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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1311 - (show annotations)
Mon Apr 22 17:35:23 2013 UTC (6 years, 7 months ago) by ph10
File size: 246075 byte(s)
Use tail recursion in maximizing character and character type repetitions, to 
reduce stack usage.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.33 xx-xxxx-201x
5 -------------------------
6
7 1. Added 'U' to some constants that are compared to unsigned integers, to
8 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
9 variables that are added to signed variables, to ensure the result is
10 signed and can be negated.
11
12 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
13 CMake config files.
14
15 3. Revise the creation of config.h.generic so that all boolean macros are
16 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
17 overriding via -D on the command line possible.
18
19 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
20 to unsigned int is reported to make a quite noticeable speed difference in
21 a specific Windows environment. Testing on Linux did also appear to show
22 some benefit (and it is clearly not harmful). Also fixed the definition of
23 Xop which should be unsigned.
24
25 5. Related to (4), changing the definition of the intermediate variable cc
26 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
27 improvements.
28
29 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
30 unnecessary spaces.
31
32 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
33 and later.
34
35 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
36
37 9. Optimizing fast_forward_start_bits in JIT.
38
39 10. Adding experimental support for callouts in JIT, and fixing some
40 issues revealed during this work. Namely:
41
42 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
43
44 (b) Minimum length was not checked before the matching is started.
45
46 11. The value of capture_last that is passed to callouts was incorrect in some
47 cases when there was a capture on one path that was subsequently abandoned
48 after a backtrack. Also, the capture_last value is now reset after a
49 recursion, since all captures are also reset in this case.
50
51 12. The interpreter no longer returns the "too many substrings" error in the
52 case when an overflowing capture is in a branch that is subsequently
53 abandoned after a backtrack.
54
55 13. In the pathological case when an offset vector of size 2 is used, pcretest
56 now prints out the matched string after a yield of 0 or 1.
57
58 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
59 Only supported by the JIT compiler at the moment.
60
61 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
62
63 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
64 offset of the starting point of the matching process, provided the offsets
65 vector is large enough.
66
67 17. The \A escape now records a lookbehind value of 1, though its execution
68 does not actually inspect the previous character. This is to ensure that,
69 in partial multi-segment matching, at least one character from the old
70 segment is retained when a new segment is processed. Otherwise, if there
71 are no lookbehinds in the pattern, \A might match incorrectly at the start
72 of a new segment.
73
74 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
75
76 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
77 gives some modest performance improvement in 8-bit mode.
78
79 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
80 be expressed in certain programming languages using Universal Character
81 Names.
82
83 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
84 which points out that "non characters" are not "characters that may not
85 appear in Unicode strings" but rather "characters that are reserved for
86 internal use and have only local meaning".
87
88 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
89 there was a conditional group that depended on an assertion, if the
90 assertion was false, the callout that immediately followed the alternation
91 in the condition was skipped when pcre_exec() was used for matching.
92
93 23. Allow an explicit callout to be inserted before an assertion that is the
94 condition for a conditional group, for compatibility with automatic
95 callouts, which always insert a callout at this point.
96
97 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
98 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
99
100 25. (*PRUNE) is now supported by the JIT compiler.
101
102 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
103
104 27. Fix the case where there are two or more SKIPs with arguments that may be
105 ignored.
106
107 28. (*SKIP) is now supported by the JIT compiler.
108
109 29. (*THEN) is now supported by the JIT compiler.
110
111 30. Update RunTest with additional test selector options.
112
113 31. The way PCRE handles backtracking verbs has been changed in two ways.
114
115 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
116 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
117 backtracking. In some cases this makes it more Perl-compatible, but Perl's
118 rather obscure rules do not always do the same thing.
119
120 (2) Previously, backtracking verbs were confined within assertions. This is
121 no longer the case for positive assertions, except for (*ACCEPT). Again,
122 this sometimes improves Perl compatibility, and sometimes does not.
123
124 32. A number of tests that were in test 2 because Perl did things differently
125 have been moved to test 1, because either Perl or PCRE has changed, and
126 these tests are now compatible.
127
128 32. Control verbs are handled in the same way in JIT and interpreter.
129
130 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
131 contained a forward subroutine reference caused a compile error.
132
133 34. Auto-detect and optimize limited repetitions in JIT.
134
135 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
136 blocking (*UTF) etc.
137
138 36. In the interpreter, maximizing pattern repetitions for characters and
139 character types now use tail recursion, which reduces stack usage.
140
141
142 Version 8.32 30-November-2012
143 -----------------------------
144
145 1. Improved JIT compiler optimizations for first character search and single
146 character iterators.
147
148 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
149 Patch by Daniel Richard G.
150
151 3. Single character iterator optimizations in the JIT compiler.
152
153 4. Improved JIT compiler optimizations for character ranges.
154
155 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
156 Reported by Giuseppe D'Angelo.
157
158 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
159 of a line, was being set incorrectly in cases where .* appeared inside
160 atomic brackets at the start of a pattern, or where there was a subsequent
161 *PRUNE or *SKIP.
162
163 7. Improved instruction cache flush for POWER/PowerPC.
164 Patch by Daniel Richard G.
165
166 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
167 grep:
168
169 (a) There is now no limit to the number of patterns to be matched.
170
171 (b) An error is given if a pattern is too long.
172
173 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
174 are now supported.
175
176 (d) --exclude-from and --include-from (multiple use) have been added.
177
178 (e) Exclusions and inclusions now apply to all files and directories, not
179 just to those obtained from scanning a directory recursively.
180
181 (f) Multiple uses of -f and --file-list are now supported.
182
183 (g) In a Windows environment, the default for -d has been changed from
184 "read" (the GNU grep default) to "skip", because otherwise the presence
185 of a directory in the file list provokes an error.
186
187 (h) The documentation has been revised and clarified in places.
188
189 9. Improve the matching speed of capturing brackets.
190
191 10. Changed the meaning of \X so that it now matches a Unicode extended
192 grapheme cluster.
193
194 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
195 out POSIX threads when JIT support is configured.
196
197 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
198
199 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
200 structure could go wrong in environments where size_t is not the same size
201 as int.
202
203 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
204 set.
205
206 15. The EBCDIC support had decayed; later updates to the code had included
207 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
208 general tidy up of EBCDIC-related issues, and the documentation was also
209 not quite right. There is now a test that can be run on ASCII systems to
210 check some of the EBCDIC-related things (but is it not a full test).
211
212 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
213 in a small tidy to the code.
214
215 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
216
217 18. If the --only-matching (-o) option in pcregrep is specified multiple
218 times, each one causes appropriate output. For example, -o1 -o2 outputs the
219 substrings matched by the 1st and 2nd capturing parentheses. A separating
220 string can be specified by --om-separator (default empty).
221
222 19. Improving the first n character searches.
223
224 20. Turn case lists for horizontal and vertical white space into macros so that
225 they are defined only once.
