983 right along the subject, even when the first match of the pattern was going to
3984 fail. The "needed" character is now not set for anchored patterns, unless it
3985 follows something in the pattern that is of non-fixed length. Thus, it still
3986 fulfils its original purpose of finding quick non-matches in cases of nested
3987 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3988
3989
3990 Version 4.0 17-Feb-03
3991 ---------------------
3992
3993 1. If a comment in an extended regex that started immediately after a meta-item
3994 extended to the end of string, PCRE compiled incorrect data. This could lead to
3995 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3996
3997 2. Moved to autoconf 2.53 and libtool 1.4.2.
3998
3999 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4000 the special perltest8 script is no longer needed - all the tests can be run
4001 from a single perltest script.
4002
4003 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4004 by \s. It has now been removed in PCRE. This means it isn't recognized as
4005 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4006 class [:space:] *does* include VT, thereby creating a mess.
4007
4008 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4009 space and tab.
4010
4011 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4012 its new features into the main test script, reducing the number of scripts.
4013
4014 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4015 were backward compatible, and made the (?i) apply to the whole pattern, as if
4016 /i were given. Now it behaves more logically, and applies the option setting
4017 only to what follows. PCRE has been changed to follow suit. However, if it
4018 finds options settings right at the start of the pattern, it extracts them into
4019 the global options, as before. Thus, they show up in the info data.
4020
4021 8. Added support for the \Q...\E escape sequence. Characters in between are
4022 treated as literals. This is slightly different from Perl in that $ and @ are
4023 also handled as literals inside the quotes. In Perl, they will cause variable
4024 interpolation. Note the following examples:
4025
4026 Pattern PCRE matches Perl matches
4027
4028 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4029 \Qabc\$xyz\E abc\$xyz abc\$xyz
4030 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4031
4032 For compatibility with Perl, \Q...\E sequences are recognized inside character
4033 classes as well as outside them.
4034
4035 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4036 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4037 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4038 signed/unsigned warnings.
4039
4040 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4041 option for pcretest, so I've replaced it by a simple function that does just
4042 that job.
4043
4044 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4045 "pcregrep -".
4046
4047 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4048 Java package. This provides some syntactic sugar for simple cases of what my
4049 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4050 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4051 item, you can use this simplified notation. Note that only makes sense with
4052 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4053 greediness, whatever the setting of the PCRE_UNGREEDY option.
4054
4055 13. A change of greediness default within a pattern was not taking effect at
4056 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4057 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4058 was abstracted outside.
4059
4060 14. PCRE now supports the \G assertion. It is true when the current matching
4061 position is at the start point of the match. This differs from \A when the
4062 starting offset is non-zero. Used with the /g option of pcretest (or similar
4063 code), it works in the same way as it does for Perl's /g option. If all
4064 alternatives of a regex begin with \G, the expression is anchored to the start
4065 match position, and the "anchored" flag is set in the compiled expression.
4066
4067 15. Some bugs concerning the handling of certain option changes within patterns
4068 have been fixed. These applied to options other than (?ims). For example,
4069 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4070 been the other way round. Some of this was related to change 7 above.
4071
4072 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4073 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4074 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4075 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4076
4077 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4078 mode. Unlike ".", it always matches newline, whatever the setting of
4079 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4080 assertions. Perl allows it, but it doesn't (in general) work because it can't
4081 calculate the length of the lookbehind. At least, that's the case for Perl
4082 5.8.0 - I've been told they are going to document that it doesn't work in
4083 future.
4084
4085 18. Added an error diagnosis for escapes that PCRE does not support: these are
4086 \L, \l, \N, \P, \p, \U, \u, and \X.
4087
4088 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4089 reading past the end of the pattern in cases such as /[abcd/.
4090
4091 20. PCRE was getting more memory than necessary for patterns with classes that
4092 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4093
4094 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4095 compiling PCRE for use with Virtual Pascal.
4096
4097 22. Small fix to the Makefile to make it work properly if the build is done
4098 outside the source tree.
4099
4100 23. Added a new extension: a condition to go with recursion. If a conditional
4101 subpattern starts with (?(R) the "true" branch is used if recursion has
4102 happened, whereas the "false" branch is used only at the top level.
4103
4104 24. When there was a very long string of literal characters (over 255 bytes
4105 without UTF support, over 250 bytes with UTF support), the computation of how
4106 much memory was required could be incorrect, leading to segfaults or other
4107 strange effects.
4108
4109 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4110 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4111 there was a subsequent back reference to those brackets. This meant that, for
4112 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4113 possible to check for precisely this case. All we can do is abandon the
4114 optimization if .* occurs inside capturing brackets when there are any back
4115 references whatsoever. (See below for a better fix that came later.)