226
227 21. This set of changes together give more compatible Unicode case-folding
228 behaviour for characters that have more than one other case when UCP
229 support is available.
230
231 (a) The Unicode property table now has offsets into a new table of sets of
232 three or more characters that are case-equivalent. The MultiStage2.py
233 script that generates these tables (the pcre_ucd.c file) now scans
234 CaseFolding.txt instead of UnicodeData.txt for character case
235 information.
236
237 (b) The code for adding characters or ranges of characters to a character
238 class has been abstracted into a generalized function that also handles
239 case-independence. In UTF-mode with UCP support, this uses the new data
240 to handle characters with more than one other case.
241
242 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
243 whose other case is greater than 256 are now correctly matched
244 caselessly. Previously, the high codepoint matched the low one, but not
245 vice versa.
246
247 (d) The processing of \h, \H, \v, and \ in character classes now makes use
248 of the new class addition function, using character lists defined as
249 macros alongside the case definitions of 20 above.
250
251 (e) Caseless back references now work with characters that have more than
252 one other case.
253
254 (f) General caseless matching of characters with more than one other case
255 is supported.
256
257 22. Unicode character properties were updated from Unicode 6.2.0
258
259 23. Improved CMake support under Windows. Patch by Daniel Richard G.
260
261 24. Add support for 32-bit character strings, and UTF-32
262
263 25. Major JIT compiler update (code refactoring and bugfixing).
264 Experimental Sparc 32 support is added.
265
266 26. Applied a modified version of Daniel Richard G's patch to create
267 pcre.h.generic and config.h.generic by "make" instead of in the
268 PrepareRelease script.
269
270 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
271 pcre_compile.c when checking for a zero character.
272
273 28. Introducing a native interface for JIT. Through this interface, the compiled
274 machine code can be directly executed. The purpose of this interface is to
275 provide fast pattern matching, so several sanity checks are not performed.
276 However, feature tests are still performed. The new interface provides
277 1.4x speedup compared to the old one.
278
279 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
280 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
281 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
282
283 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
284 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
285 the "old" RFC 2279). Instead, it ended up passing a negative length to
286 pcre_exec().
287
288 31. Add support for GCC's visibility feature to hide internal functions.
289
290 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
291 "unknown -C option" after outputting 0 or 1.
292
293 33. There is now support for generating a code coverage report for the test
294 suite in environments where gcc is the compiler and lcov is installed. This
295 is mainly for the benefit of the developers.
296
297 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
298 unaddressable using valgrind annotations, allowing valgrind to detect
299 invalid memory accesses. This is mainly for the benefit of the developers.
300
301 25. (*UTF) can now be used to start a pattern in any of the three libraries.
302
303 26. Give configure error if --enable-cpp but no C++ compiler found.
304
305
306 Version 8.31 06-July-2012
307 -------------------------
308
309 1. Fixing a wrong JIT test case and some compiler warnings.
310
311 2. Removed a bashism from the RunTest script.
312
313 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
314 to unsigned type, result still unsigned" that was given by an MS compiler
315 on encountering the code "-sizeof(xxx)".
316
317 4. Partial matching support is added to the JIT compiler.
318
319 5. Fixed several bugs concerned with partial matching of items that consist
320 of more than one character:
321
322 (a) /^(..)\1/ did not partially match "aba" because checking references was
323 done on an "all or nothing" basis. This also applied to repeated
324 references.
325
326 (b) \R did not give a hard partial match if \r was found at the end of the
327 subject.
328
329 (c) \X did not give a hard partial match after matching one or more
330 characters at the end of the subject.
331
332 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
333 a partial match for the string "\r".
334
335 (e) When newline was set to CRLF, the metacharacter "." did not recognize
336 a partial match for a CR character at the end of the subject string.
337
338 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
339 running pcretest, the text "(JIT)" added to the output whenever JIT is
340 actually used to run the match.
341
342 7. Individual JIT compile options can be set in pcretest by following -s+[+]
343 or /S+[+] with a digit between 1 and 7.
344
345 8. OP_NOT now supports any UTF character not just single-byte ones.
346
347 9. (*MARK) control verb is now supported by the JIT compiler.
348
349 10. The command "./RunTest list" lists the available tests without actually
350 running any of them. (Because I keep forgetting what they all are.)
351
352 11. Add PCRE_INFO_MAXLOOKBEHIND.
353
354 12. Applied a (slightly modified) user-supplied patch that improves performance
355 when the heap is used for recursion (compiled with --disable-stack-for-
356 recursion). Instead of malloc and free for each heap frame each time a
357 logical recursion happens, frames are retained on a chain and re-used where
358 possible. This sometimes gives as much as 30% improvement.
359
360 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
361 call.
362
363 14. As documented, (*COMMIT) is now confined to within a positive assertion.
364
365 15. It is now possible to link pcretest with libedit as an alternative to
366 libreadline.
367
368 16. (*COMMIT) control verb is now supported by the JIT compiler.
369
370 17. The Unicode data tables have been updated to Unicode 6.1.0.
371
372 18. Added --file-list option to pcregrep.
373
374 19. Added binary file support to pcregrep, including the -a, --binary-files,
375 -I, and --text options.
376
377 20. The madvise function is renamed for posix_madvise for QNX compatibility
378 reasons. Fixed by Giuseppe D'Angelo.
379
380 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
381
382 22. Changed the option for creating symbolic links for 16-bit man pages from
383 -s to -sf so that re-installing does not cause issues.
384
385 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
386
387 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
388 matches in certain environments (the workspace was not being correctly
389 retained). Also added to pcre_dfa_exec() a simple plausibility check on
390 some of the workspace data at the beginning of a restart.
391
392 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
393 was not doing so when it should - probably a typo introduced by SVN 528
394 (change 8.10/14).
395
396 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
397 \w+ when the character tables indicated that \x{c4} was a word character.
398 There were several related cases, all because the tests for doing a table
399 lookup were testing for characters less than 127 instead of 255.
400
401 27. If a pattern contains capturing parentheses that are not used in a match,
402 their slots in the ovector are set to -1. For those that are higher than
403 any matched groups, this happens at the end of processing. In the case when
404 there were back references that the ovector was too small to contain
405 (causing temporary malloc'd memory to be used during matching), and the
406 highest capturing number was not used, memory off the end of the ovector
407 was incorrectly being set to -1. (It was using the size of the temporary
408 memory instead of the true size.)
409
410 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
411 ovector size, it uses memory at the end of the block that it has got.
412
413 29. Check for an overlong MARK name and give an error at compile time. The
414 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
415
416 30. JIT compiler update.
417
418 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
419 Rill for the patch.
420
421 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
422
423 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
424
425 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
426 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
427 was enabled.
428
429 35. Improve JIT code generation for greedy plus quantifier.
430
431 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
432 1 to "aa" instead of to an empty string. The bug affected repeated groups
433 that could potentially match an empty string.
434
435 37. Optimizing single character iterators in JIT.
436
437 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
438 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
439 codepoints that are too big for the mode are faulted, and in a UTF mode,
440 disallowed codepoints are also faulted.
441
442 39. If PCRE was compiled with UTF support, in three places in the DFA
443 matcher there was code that should only have been obeyed in UTF mode, but
444 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
445 processing when bytes with values greater than 127 were present. In 16-bit
446 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
447 both cases the values are those that cannot be the first data item in a UTF
448 character. The three items that might have provoked this were recursions,
449 possessively repeated groups, and atomic groups.
450
451 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
452 and pcregrep, because some OS require shared objects to be explicitly
453 passed to ld, causing the link step to fail if they are not.
454
455 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
456 mode, patterns that started with \h* or \R* might be incorrectly matched.
457
458
459 Version 8.30 04-February-2012
460 -----------------------------
461
462 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
463 name is defined in ctype.h.
464
465 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
466 only in quite long subpatterns.
467
468 3. Removed the function pcre_info(), which has been obsolete and deprecated
469 since it was replaced by pcre_fullinfo() in February 2000.
470
471 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
472 match a (*MARK), and the match failed at the start of the subject, a
473 reference to memory before the start of the subject could occur. This bug
474 was introduced by fix 17 of release 8.21.
475
476 5. A reference to an unset group with zero minimum repetition was giving
477 totally wrong answers (in non-JavaScript-compatibility mode). For example,
478 /(another)?(\1?)test/ matched against "hello world test". This bug was
479 introduced in release 8.13.
480
481 6. Add support for 16-bit character strings (a large amount of work involving
482 many changes and refactorings).
483
484 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
485 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
486 from a file.
487
488 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
489 rounding is not applied in this particular case).
490
491 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
492 if they appear, or are escaped, in patterns.
493
494 10. Get rid of a number of -Wunused-but-set-variable warnings.
495
496 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
497 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
498 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
499 also returns the mark "x". This bug applied to capturing parentheses,
500 non-capturing parentheses, and atomic parentheses. It also applied to some
501 assertions.
502
503 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
504 information out of configure.ac instead of relying on pcre.h.generic, which
505 is not stored in the repository.
506
507 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
508 -lreadline.
509
510 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
511
512 15. Applied Graycode's patch to put the top-level frame on the stack rather
513 than the heap when not using the stack for recursion. This gives a
514 performance improvement in many cases when recursion is not deep.
515
516 16. Experimental code added to "pcretest -C" to output the stack frame size.
517
518
519 Version 8.21 12-Dec-2011
520 ------------------------
521
522 1. Updating the JIT compiler.
523
524 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
525 are added as well.
526
527 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
528 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
529 calling _pcre_jit_exec. Some extra comments are added.
530
531 4. (*MARK) settings inside atomic groups that do not contain any capturing
532 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
533 was introduced by change 18 for 8.20.
534
535 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
536 ECMA-262 standard.
537
538 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
539 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
540 This bug was probably introduced by change 9 of 8.13.
541
542 7. While fixing 6 above, I noticed that a number of other items were being
543 incorrectly rejected as "not fixed length". This arose partly because newer
544 opcodes had not been added to the fixed-length checking code. I have (a)
545 corrected the bug and added tests for these items, and (b) arranged for an
546 error to occur if an unknown opcode is encountered while checking for fixed
547 length instead of just assuming "not fixed length". The items that were
548 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
549 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
550 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
551
552 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
553 being incorrectly compiled and would have given unpredicatble results.
554
555 9. A possessively repeated subpattern with minimum repeat count greater than
556 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
557 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
558 the first (A) could occur when it should not.
559
560 10. Add a cast and remove a redundant test from the code.
561
562 11. JIT should use pcre_malloc/pcre_free for allocation.
563
564 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
565 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
566 is anything other than /usr, -L is still shown).
567
568 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
569
570 14. Perl does not support \N without a following name in a [] class; PCRE now
571 also gives an error.
572
573 15. If a forward reference was repeated with an upper limit of around 2000,
574 it caused the error "internal error: overran compiling workspace". The
575 maximum number of forward references (including repeats) was limited by the
576 internal workspace, and dependent on the LINK_SIZE. The code has been
577 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
578 the default depends on LINK_SIZE. There is a new upper limit (for safety)
579 of around 200,000 forward references. While doing this, I also speeded up
580 the filling in of repeated forward references.
581
582 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
583 incorrectly expecting the subject to contain another "a" after the start.
584
585 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
586 in the match, the SKIP should be ignored. This was not happening; instead
587 the SKIP was being treated as NOMATCH. For patterns such as
588 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
589 tested.
590
591 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
592 now much more compatible with Perl, in particular in cases where the result
593 is a non-match for a non-anchored pattern. For example, if
594 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
595 "m", where previously it did not return a name. A side effect of this
596 change is that for partial matches, the last encountered mark name is
597 returned, as for non matches. A number of tests that were previously not
598 Perl-compatible have been moved into the Perl-compatible test files. The
599 refactoring has had the pleasing side effect of removing one argument from
600 the match() function, thus reducing its stack requirements.
601
602 19. If the /S+ option was used in pcretest to study a pattern using JIT,
603 subsequent uses of /S (without +) incorrectly behaved like /S+.
604
605 21. Retrieve executable code size support for the JIT compiler and fixing
606 some warnings.
607
608 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
609 not work when the shorter character appeared right at the end of the
610 subject string.
611
612 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
613 systems.
614
615 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
616 output it when the /M option is used in pcretest.
617
618 25. The CheckMan script was not being included in the distribution. Also, added
619 an explicit "perl" to run Perl scripts from the PrepareRelease script
620 because this is reportedly needed in Windows.
621
622 26. If study data was being save in a file and studying had not found a set of
623 "starts with" bytes for the pattern, the data written to the file (though
624 never used) was taken from uninitialized memory and so caused valgrind to
625 complain.
626
627 27. Updated RunTest.bat as provided by Sheri Pierce.
628
629 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
630
631 29. Computation of memory usage for the table of capturing group names was
632 giving an unnecessarily large value.
633
634
635 Version 8.20 21-Oct-2011
636 ------------------------
637
638 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
639 a POSIX class. After further experiments with Perl, which convinced me that
640 Perl has bugs and confusions, a closing square bracket is no longer allowed
641 in a POSIX name. This bug also affected patterns with classes that started
642 with full stops.
643
644 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
645 captured substring, but while checking the failing first alternative,
646 substring 1 is temporarily captured. If the output vector supplied to
647 pcre_exec() was not big enough for this capture, the yield of the function
648 was still zero ("insufficient space for captured substrings"). This cannot
649 be totally fixed without adding another stack variable, which seems a lot
650 of expense for a edge case. However, I have improved the situation in cases
651 such as /(a)(b)x|abc/ matched against "abc", where the return code
652 indicates that fewer than the maximum number of slots in the ovector have
653 been set.