4116
4117 26. The handling of the optimization for finding the first character of a
4118 non-anchored pattern, and for finding a character that is required later in the
4119 match were failing in some cases. This didn't break the matching; it just
4120 failed to optimize when it could. The way this is done has been re-implemented.
4121
4122 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4123
4124 28. Added a new feature that provides some of the functionality that Perl
4125 provides with (?{...}). The facility is termed a "callout". The way it is done
4126 in PCRE is for the caller to provide an optional function, by setting
4127 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4128 global variable. By default it is unset, which disables all calling out. To get
4129 the function called, the regex must include (?C) at appropriate points. This
4130 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4131 This provides a means of identifying different callout points. When PCRE
4132 reaches such a point in the regex, if pcre_callout has been set, the external
4133 function is called. It is provided with data in a structure called
4134 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4135 matching continues; if it returns a non-zero value, the match at the current
4136 point fails. However, backtracking will occur if possible. [This was changed
4137 later and other features added - see item 49 below.]
4138
4139 29. pcretest is upgraded to test the callout functionality. It provides a
4140 callout function that displays information. By default, it shows the start of
4141 the match and the current position in the text. There are some new data escapes
4142 to vary what happens:
4143
4144 \C+ in addition, show current contents of captured substrings
4145 \C- do not supply a callout function
4146 \C!n return 1 when callout number n is reached
4147 \C!n!m return 1 when callout number n is reached for the mth time
4148
4149 30. If pcregrep was called with the -l option and just a single file name, it
4150 output "<stdin>" if a match was found, instead of the file name.
4151
4152 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4153 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4154 pcre_exec(). This saves a malloc/free per call. The default value of
4155 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4156 when configuring.
4157
4158 32. The default maximum size of a compiled pattern is 64K. There have been a
4159 few cases of people hitting this limit. The code now uses macros to handle the
4160 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4161 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4162 configuring. Tests 2 and 5 work only with 2-byte links because they output
4163 debugging information about compiled patterns.
4164
4165 33. Internal code re-arrangements:
4166
4167 (a) Moved the debugging function for printing out a compiled regex into
4168 its own source file (printint.c) and used #include to pull it into
4169 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4170 separate copies.
4171
4172 (b) Defined the list of op-code names for debugging as a macro in
4173 internal.h so that it is next to the definition of the opcodes.
4174
4175 (c) Defined a table of op-code lengths for simpler skipping along compiled
4176 code. This is again a macro in internal.h so that it is next to the
4177 definition of the opcodes.
4178
4179 34. Added support for recursive calls to individual subpatterns, along the
4180 lines of Robin Houston's patch (but implemented somewhat differently).
4181
4182 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4183 allow it to read and process whole directories in Win32. This code was
4184 contributed by Lionel Fourquaux; it has not been tested by me.
4185
4186 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4187 used to name a group. Names consist of alphanumerics and underscores, and must
4188 be unique. Back references use the syntax (?P=name) and recursive calls use
4189 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4190 numbers. The function pcre_fullinfo() can be used after compilation to extract
4191 a name/number map. There are three relevant calls:
4192
4193 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4194 PCRE_INFO_NAMECOUNT yields the number of entries
4195 PCRE_INFO_NAMETABLE yields a pointer to the map.
4196
4197 The map is a vector of fixed-size entries. The size of each entry depends on
4198 the length of the longest name used. The first two bytes of each entry are the
4199 group number, most significant byte first. There follows the corresponding
4200 name, zero terminated. The names are in alphabetical order.
4201
4202 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4203 case instead of 255. Making it the same both with and without UTF-8 support
4204 means that the same test output works with both.
4205
4206 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4207 calling malloc() with a zero argument.
4208
4209 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4210 optimization. I've improved things by keeping a bitmap of backreferences with
4211 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4212 fact referenced, the optimization can be applied. It is unlikely that a
4213 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4214 the match to follow \n) will appear inside brackets with a number greater than
4215 31, but if it does, any back reference > 31 suppresses the optimization.
4216
4217 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4218 of disabling numbered capturing parentheses. Any opening parenthesis that is
4219 not followed by ? behaves as if it were followed by ?: but named parentheses
4220 can still be used for capturing (and they will acquire numbers in the usual
4221 way).
4222
4223 41. Redesigned the return codes from the match() function into yes/no/error so
4224 that errors can be passed back from deep inside the nested calls. A malloc
4225 failure while inside a recursive subpattern call now causes the
4226 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4227
4228 42. It is now possible to set a limit on the number of times the match()
4229 function is called in a call to pcre_exec(). This facility makes it possible to
4230 limit the amount of recursion and backtracking, though not in a directly
4231 obvious way, because the match() function is used in a number of different
4232 circumstances. The count starts from zero for each position in the subject
4233 string (for non-anchored patterns). The default limit is, for compatibility, a
4234 large number, namely 10 000 000. You can change this in two ways:
4235
4236 (a) When configuring PCRE before making, you can use --with-match-limit=n
4237 to set a default value for the compiled library.
4238
4239 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4240 a different value is set. See 45 below.
4241
4242 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4243
4244 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4245 of things that can be changed at compile time. The first argument specifies
4246 what is wanted and the second points to where the information is to be placed.
4247 The current list of available information is:
4248
4249 PCRE_CONFIG_UTF8
4250
4251 The output is an integer that is set to one if UTF-8 support is available;
4252 otherwise it is set to zero.
4253
4254 PCRE_CONFIG_NEWLINE
4255
4256 The output is an integer that it set to the value of the code that is used for
4257 newline. It is either LF (10) or CR (13).
4258
4259 PCRE_CONFIG_LINK_SIZE
4260
4261 The output is an integer that contains the number of bytes used for internal
4262 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4263
4264 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4265
4266 The output is an integer that contains the threshold above which the POSIX
4267 interface uses malloc() for output vectors. See item 31 above.
4268
4269 PCRE_CONFIG_MATCH_LIMIT
4270
4271 The output is an unsigned integer that contains the default limit of the number
4272 of match() calls in a pcre_exec() execution. See 42 above.
4273
4274 44. pcretest has been upgraded by the addition of the -C option. This causes it
4275 to extract all the available output from the new pcre_config() function, and to
4276 output it. The program then exits immediately.
4277
4278 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4279 order to support additional features. One way would have been to define
4280 pcre_exec2() (for example) with extra arguments, but this would not have been
4281 extensible, and would also have required all calls to the original function to
4282 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4283 is used for passing in "extra" data from pcre_study().
4284
4285 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4286 contains the following fields:
4287
4288 flags a bitmap indicating which of the following fields are set
4289 study_data opaque data from pcre_study()
4290 match_limit a way of specifying a limit on match() calls for a specific
4291 call to pcre_exec()
4292 callout_data data for callouts (see 49 below)
4293
4294 The flag bits are also defined in pcre.h, and are
4295
4296 PCRE_EXTRA_STUDY_DATA
4297 PCRE_EXTRA_MATCH_LIMIT
4298 PCRE_EXTRA_CALLOUT_DATA
4299
4300 The pcre_study() function now returns one of these new pcre_extra blocks, with
4301 the actual study data pointed to by the study_data field, and the
4302 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4303 before. That is, this change is entirely upwards-compatible and requires no
4304 change to existing code.
4305
4306 If you want to pass in additional data to pcre_exec(), you can either place it
4307 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4308 block.
4309
4310 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4311 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4312 times with different match limits, until it finds the minimum value needed for
4313 pcre_exec() to complete. The value is then output. This can be instructive; for
4314 most simple matches the number is quite small, but for pathological cases it
4315 gets very large very quickly.
4316
4317 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4318 returns the size of the data block pointed to by the study_data field in a
4319 pcre_extra block, that is, the value that was passed as the argument to
4320 pcre_malloc() when PCRE was getting memory in which to place the information
4321 created by pcre_study(). The fourth argument should point to a size_t variable.
4322 pcretest has been extended so that this information is shown after a successful
4323 pcre_study() call when information about the compiled regex is being displayed.
4324
4325 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4326 because what follows is always an absolute path. (Later: it turns out that this
4327 is more than cosmetic for MinGW, because it doesn't like empty path
4328 components.)
4329
4330 49. Some changes have been made to the callout feature (see 28 above):
4331
4332 (i) A callout function now has three choices for what it returns:
4333
4334 0 => success, carry on matching
4335 > 0 => failure at this point, but backtrack if possible
4336 < 0 => serious error, return this value from pcre_exec()
4337
4338 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4339 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4340 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4341 use by callout functions. It will never be used by PCRE itself.
4342
4343 (ii) The pcre_extra structure (see 45 above) has a void * field called
4344 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4345 pcre_callout_block structure has a field of the same name. The contents of
4346 the field passed in the pcre_extra structure are passed to the callout
4347 function in the corresponding field in the callout block. This makes it
4348 easier to use the same callout-containing regex from multiple threads. For
4349 testing, the pcretest program has a new data escape
4350
4351 \C*n pass the number n (may be negative) as callout_data
4352
4353 If the callout function in pcretest receives a non-zero value as
4354 callout_data, it returns that value.
4355
4356 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4357 there were some redundant $(CFLAGS) in commands that are now specified as
4358 $(LINK), which already includes $(CFLAGS).
4359
4360 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4361 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4362 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4363 one-byte characters throughout. Note that case-insensitive matching applies
4364 only to characters whose values are less than 256. PCRE doesn't support the
4365 notion of cases for higher-valued characters.