654
655 3. Related to (2) above: when there are more back references in a pattern than
656 slots in the output vector, pcre_exec() uses temporary memory during
657 matching, and copies in the captures as far as possible afterwards. It was
658 using the entire output vector, but this conflicts with the specification
659 that only 2/3 is used for passing back captured substrings. Now it uses
660 only the first 2/3, for compatibility. This is, of course, another edge
661 case.
662
663 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
664 main code base, and can be used by building with --enable-jit. When this is
665 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
666 runtime --no-jit option is given.
667
668 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
669 ovector, the return from the function was zero, implying that there were
670 other matches that did not fit. The correct "exactly full" value is now
671 returned.
672
673 6. If a subpattern that was called recursively or as a subroutine contained
674 (*PRUNE) or any other control that caused it to give a non-standard return,
675 invalid errors such as "Error -26 (nested recursion at the same subject
676 position)" or even infinite loops could occur.
677
678 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
679 computing the minimum length on reaching *ACCEPT, and so ended up with the
680 wrong value of 1 rather than 0. Further investigation indicates that
681 computing a minimum subject length in the presence of *ACCEPT is difficult
682 (think back references, subroutine calls), and so I have changed the code
683 so that no minimum is registered for a pattern that contains *ACCEPT.
684
685 8. If (*THEN) was present in the first (true) branch of a conditional group,
686 it was not handled as intended. [But see 16 below.]
687
688 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
689 Sheri Pierce.
690
691 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
692 the first byte in a match must be "a".
693
694 11. Change 17 for 8.13 increased the recursion depth for patterns like
695 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
696 pattern contains any instances of (*THEN). If it does not, the old
697 optimizations are restored. It would be nice to do this on a per-group
698 basis, but at the moment that is not feasible.
699
700 12. In some environments, the output of pcretest -C is CRLF terminated. This
701 broke RunTest's code that checks for the link size. A single white space
702 character after the value is now allowed for.
703
704 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
705 For "fr", it uses the Windows-specific input and output files.
706
707 14. If (*THEN) appeared in a group that was called recursively or as a
708 subroutine, it did not work as intended. [But see next item.]
709
710 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
711 pattern fragments (but not containing any | characters). If A and B are
712 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
713 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
714 D. In other words, Perl considers parentheses that do not contain any |
715 characters to be part of a surrounding alternative, whereas PCRE was
716 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
717 differently. PCRE now behaves in the same way as Perl, except in the case
718 of subroutine/recursion calls such as (?1) which have in any case always
719 been different (but PCRE had them first :-).
720
721 16. Related to 15 above: Perl does not treat the | in a conditional group as
722 creating alternatives. Such a group is treated in the same way as an
723 ordinary group without any | characters when processing (*THEN). PCRE has
724 been changed to match Perl's behaviour.
725
726 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
727 RunGrepTest script failed.
728
729 18. Change 22 for version 13 caused atomic groups to use more stack. This is
730 inevitable for groups that contain captures, but it can lead to a lot of
731 stack use in large patterns. The old behaviour has been restored for atomic
732 groups that do not contain any capturing parentheses.
733
734 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
735 suppress the check for a minimum subject length at run time. (If it was
736 given to pcre_exec() or pcre_dfa_exec() it did work.)
737
738 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
739 fail to work when decoding hex characters in data strings in EBCDIC
740 environments.
741
742 21. It appears that in at least one Mac OS environment, the isxdigit() function
743 is implemented as a macro that evaluates to its argument more than once,
744 contravening the C 90 Standard (I haven't checked a later standard). There
745 was an instance in pcretest which caused it to go wrong when processing
746 \x{...} escapes in subject strings. The has been rewritten to avoid using
747 things like p++ in the argument of isxdigit().
748
749
750 Version 8.13 16-Aug-2011
751 ------------------------
752
753 1. The Unicode data tables have been updated to Unicode 6.0.0.
754
755 2. Two minor typos in pcre_internal.h have been fixed.
756
757 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
758 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
759 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
760
761 4. There were a number of related bugs in the code for matching backrefences
762 caselessly in UTF-8 mode when codes for the characters concerned were
763 different numbers of bytes. For example, U+023A and U+2C65 are an upper
764 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
765 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
766 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
767 2-byte code at the end of the subject (it thought there wasn't enough data
768 left).
769
770 5. Comprehensive information about what went wrong is now returned by
771 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
772 as the output vector has at least 2 elements. The offset of the start of
773 the failing character and a reason code are placed in the vector.
774
775 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
776 now returned is for the first byte of the failing character, instead of the
777 last byte inspected. This is an incompatible change, but I hope it is small
778 enough not to be a problem. It makes the returned offset consistent with
779 pcre_exec() and pcre_dfa_exec().
780
781 7. pcretest now gives a text phrase as well as the error number when
782 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
783 failure, the offset and reason code are output.
784
785 8. When \R was used with a maximizing quantifier it failed to skip backwards
786 over a \r\n pair if the subsequent match failed. Instead, it just skipped
787 back over a single character (\n). This seems wrong (because it treated the
788 two characters as a single entity when going forwards), conflicts with the
789 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
790 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
791 has been changed.
792
793 9. Some internal refactoring has changed the processing so that the handling
794 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
795 time (the PCRE_DOTALL option was changed this way some time ago: version
796 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
797 which was always a bit of a fudge. It also means that there is one less
798 argument for the match() function, which reduces its stack requirements
799 slightly. This change also fixes an incompatibility with Perl: the pattern
800 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
801
802 10. More internal refactoring has drastically reduced the number of recursive
803 calls to match() for possessively repeated groups such as (abc)++ when
804 using pcre_exec().
805
806 11. While implementing 10, a number of bugs in the handling of groups were
807 discovered and fixed:
808
809 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
810 (a|)*(?1) gave a compile-time internal error.
811 ((a|)+)+ did not notice that the outer group could match an empty string.
812 (^a|^)+ was not marked as anchored.
813 (.*a|.*)+ was not marked as matching at start or after a newline.
814
815 12. Yet more internal refactoring has removed another argument from the match()
816 function. Special calls to this function are now indicated by setting a
817 value in a variable in the "match data" data block.
818
819 13. Be more explicit in pcre_study() instead of relying on "default" for
820 opcodes that mean there is no starting character; this means that when new
821 ones are added and accidentally left out of pcre_study(), testing should
822 pick them up.
823
824 14. The -s option of pcretest has been documented for ages as being an old
825 synonym of -m (show memory usage). I have changed it to mean "force study
826 for every regex", that is, assume /S for every regex. This is similar to -i
827 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
828 using it. It makes it easier to run collections of tests with and without
829 study enabled, and thereby test pcre_study() more easily. All the standard
830 tests are now run with and without -s (but some patterns can be marked as
831 "never study" - see 20 below).
832
833 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
834 restoration of the capturing data to the outer values was not happening
835 correctly.
836
837 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
838 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
839 pattern had matched an empty string, and so incorrectly returned a no
840 match.