4366
4367 (i) A character class whose characters are all within 0-255 is handled as
4368 a bit map, and the map is inverted for negative classes. Previously, a
4369 character > 255 always failed to match such a class; however it should
4370 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4371
4372 (ii) A negated character class with a single character < 255 is coded as
4373 "not this character" (OP_NOT). This wasn't working properly when the test
4374 character was multibyte, either singly or repeated.
4375
4376 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4377 mode, for example: \x{100}{2,3}.
4378
4379 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4380 singly or repeated) now correctly test multibyte characters. However,
4381 PCRE doesn't recognize any characters with values greater than 255 as
4382 digits, spaces, or word characters. Such characters always match \D, \S,
4383 and \W, and never match \d, \s, or \w.
4384
4385 (v) Classes may now contain characters and character ranges with values
4386 greater than 255. For example: [ab\x{100}-\x{400}].
4387
4388 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4389 PCRE in UTF-8 mode.
4390
4391 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4392 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4393 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4394 value.)
4395
4396 53. The single man page has become too large. I have therefore split it up into
4397 a number of separate man pages. These also give rise to individual HTML pages;
4398 these are now put in a separate directory, and there is an index.html page that
4399 lists them all. Some hyperlinking between the pages has been installed.
4400
4401 54. Added convenience functions for handling named capturing parentheses.
4402
4403 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4404 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4405 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4406 are faulted.
4407
4408 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4409 calling configure. These values are used when compiling the dftables.c program
4410 which is run to generate the source of the default character tables. They
4411 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4412 you will need to set these values.
4413
4414 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4415
4416
4417 Version 3.9 02-Jan-02
4418 ---------------------
4419
4420 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4421
4422 2. If --disable-static was given, the building process failed when trying to
4423 build pcretest and pcregrep. (For some reason it was using libtool to compile
4424 them, which is not right, as they aren't part of the library.)
4425
4426
4427 Version 3.8 18-Dec-01
4428 ---------------------
4429
4430 1. The experimental UTF-8 code was completely screwed up. It was packing the
4431 bytes in the wrong order. How dumb can you get?
4432
4433
4434 Version 3.7 29-Oct-01
4435 ---------------------
4436
4437 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4438 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4439 this didn't happen under Solaris 8, where I normally test things.
4440
4441 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4442 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4443 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4444 this fix an hour or so after the initial 3.7 release.)
4445
4446
4447 Version 3.6 23-Oct-01
4448 ---------------------
4449
4450 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4451 offsets passed as NULL with zero offset count.
4452
4453 2. The config.guess and config.sub files had not been updated when I moved to
4454 the latest autoconf.
4455
4456
4457 Version 3.5 15-Aug-01
4458 ---------------------
4459
4460 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4461 had been forgotten.
4462
4463 2. By using declared but undefined structures, we can avoid using "void"
4464 definitions in pcre.h while keeping the internal definitions of the structures
4465 private.
4466
4467 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4468 user point of view, this means that both static and shared libraries are built
4469 by default, but this can be individually controlled. More of the work of
4470 handling this static/shared cases is now inside libtool instead of PCRE's make
4471 file.
4472
4473 4. The pcretest utility is now installed along with pcregrep because it is
4474 useful for users (to test regexs) and by doing this, it automatically gets
4475 relinked by libtool. The documentation has been turned into a man page, so
4476 there are now .1, .txt, and .html versions in /doc.
4477
4478 5. Upgrades to pcregrep:
4479 (i) Added long-form option names like gnu grep.
4480 (ii) Added --help to list all options with an explanatory phrase.
4481 (iii) Added -r, --recursive to recurse into sub-directories.
4482 (iv) Added -f, --file to read patterns from a file.
4483
4484 6. pcre_exec() was referring to its "code" argument before testing that
4485 argument for NULL (and giving an error if it was NULL).
4486
4487 7. Upgraded Makefile.in to allow for compiling in a different directory from
4488 the source directory.
4489
4490 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4491 options bits, the pointer it was passed was to an int instead of to an unsigned
4492 long int. This mattered only on 64-bit systems.
4493
4494 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4495 generated) instead of pcre.in, which it its source. Also made the same change
4496 in several of the .c files.
4497
4498 10. A new release of gcc defines printf() as a macro, which broke pcretest
4499 because it had an ifdef in the middle of a string argument for printf(). Fixed
4500 by using separate calls to printf().
4501
4502 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4503 script, to force use of CR or LF instead of \n in the source. On non-Unix
4504 systems, the value can be set in config.h.
4505
4506 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4507 absolute limit. Changed the text of the error message to make this clear, and
4508 likewise updated the man page.
4509
4510 13. The limit of 99 on the number of capturing subpatterns has been removed.
4511 The new limit is 65535, which I hope will not be a "real" limit.
4512
4513
4514 Version 3.4 22-Aug-00
4515 ---------------------
4516
4517 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4518
4519 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4520
4521
4522 Version 3.3 01-Aug-00
4523 ---------------------
4524
4525 1. If an octal character was given, but the value was greater than \377, it
4526 was not getting masked to the least significant bits, as documented. This could
4527 lead to crashes in some systems.
4528
4529 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4530 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4531
4532 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4533 These just pass their arguments on to (pcre_free)(), but they are provided
4534 because some uses of PCRE bind it to non-C systems that can call its functions,
4535 but cannot call free() or pcre_free() directly.
4536
4537 4. Add "make test" as a synonym for "make check". Corrected some comments in
4538 the Makefile.
4539
4540 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4541 Makefile.
4542
4543 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4544 command called pgrep for grepping around the active processes.
4545
4546 7. Added the beginnings of support for UTF-8 character strings.
4547
4548 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4549 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4550 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4551 out for the ar command.)
4552
4553
4554 Version 3.2 12-May-00
4555 ---------------------
4556
4557 This is purely a bug fixing release.
4558
4559 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4560 of ZA. This was just one example of several cases that could provoke this bug,
4561 which was introduced by change 9 of version 2.00. The code for breaking
4562 infinite loops after an iteration that matches an empty string was't working
4563 correctly.
4564
4565 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4566 when matched against abbab (for example). After matching an empty string, it
4567 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4568 caused it to match further down the string than it should.
4569
4570 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4571 was there because it doesn't seem to be needed, and it causes trouble on some
4572 systems, as it is not a Standard C header. It has been removed.
4573
4574 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4575 were reported on the Macintosh. The changes were from
4576
4577 while ((c = *(++ptr)) != 0 && c != '\n');
4578 to
4579 while ((c = *(++ptr)) != 0 && c != '\n') ;
4580
4581 Totally extraordinary, but if that's what it takes...
4582
4583 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4584 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4585 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4586 assumes the way PCRE uses memmove() (always moving upwards).
4587
4588 6. PCRE is being used in one environment where strchr() is not available. There
4589 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4590 faster code anyway.
4591
4592
4593 Version 3.1 09-Feb-00
4594 ---------------------
4595
4596 The only change in this release is the fixing of some bugs in Makefile.in for
4597 the "install" target:
4598
4599 (1) It was failing to install pcreposix.h.
4600
4601 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4602
4603
4604 Version 3.0 01-Feb-00
4605 ---------------------
4606
4607 1. Add support for the /+ modifier to perltest (to output $` like it does in
4608 pcretest).
4609
4610 2. Add support for the /g modifier to perltest.
4611
4612 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4613 matches null strings.
4614
4615 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4616 pattern. Perl treats empty patterns specially - it reuses the most recent
4617 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4618 effect.
4619
4620 5. The POSIX interface was broken in that it was just handing over the POSIX
4621 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4622 required a bigger vector, with some working space on the end. This means that
4623 the POSIX wrapper now has to get and free some memory, and copy the results.
4624
4625 6. Added some simple autoconf support, placing the test data and the
4626 documentation in separate directories, re-organizing some of the
4627 information files, and making it build pcre-config (a GNU standard). Also added
4628 libtool support for building PCRE as a shared library, which is now the
4629 default.
4630
4631 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4632 09 are not valid octal constants. Single digits will be used for minor values
4633 less than 10.
4634
4635 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4636 existing programs that set these in the POSIX interface can use PCRE without
4637 modification.
4638
4639 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4640 return all that pcre_info() returns, plus additional data. The pcre_info()
4641 function is retained for compatibility, but is considered to be obsolete.
4642
4643 10. Added experimental recursion feature (?R) to handle one common case that
4644 Perl 5.6 will be able to do with (?p{...}).
4645
4646 11. Added support for POSIX character classes like [:alpha:], which Perl is
4647 adopting.
4648
4649
4650 Version 2.08 31-Aug-99
4651 ----------------------
4652
4653 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4654 trying to match at the startoffset position, but instead was moving forward to
4655 the next newline as if a previous match had failed.
4656
4657 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4658 and could get into a loop if a null string was matched other than at the start
4659 of the subject.
4660
4661 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4662 be distinguished at compile time, and for completeness also added PCRE_DATE.
4663
4664 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4665 in GnuWin32 environments.
4666
4667
4668 Version 2.07 29-Jul-99
4669 ----------------------
4670
4671 1. The documentation is now supplied in plain text form and HTML as well as in
4672 the form of man page sources.
4673
4674 2. C++ compilers don't like assigning (void *) values to other pointer types.
4675 In particular this affects malloc(). Although there is no problem in Standard
4676 C, I've put in casts to keep C++ compilers happy.
4677
4678 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4679 should be (const char *).
4680
4681 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4682 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4683 However, I haven't made this a standard facility. The documentation doesn't
4684 mention it, and the Makefile doesn't support it.
4685
4686 5. The Makefile now contains an "install" target, with editable destinations at
4687 the top of the file. The pcretest program is not installed.