841
842 17. There was optimizing code for the last branch of non-capturing parentheses,
843 and also for the obeyed branch of a conditional subexpression, which used
844 tail recursion to cut down on stack usage. Unfortunately, now that there is
845 the possibility of (*THEN) occurring in these branches, tail recursion is
846 no longer possible because the return has to be checked for (*THEN). These
847 two optimizations have therefore been removed. [But see 8.20/11 above.]
848
849 18. If a pattern containing \R was studied, it was assumed that \R always
850 matched two bytes, thus causing the minimum subject length to be
851 incorrectly computed because \R can also match just one byte.
852
853 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
854 was incorrectly computed.
855
856 20. If /S is present twice on a test pattern in pcretest input, it now
857 *disables* studying, thereby overriding the use of -s on the command line
858 (see 14 above). This is necessary for one or two tests to keep the output
859 identical in both cases.
860
861 21. When (*ACCEPT) was used in an assertion that matched an empty string and
862 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
863
864 22. When an atomic group that contained a capturing parenthesis was
865 successfully matched, but the branch in which it appeared failed, the
866 capturing was not being forgotten if a higher numbered group was later
867 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
868 group 1 to "a", when in fact it should be unset. This applied to multi-
869 branched capturing and non-capturing groups, repeated or not, and also to
870 positive assertions (capturing in negative assertions does not happen
871 in PCRE) and also to nested atomic groups.
872
873 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
874 subject after a captured substring, to make it easier to tell which of a
875 number of identical substrings has been captured.
876
877 24. The way atomic groups are processed by pcre_exec() has been changed so that
878 if they are repeated, backtracking one repetition now resets captured
879 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
880 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
881 "aaa". Previously, it would have been "a". As part of this code
882 refactoring, the way recursive calls are handled has also been changed.
883
884 25. If an assertion condition captured any substrings, they were not passed
885 back unless some other capturing happened later. For example, if
886 (?(?=(a))a) was matched against "a", no capturing was returned.
887
888 26. When studying a pattern that contained subroutine calls or assertions,
889 the code for finding the minimum length of a possible match was handling
890 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
891 group 1 called group 2 while simultaneously a separate group 2 called group
892 1). A stack overflow occurred in this case. I have fixed this by limiting
893 the recursion depth to 10.
894
895 27. Updated RunTest.bat in the distribution to the version supplied by Tom
896 Fortmann. This supports explicit test numbers on the command line, and has
897 argument validation and error reporting.
898
899 28. An instance of \X with an unlimited repeat could fail if at any point the
900 first character it looked at was a mark character.
901
902 29. Some minor code refactoring concerning Unicode properties and scripts
903 should reduce the stack requirement of match() slightly.
904
905 30. Added the '=' option to pcretest to check the setting of unused capturing
906 slots at the end of the pattern, which are documented as being -1, but are
907 not included in the return count.
908
909 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
910 compiled something random. Now it gives a compile-time error (as does
911 Perl).
912
913 32. A *MARK encountered during the processing of a positive assertion is now
914 recorded and passed back (compatible with Perl).
915
916 33. If --only-matching or --colour was set on a pcregrep call whose pattern
917 had alternative anchored branches, the search for a second match in a line
918 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
919 matched the line "0102" twice. The same bug affected patterns that started
920 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
921 twice.
922
923 34. Previously, PCRE did not allow quantification of assertions. However, Perl
924 does, and because of capturing effects, quantifying parenthesized
925 assertions may at times be useful. Quantifiers are now allowed for
926 parenthesized assertions.
927
928 35. A minor code tidy in pcre_compile() when checking options for \R usage.
929
930 36. \g was being checked for fancy things in a character class, when it should
931 just be a literal "g".
932
933 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
934 appearance of a nested POSIX class supersedes an apparent external class.
935 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
936 unescaped square brackets may also appear as part of class names. For
937 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
938 more like Perl. (But see 8.20/1 above.)
939
940 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
941 was because it thought it was \N{name}, which is not supported).
942
943 39. Add minix to OS list not supporting the -S option in pcretest.
944
945 40. PCRE tries to detect cases of infinite recursion at compile time, but it
946 cannot analyze patterns in sufficient detail to catch mutual recursions
947 such as ((?1))((?2)). There is now a runtime test that gives an error if a
948 subgroup is called recursively as a subpattern for a second time at the
949 same position in the subject string. In previous releases this might have
950 been caught by the recursion limit, or it might have run out of stack.
951
952 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
953 happen only once. PCRE was, however incorrectly giving a compile time error
954 "recursive call could loop indefinitely" because it cannot analyze the
955 pattern in sufficient detail. The compile time test no longer happens when
956 PCRE is compiling a conditional subpattern, but actual runaway loops are
957 now caught at runtime (see 40 above).
958
959 42. It seems that Perl allows any characters other than a closing parenthesis
960 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
961 has been changed to be the same.
962
963 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
964 as not to get warnings when autogen.sh is called. Also changed
965 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
966
967 44. To help people who use pcregrep to scan files containing exceedingly long
968 lines, the following changes have been made:
969
970 (a) The default value of the buffer size parameter has been increased from
971 8K to 20K. (The actual buffer used is three times this size.)
972
973 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
974 PCRE is built.
975
976 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
977 to be set at run time.
978
979 (d) Numerical values in pcregrep options can be followed by K or M, for
980 example --buffer-size=50K.
981
982 (e) If a line being scanned overflows pcregrep's buffer, an error is now
983 given and the return code is set to 2.
984
985 45. Add a pointer to the latest mark to the callout data block.
986
987 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
988 partial match of an empty string instead of no match. This was specific to
989 the use of ".".
990
991 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
992 complete match instead of a partial match. This bug was dependent on both
993 the PCRE_UTF8 and PCRE_DOTALL options being set.
994
995 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
996 starting byte set, because \b was not being ignored.
997
998
999 Version 8.12 15-Jan-2011
1000 ------------------------
1001
1002 1. Fixed some typos in the markup of the man pages, and wrote a script that
1003 checks for such things as part of the documentation building process.
1004
1005 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1006 --match-limit and --recursion-limit options (added for 8.11). In
1007 particular, this made one of the standard tests fail. (The integer value
1008 went into the wrong half of a long int.)
1009
1010 3. If the --colour option was given to pcregrep with -v (invert match), it
1011 did strange things, either producing crazy output, or crashing. It should,
1012 of course, ignore a request for colour when reporting lines that do not
1013 match.
1014
1015 4. Another pcregrep bug caused similar problems if --colour was specified with
1016 -M (multiline) and the pattern match finished with a line ending.
1017
1018 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1019 matched in multiline mode, the following line was shown as part of the
1020 match. This seems wrong, so I have changed it.
1021
1022 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1023 the check for further matches in the same line (so they could be coloured)
1024 to overrun the end of the current line. If another match was found, it was
1025 incorrectly shown (and then shown again when found in the next line).