4688
4689 6. pgrep -V now gives the PCRE version number and date.
4690
4691 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4692 causing the entire string to be ignored, instead of just the last character.
4693
4694 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4695 non-matching string, it can take a very, very long time, even for strings of
4696 quite modest length, because of the nested recursion. PCRE now does better in
4697 some of these cases. It does this by remembering the last required literal
4698 character in the pattern, and pre-searching the subject to ensure it is present
4699 before running the real match. In other words, it applies a heuristic to detect
4700 some types of certain failure quickly, and in the above example, if presented
4701 with a string that has no trailing " it gives "no match" very quickly.
4702
4703 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4704 other alternatives are tried instead.
4705
4706
4707 Version 2.06 09-Jun-99
4708 ----------------------
4709
4710 1. Change pcretest's output for amount of store used to show just the code
4711 space, because the remainder (the data block) varies in size between 32-bit and
4712 64-bit systems.
4713
4714 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4715 start matching at. This allows lookbehinds to work when searching for multiple
4716 occurrences in a string.
4717
4718 3. Added additional options to pcretest for testing multiple occurrences:
4719
4720 /+ outputs the rest of the string that follows a match
4721 /g loops for multiple occurrences, using the new startoffset argument
4722 /G loops for multiple occurrences by passing an incremented pointer
4723
4724 4. PCRE wasn't doing the "first character" optimization for patterns starting
4725 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4726 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4727 the letter 'x'. On long subject strings, this gives a significant speed-up.
4728
4729
4730 Version 2.05 21-Apr-99
4731 ----------------------
4732
4733 1. Changed the type of magic_number from int to long int so that it works
4734 properly on 16-bit systems.
4735
4736 2. Fixed a bug which caused patterns starting with .* not to work correctly
4737 when the subject string contained newline characters. PCRE was assuming
4738 anchoring for such patterns in all cases, which is not correct because .* will
4739 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4740 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4741 must be retried after every newline in the subject.
4742
4743
4744 Version 2.04 18-Feb-99
4745 ----------------------
4746
4747 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4748 computation of the store needed to hold the pattern was incorrect (too large).
4749 If such patterns were nested a few deep, this could multiply and become a real
4750 problem.
4751
4752 2. Added /M option to pcretest to show the memory requirement of a specific
4753 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4754
4755 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4756 compiled in such a way that the backtracking after subsequent failure was
4757 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4758 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4759
4760
4761 Version 2.03 02-Feb-99
4762 ----------------------
4763
4764 1. Fixed typo and small mistake in man page.
4765
4766 2. Added 4th condition (GPL supersedes if conflict) and created separate
4767 LICENCE file containing the conditions.
4768
4769 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4770 Perl, that is the internal \ allows the delimiter to be included in the
4771 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4772 the final delimiter, add \ to the end of the pattern (to test the error).
4773
4774 4. Added the convenience functions for extracting substrings after a successful
4775 match. Updated pcretest to make it able to test these functions.
4776
4777
4778 Version 2.02 14-Jan-99
4779 ----------------------
4780
4781 1. Initialized the working variables associated with each extraction so that
4782 their saving and restoring doesn't refer to uninitialized store.
4783
4784 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4785 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4786 fix the problem.
4787
4788 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4789 calls, and wasn't printing the correct value for compiling calls. Increased the
4790 default value of LOOPREPEAT, and the number of significant figures in the
4791 times.
4792
4793 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4794
4795 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4796 a building problem on Windows NT with a FAT file system.
4797
4798
4799 Version 2.01 21-Oct-98
4800 ----------------------
4801
4802 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4803 to character tables built by pcre_maketables() in the current locale. If NULL
4804 is passed, the default tables are used.
4805
4806
4807 Version 2.00 24-Sep-98
4808 ----------------------
4809
4810 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4811 it any more.
4812
4813 2. Allow quantification of (?>) groups, and make it work correctly.
4814
4815 3. The first character computation wasn't working for (?>) groups.
4816
4817 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4818 end of the subject) and add 5.005's \z, which really does match only at the
4819 very end of the subject.
4820
4821 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4822
4823 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4824 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4825 localized options. All options to pcre_study() were also removed.
4826
4827 7. Add other new features from 5.005:
4828
4829 $(?<= positive lookbehind
4830 $(?<! negative lookbehind
4831 (?imsx-imsx) added the unsetting capability
4832 such a setting is global if at outer level; local otherwise
4833 (?imsx-imsx:) non-capturing groups with option setting
4834 (?(cond)re|re) conditional pattern matching
4835
4836 A backreference to itself in a repeated group matches the previous
4837 captured string.
4838
4839 8. General tidying up of studying (both automatic and via "study")
4840 consequential on the addition of new assertions.
4841
4842 9. As in 5.005, unlimited repeated groups that could match an empty substring
4843 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4844 runtime if any iteration does actually match an empty substring.
4845
4846 10. Include the RunTest script in the distribution.
4847
4848 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4849 discrepancies, some of which were old and were also with respect to 5.004. They
4850 have now been fixed.
4851
4852
4853 Version 1.09 28-Apr-98
4854 ----------------------
4855
4856 1. A negated single character class followed by a quantifier with a minimum
4857 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4858 program crashes, or just wrong answers. This did not apply to negated classes
4859 containing more than one character, or to minima other than one.
4860
4861
4862 Version 1.08 27-Mar-98
4863 ----------------------
4864
4865 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4866
4867 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4868 latter must appear before anything that relies on it in the pattern.
4869
4870
4871 Version 1.07 16-Feb-98
4872 ----------------------
4873
4874 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4875 repeat of a potentially empty string).
4876
4877
4878 Version 1.06 23-Jan-98
4879 ----------------------
4880
4881 1. Added Markus Oberhumer's little patches for C++.
4882
4883 2. Literal strings longer than 255 characters were broken.
4884
4885
4886 Version 1.05 23-Dec-97
4887 ----------------------
4888
4889 1. Negated character classes containing more than one character were failing if
4890 PCRE_CASELESS was set at run time.
4891
4892
4893 Version 1.04 19-Dec-97
4894 ----------------------
4895
4896 1. Corrected the man page, where some "const" qualifiers had been omitted.
4897
4898 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4899 input syntax.
4900
4901 3. Fixed memory leak which occurred when a regex with back references was
4902 matched with an offsets vector that wasn't big enough. The temporary memory
4903 that is used in this case wasn't being freed if the match failed.
4904
4905 4. Tidied pcretest to ensure it frees memory that it gets.
4906
4907 5. Temporary memory was being obtained in the case where the passed offsets
4908 vector was exactly big enough.
4909
4910 6. Corrected definition of offsetof() from change 5 below.
4911
4912 7. I had screwed up change 6 below and broken the rules for the use of
4913 setjmp(). Now fixed.
4914
4915
4916 Version 1.03 18-Dec-97
4917 ----------------------
4918
4919 1. A erroneous regex with a missing opening parenthesis was correctly
4920 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4921 on some systems.
4922
4923 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4924 it was reported that one broken compiler failed on the former because "code" is
4925 also an independent variable.
4926
4927 3. The erroneous regex a[]b caused an array overrun reference.
4928
4929 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4930 fail on data ending with that character. (It was going on too far, and checking
4931 the next character, typically a binary zero.) This was specific to the
4932 optimized code for single-character negative classes.
4933
4934 5. Added a contributed patch from the TIN world which does the following:
4935
4936 + Add an undef for memmove, in case the the system defines a macro for it.
4937
4938 + Add a definition of offsetof(), in case there isn't one. (I don't know
4939 the reason behind this - offsetof() is part of the ANSI standard - but
4940 it does no harm).
4941
4942 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4943 most of the places where whitespace preceded '#'. I have given up and
4944 allowed the remaining 2 cases to be at the margin.
4945
4946 + Rename some variables in pcre to eliminate shadowing. This seems very
4947 pedantic, but does no harm, of course.
4948
4949 6. Moved the call to setjmp() into its own function, to get rid of warnings
4950 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4951
4952 7. Constructs such as \d{8,} were compiling into the equivalent of
4953 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4954 outcome, but in this particular case used more store than had been allocated,
4955 which caused the bug to be discovered because it threw up an internal error.
4956
4957 8. The debugging code in both pcre and pcretest for outputting the compiled
4958 form of a regex was going wrong in the case of back references followed by
4959 curly-bracketed repeats.
4960
4961
4962 Version 1.02 12-Dec-97
4963 ----------------------
4964
4965 1. Typos in pcre.3 and comments in the source fixed.
4966
4967 2. Applied a contributed patch to get rid of places where it used to remove
4968 'const' from variables, and fixed some signed/unsigned and uninitialized
4969 variable warnings.
4970
4971 3. Added the "runtest" target to Makefile.
4972
4973 4. Set default compiler flag to -O2 rather than just -O.
4974
4975
4976 Version 1.01 19-Nov-97
4977 ----------------------
4978
4979 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4980 like /([ab]*)*/, that is, for classes with more than one character in them.
4981
4982 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4983 as /((?>a*))*/ (a PCRE_EXTRA facility).
4984
4985
4986 Version 1.00 18-Nov-97
4987 ----------------------
4988
4989 1. Added compile-time macros to support systems such as SunOS4 which don't have
4990 memmove() or strerror() but have other things that can be used instead.
4991
4992 2. Arranged that "make clean" removes the executables.