1026
1027 7. If pcregrep was compiled under Windows, there was a reference to the
1028 function pcregrep_exit() before it was defined. I am assuming this was
1029 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1030 reported by a user. I've moved the definition above the reference.
1031
1032
1033 Version 8.11 10-Dec-2010
1034 ------------------------
1035
1036 1. (*THEN) was not working properly if there were untried alternatives prior
1037 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1038 backtracked to try for "b" instead of moving to the next alternative branch
1039 at the same level (in this case, to look for "c"). The Perl documentation
1040 is clear that when (*THEN) is backtracked onto, it goes to the "next
1041 alternative in the innermost enclosing group".
1042
1043 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1044 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1045 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1046 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1047 (*THEN).
1048
1049 3. If \s appeared in a character class, it removed the VT character from
1050 the class, even if it had been included by some previous item, for example
1051 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1052 of \s, but is part of the POSIX "space" class.)
1053
1054 4. A partial match never returns an empty string (because you can always
1055 match an empty string at the end of the subject); however the checking for
1056 an empty string was starting at the "start of match" point. This has been
1057 changed to the "earliest inspected character" point, because the returned
1058 data for a partial match starts at this character. This means that, for
1059 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1060 (previously it gave "no match").
1061
1062 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1063 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1064 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1065 has an implication that the given string is incomplete (because a partial
1066 match is preferred over a full match). For this reason, these items now
1067 give a partial match in this situation. [Aside: previously, the one case
1068 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1069 match rather than a full match, which was wrong by the old rules, but is
1070 now correct.]
1071
1072 6. There was a bug in the handling of #-introduced comments, recognized when
1073 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1074 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1075 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1076 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1077 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1078 places in pcre_compile().
1079
1080 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1081 comments when looking ahead for named forward references to subpatterns,
1082 the only newline sequence it recognized was NL. It now handles newlines
1083 according to the set newline convention.
1084
1085 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1086 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1087 cater for a lack of strerror(). These oversights have been fixed.
1088
1089 9. Added --match-limit and --recursion-limit to pcregrep.
1090
1091 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1092
1093 11. When the -o option was used, pcregrep was setting a return code of 1, even
1094 when matches were found, and --line-buffered was not being honoured.
1095
1096 12. Added an optional parentheses number to the -o and --only-matching options
1097 of pcregrep.
1098
1099 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1100 can match an empty string. The code to do it in pcretest and pcredemo
1101 needed fixing:
1102
1103 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1104 only one byte after an empty string match just before CRLF (this case
1105 just got forgotten; "any" and "anycrlf" were OK).
1106
1107 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1108 mode when an empty string match preceded an ASCII character followed by
1109 a non-ASCII character. (The code for advancing by one character rather
1110 than one byte was nonsense.)
1111
1112 (c) The pcredemo.c sample program did not have any code at all to handle
1113 the cases when CRLF is a valid newline sequence.
1114
1115 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1116 as a starting offset was within the subject string. There is now a new
1117 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1118 negative or greater than the length of the string. In order to test this,
1119 pcretest is extended to allow the setting of negative starting offsets.
1120
1121 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1122 starting offset points to the beginning of a UTF-8 character was
1123 unnecessarily clumsy. I tidied it up.
1124
1125 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1126 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1127
1128 17. Nobody had reported that the --include_dir option, which was added in
1129 release 7.7 should have been called --include-dir (hyphen, not underscore)
1130 for compatibility with GNU grep. I have changed it to --include-dir, but
1131 left --include_dir as an undocumented synonym, and the same for
1132 --exclude-dir, though that is not available in GNU grep, at least as of
1133 release 2.5.4.
1134
1135 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1136 characters from a string of bytes) have been redefined so as not to use
1137 loops, in order to improve performance in some environments. At the same
1138 time, I abstracted some of the common code into auxiliary macros to save
1139 repetition (this should not affect the compiled code).
1140
1141 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1142 compile-time error is now given if \c is not followed by an ASCII
1143 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1144 different, and any byte value is allowed.)
1145
1146 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1147 START_OPTIMIZE option, which is now allowed at compile time - but just
1148 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1149 to pcregrep and other applications that have no direct access to PCRE
1150 options. The new /Y option in pcretest sets this option when calling
1151 pcre_compile().
1152
1153 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1154 back references. Groups containing recursive back references were forced to
1155 be atomic by that change, but in the case of named groups, the amount of
1156 memory required was incorrectly computed, leading to "Failed: internal
1157 error: code overflow". This has been fixed.
1158
1159 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1160 pcretest.c, to avoid build problems in some Borland environments.
1161
1162
1163 Version 8.10 25-Jun-2010
1164 ------------------------
1165
1166 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1167 THEN.
1168
1169 2. (*ACCEPT) was not working when inside an atomic group.
1170
1171 3. Inside a character class, \B is treated as a literal by default, but
1172 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1173 causes the error). The code is unchanged, but I tidied the documentation.
1174
1175 4. Inside a character class, PCRE always treated \R and \X as literals,
1176 whereas Perl faults them if its -w option is set. I have changed PCRE so
1177 that it faults them when PCRE_EXTRA is set.
1178
1179 5. Added support for \N, which always matches any character other than
1180 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1181
1182 6. When compiling pcregrep with newer versions of gcc which may have
1183 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1184 declared with attribute warn_unused_result" were given. Just casting the
1185 result to (void) does not stop the warnings; a more elaborate fudge is
1186 needed. I've used a macro to implement this.
1187
1188 7. Minor change to pcretest.c to avoid a compiler warning.
1189
1190 8. Added four artifical Unicode properties to help with an option to make
1191 \s etc use properties (see next item). The new properties are: Xan
1192 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1193
1194 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1195 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1196 this option. Modified pcretest to add /W to test this facility. Added
1197 REG_UCP to make it available via the POSIX interface.
1198
1199 10. Added --line-buffered to pcregrep.
1200
1201 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1202 studied, and the match started with a letter with a code point greater than
1203 127 whose first byte was different to the first byte of the other case of
1204 the letter, the other case of this starting letter was not recognized
1205 (#976).
1206
1207 12. If a pattern that was studied started with a repeated Unicode property
1208 test, for example, \p{Nd}+, there was the theoretical possibility of
1209 setting up an incorrect bitmap of starting bytes, but fortunately it could
1210 not have actually happened in practice until change 8 above was made (it
1211 added property types that matched character-matching opcodes).
1212
1213 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1214 possible starting bytes for non-anchored patterns.
1215
1216 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1217 \R, and also a number of cases that involve Unicode properties, both
1218 explicit and implicit when PCRE_UCP is set.
1219
1220 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1221 input, it could crash or give wrong results if characters with values
1222 greater than 0xc0 were present in the subject string. (Detail: it assumed
1223 UTF-8 input when processing these items.)
1224
1225 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1226 size_t is 64-bit (#991).
1227
1228 17. Added a check for running out of memory when PCRE is compiled with
1229 --disable-stack-for-recursion (#990).