4993
4994
4995 Version 0.99 27-Oct-97
4996 ----------------------
4997
4998 1. Fixed bug in code for optimizing classes with only one character. It was
4999 initializing a 32-byte map regardless, which could cause it to run off the end
5000 of the memory it had got.
5001
5002 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5003
5004
5005 Version 0.98 22-Oct-97
5006 ----------------------
5007
5008 1. Fixed bug in code for handling temporary memory usage when there are more
5009 back references than supplied space in the ovector. This could cause segfaults.
5010
5011
5012 Version 0.97 21-Oct-97
5013 ----------------------
5014
5015 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5016
5017 2. Optimized negated single characters not to use a bit map.
5018
5019 3. Brought error texts together as macro definitions; clarified some of them;
5020 fixed one that was wrong - it said "range out of order" when it meant "invalid
5021 escape sequence".
5022
5023 4. Changed some char * arguments to const char *.
5024
5025 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5026
5027 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5028 pcretest.
5029
5030
5031 Version 0.96 16-Oct-97
5032 ----------------------
5033
5034 1. Added a simple "pgrep" utility to the distribution.
5035
5036 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5037 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5038 where "ddd" means "one or more decimal digits".
5039
5040 3. Fixed serious bug. If a pattern had a back reference, but the call to
5041 pcre_exec() didn't supply a large enough ovector to record the related
5042 identifying subpattern, the match always failed. PCRE now remembers the number
5043 of the largest back reference, and gets some temporary memory in which to save
5044 the offsets during matching if necessary, in order to ensure that
5045 backreferences always work.
5046
5047 4. Increased the compatibility with Perl in a number of ways:
5048
5049 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5050 to request this handling. The option can be set at compile or exec time.
5051
5052 (b) $ matches before a terminating newline by default; an option
5053 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5054 mode). The option can be set at compile or exec time.
5055
5056 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5057 the same as Perl's. If the decimal number it represents is less than 10
5058 or there aren't that many previous left capturing parentheses, an octal
5059 escape is read. Inside a character class, it's always an octal escape,
5060 even if it is a single digit.
5061
5062 (d) An escaped but undefined alphabetic character is taken as a literal,
5063 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5064 escapes.
5065
5066 (e) {0} is now permitted. (The previous item is removed from the compiled
5067 pattern).
5068
5069 5. Changed all the names of code files so that the basic parts are no longer
5070 than 10 characters, and abolished the teeny "globals.c" file.
5071
5072 6. Changed the handling of character classes; they are now done with a 32-byte
5073 bit map always.
5074
5075 7. Added the -d and /D options to pcretest to make it possible to look at the
5076 internals of compilation without having to recompile pcre.
5077
5078
5079 Version 0.95 23-Sep-97
5080 ----------------------
5081
5082 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5083 \x20 at the start of a run of normal characters. These were being treated as
5084 real characters, instead of the source characters being re-checked.
5085
5086
5087 Version 0.94 18-Sep-97
5088 ----------------------
5089
5090 1. The functions are now thread-safe, with the caveat that the global variables
5091 containing pointers to malloc() and free() or alternative functions are the
5092 same for all threads.
5093
5094 2. Get pcre_study() to generate a bitmap of initial characters for non-
5095 anchored patterns when this is possible, and use it if passed to pcre_exec().
5096
5097
5098 Version 0.93 15-Sep-97
5099 ----------------------
5100
5101 1. /(b)|(:+)/ was computing an incorrect first character.
5102
5103 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5104 but not actually doing anything yet.
5105
5106 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5107 as Perl does (e.g. [-az] or [az-]).
5108
5109 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5110 all possible positions.
5111
5112 5. Split up into different modules to avoid including unneeded functions in a
5113 compiled binary. However, compile and exec are still in one module. The "study"
5114 function is split off.
5115
5116 6. The character tables are now in a separate module whose source is generated
5117 by an auxiliary program - but can then be edited by hand if required. There are
5118 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5119 toupper() in the code.
5120
5121 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5122 make them global. Abolish the function for setting them, as the caller can now
5123 set them directly.
5124
5125
5126 Version 0.92 11-Sep-97
5127 ----------------------
5128
5129 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5130 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5131
5132 2. Caseless matching was not working in character classes if the characters in
5133 the pattern were in upper case.
5134
5135 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5136
5137 4. Make PCRE_ANCHORED public and accept as a compile option.
5138
5139 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5140 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5141 pass them.
5142
5143 6. Give an error if bad option bits passed at compile or run time.
5144
5145 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5146 pcretest to cause it to pass that flag.
5147
5148 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5149 options, and the first character, if set.
5150
5151 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5152
5153
5154 Version 0.91 10-Sep-97
5155 ----------------------
5156
5157 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5158 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5159
5160 2. PCRE was looping on encountering an indefinitely repeated back reference to
5161 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5162 Perl does - treats the match as successful.
5163
5164 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5