1230
1231 18. If the last data line in a file for pcretest does not have a newline on
1232 the end, a newline was missing in the output.
1233
1234 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1235 less than 128) in its various bitmaps. However, there is a facility for
1236 generating tables according to the current locale when PCRE is compiled. It
1237 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1238 characters, are recognized by isspace() and therefore were getting set in
1239 these tables, and indeed these tables seem to approximate to ISO 8859. This
1240 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1241 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1242 which of course cannot start UTF-8 characters. I have changed the code so
1243 that only real ASCII characters (less than 128) and the correct starting
1244 bytes for UTF-8 encodings are set for characters greater than 127 when in
1245 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1246 altogether.)
1247
1248 20. Added the /T option to pcretest so as to be able to run tests with non-
1249 standard character tables, thus making it possible to include the tests
1250 used for 19 above in the standard set of tests.
1251
1252 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1253 reference to a subpattern the other side of a comment that contains an
1254 opening parenthesis caused either an internal compiling error, or a
1255 reference to the wrong subpattern.
1256
1257
1258 Version 8.02 19-Mar-2010
1259 ------------------------
1260
1261 1. The Unicode data tables have been updated to Unicode 5.2.0.
1262
1263 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1264 configured.
1265
1266 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1267 original author of that file, following a query about its status.
1268
1269 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1270 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1271
1272 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1273 quantifier applied to a forward-referencing subroutine call, could compile
1274 incorrect code or give the error "internal error: previously-checked
1275 referenced subpattern not found".
1276
1277 6. Both MS Visual Studio and Symbian OS have problems with initializing
1278 variables to point to external functions. For these systems, therefore,
1279 pcre_malloc etc. are now initialized to local functions that call the
1280 relevant global functions.
1281
1282 7. There were two entries missing in the vectors called coptable and poptable
1283 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1284 I've fixed the data, and added a kludgy way of testing at compile time that
1285 the lengths are correct (equal to the number of opcodes).
1286
1287 8. Following on from 7, I added a similar kludge to check the length of the
1288 eint vector in pcreposix.c.
1289
1290 9. Error texts for pcre_compile() are held as one long string to avoid too
1291 much relocation at load time. To find a text, the string is searched,
1292 counting zeros. There was no check for running off the end of the string,
1293 which could happen if a new error number was added without updating the
1294 string.
1295
1296 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1297
1298 11. \K was not working if it appeared in an atomic group or in a group that
1299 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1300 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1301 the assertion is positive, but not if it is negative.
1302
1303 12. Change 11 fortuitously reduced the size of the stack frame used in the
1304 "match()" function of pcre_exec.c by one pointer. Forthcoming
1305 implementation of support for (*MARK) will need an extra pointer on the
1306 stack; I have reserved it now, so that the stack frame size does not
1307 decrease.
1308
1309 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1310 item in branch that calls a recursion is a subroutine call - as in the
1311 second branch in the above example - was incorrectly given the compile-
1312 time error "recursive call could loop indefinitely" because pcre_compile()
1313 was not correctly checking the subroutine for matching a non-empty string.
1314
1315 14. The checks for overrunning compiling workspace could trigger after an
1316 overrun had occurred. This is a "should never occur" error, but it can be
1317 triggered by pathological patterns such as hundreds of nested parentheses.
1318 The checks now trigger 100 bytes before the end of the workspace.
1319
1320 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1321
1322
1323 Version 8.01 19-Jan-2010
1324 ------------------------
1325
1326 1. If a pattern contained a conditional subpattern with only one branch (in
1327 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1328 computed the wrong minimum data length (which is of course zero for such
1329 subpatterns). This could cause incorrect "no match" results.
1330
1331 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1332 the pattern is reset in the first branch, pcre_compile() failed with
1333 "internal error: code overflow at offset...". This happened only when
1334 the reset was to the original external option setting. (An optimization
1335 abstracts leading options settings into an external setting, which was the
1336 cause of this.)
1337
1338 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1339 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1340 assertion pattern did not match (meaning that the assertion was true), it
1341 was incorrectly treated as false if the SKIP had been reached during the
1342 matching. This also applied to assertions used as conditions.
1343
1344 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1345 assertion subpattern, including such a pattern used as a condition,
1346 unpredictable results occurred, instead of the error return
1347 PCRE_ERROR_DFA_UITEM.
1348
1349 5. The C++ GlobalReplace function was not working like Perl for the special
1350 situation when an empty string is matched. It now does the fancy magic
1351 stuff that is necessary.
1352
1353 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1354 removed. (These were left over from very, very early versions of PCRE.)
1355
1356 7. Some cosmetic changes to the code to make life easier when compiling it
1357 as part of something else:
1358
1359 (a) Change DEBUG to PCRE_DEBUG.
1360
1361 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1362 called "current" as "current_branch", to prevent a collision with the
1363 Linux macro when compiled as a kernel module.
1364
1365 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1366 prevent a collision with the Linux macro when compiled as a kernel
1367 module.
1368
1369 8. In pcre_compile() there are some checks for integer overflows that used to
1370 cast potentially large values to (double). This has been changed to that
1371 when building, a check for int64_t is made, and if it is found, it is used
1372 instead, thus avoiding the use of floating point arithmetic. (There is no
1373 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1374 double.
1375
1376 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1377 2005 (difference between two addresses compared to an unsigned value).
1378
1379 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1380 custom one, because of the following reported problem in Windows:
1381
1382 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1383 under Win32.
1384 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1385 therefore missing the function definition.
1386 - The compiler thus generates a "C" signature for the test function.
1387 - The linker fails to find the "C" function.
1388 - PCRE fails to configure if asked to do so against libbz2.
1389
1390 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1391 messages were output:
1392
1393 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1394 rerunning libtoolize, to keep the correct libtool macros in-tree.
1395 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1396
1397 I have done both of these things.
1398
1399 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1400 most of the time, it *can* run out if it is given a pattern that contains a
1401 runaway infinite recursion. I updated the discussion in the pcrestack man
1402 page.
1403
1404 13. Now that we have gone to the x.xx style of version numbers, the minor
1405 version may start with zero. Using 08 or 09 is a bad idea because users
1406 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1407 interpreted as invalid octal numbers. I've updated the previous comment in
1408 configure.ac, and also added a check that gives an error if 08 or 09 are
1409 used.
1410
1411 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1412 causing partial matching to fail when the end of the subject matched \W
1413 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1414
1415 15. There were some discrepancies between the declarations in pcre_internal.h
1416 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1417 their definitions. The declarations used "const uschar *" and the
1418 definitions used USPTR. Even though USPTR is normally defined as "const
1419 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1420 reported that: "This difference in casting confuses some C++ compilers, for
1421 example, SunCC recognizes above declarations as different functions and
1422 generates broken code for hbpcre." I have changed the declarations to use
1423 USPTR.
1424
1425 16. GNU libtool is named differently on some systems. The autogen.sh script now
1426 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1427 (FreeBSD).
1428
1429 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1430 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1431 comment: "Figure out how to create a longlong from a string: strtoll and
1432 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1433 instance, but it only takes 2 args instead of 3!"
1434
1435 18. A subtle bug concerned with back references has been fixed by a change of
1436 specification, with a corresponding code fix. A pattern such as
1437 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1438 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1439 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1440 same bug. Such groups have to be quantified to be useful, or contained
1441 inside another quantified group. (If there's no repetition, the reference
1442 can never match.) The problem arises because, having left the group and
1443 moved on to the rest of the pattern, a later failure that backtracks into
1444 the group uses the captured value from the final iteration of the group
1445 rather than the correct earlier one. I have fixed this in PCRE by forcing
1446 any group that contains a reference to itself to be an atomic group; that
1447 is, there cannot be any backtracking into it once it has completed. This is
1448 similar to recursive and subroutine calls.
1449
1450
1451 Version 8.00 19-Oct-09
1452 ----------------------
1453
1454 1. The table for translating pcre_compile() error codes into POSIX error codes
1455 was out-of-date, and there was no check on the pcre_compile() error code
1456 being within the table. This could lead to an OK return being given in
1457 error.
1458
1459 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1460 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1461 in a Windows environment.
1462
1463 3. The pcregrep --count option prints the count for each file even when it is
1464 zero, as does GNU grep. However, pcregrep was also printing all files when
1465 --files-with-matches was added. Now, when both options are given, it prints
1466 counts only for those files that have at least one match. (GNU grep just
1467 prints the file name in this circumstance, but including the count seems
1468 more useful - otherwise, why use --count?) Also ensured that the
1469 combination -clh just lists non-zero counts, with no names.
1470
1471 4. The long form of the pcregrep -F option was incorrectly implemented as
1472 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1473 but it seems right to fix it, and I didn't think it was worth preserving
1474 the old behaviour.
1475
1476 5. The command line items --regex=pattern and --regexp=pattern were not
1477 recognized by pcregrep, which required --regex pattern or --regexp pattern
1478 (with a space rather than an '='). The man page documented the '=' forms,
1479 which are compatible with GNU grep; these now work.
1480
1481 6. No libpcreposix.pc file was created for pkg-config; there was just
1482 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1483
1484 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1485 when UCP support is not needed, by modifying the Python script that
1486 generates it from Unicode data files. This should not matter if the module
1487 is correctly used as a library, but I received one complaint about 50K of
1488 unwanted data. My guess is that the person linked everything into his
1489 program rather than using a library. Anyway, it does no harm.
1490
1491 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1492 was a minimum greater than 1 for a wide character in a possessive
1493 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1494 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1495 character. Chaos in the form of incorrect output or a compiling loop could
1496 result.
1497
1498 9. The restrictions on what a pattern can contain when partial matching is
1499 requested for pcre_exec() have been removed. All patterns can now be
1500 partially matched by this function. In addition, if there are at least two
1501 slots in the offset vector, the offset of the earliest inspected character
1502 for the match and the offset of the end of the subject are set in them when
1503 PCRE_ERROR_PARTIAL is returned.
1504
1505 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1506 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1507 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1508 and may be more useful for multi-segment matching.
1509
1510 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1511 used to be given if ever the end of the subject was reached; now it is
1512 given only if matching could not proceed because another character was
1513 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1514 string "Z", which now yields "no match" instead of "partial match". In the
1515 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1516 final character ended with (*FAIL).
1517
1518 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1519 if the pattern had a "must contain" character that was already found in the
1520 earlier partial match, unless partial matching was again requested. For
1521 example, with the pattern /dog.(body)?/, the "must contain" character is
1522 "g". If the first part-match was for the string "dog", restarting with
1523 "sbody" failed. This bug has been fixed.
1524
1525 13. The string returned by pcre_dfa_exec() after a partial match has been
1526 changed so that it starts at the first inspected character rather than the
1527 first character of the match. This makes a difference only if the pattern
1528 starts with a lookbehind assertion or \b or \B (\K is not supported by
1529 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1530 matching functions compatible, and I think it's the right thing to do.
1531
1532 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1533 so that the demonstration program is easily available in environments where
1534 PCRE has not been installed from source.
1535
1536 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1537 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1538 library.
1539
1540 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1541 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1542 is not the first non-POSIX option to be added. Clearly some people find
1543 these options useful.
1544
1545 17. If a caller to the POSIX matching function regexec() passes a non-zero
1546 value for nmatch with a NULL value for pmatch, the value of
1547 nmatch is forced to zero.
1548
1549 18. RunGrepTest did not have a test for the availability of the -u option of
1550 the diff command, as RunTest does. It now checks in the same way as
1551 RunTest, and also checks for the -b option.
1552
1553 19. If an odd number of negated classes containing just a single character
1554 interposed, within parentheses, between a forward reference to a named
1555 subpattern and the definition of the subpattern, compilation crashed with
1556 an internal error, complaining that it could not find the referenced
1557 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1558 [The bug was that it was starting one character too far in when skipping
1559 over the character class, thus treating the ] as data rather than
1560 terminating the class. This meant it could skip too much.]
1561
1562 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1563 /g option in pcretest when the pattern contains \K, which makes it possible
1564 to have an empty string match not at the start, even when the pattern is
1565 anchored. Updated pcretest and pcredemo to use this option.
1566
1567 21. If the maximum number of capturing subpatterns in a recursion was greater
1568 than the maximum at the outer level, the higher number was returned, but
1569 with unset values at the outer level. The correct (outer level) value is
1570 now given.
1571
1572 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1573 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1574 make it do so. The string so far is captured, making this feature
1575 compatible with Perl.
1576
1577 23. The tests have been re-organized, adding tests 11 and 12, to make it
1578 possible to check the Perl 5.10 features against Perl 5.10.
1579
1580 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1581 pattern matches a fixed length string. PCRE did not allow this; now it
1582 does. Neither allows recursion.
1583
1584 25. I finally figured out how to implement a request to provide the minimum
1585 length of subject string that was needed in order to match a given pattern.
1586 (It was back references and recursion that I had previously got hung up
1587 on.) This code has now been added to pcre_study(); it finds a lower bound
1588 to the length of subject needed. It is not necessarily the greatest lower
1589 bound, but using it to avoid searching strings that are too short does give
1590 some useful speed-ups. The value is available to calling programs via
1591 pcre_fullinfo().
1592
1593 26. While implementing 25, I discovered to my embarrassment that pcretest had
1594 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1595 study optimizations had never been tested with that matching function.
1596 Oops. What is worse, even when it was passed study data, there was a bug in
1597 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1598 were also very few tests of studied patterns with pcre_dfa_exec().
1599
1600 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1601 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1602 on the other side of the coin, they are no longer allowed to have different
1603 names, because these cannot be distinguished in PCRE, and this has caused
1604 confusion. (This is a difference from Perl.)
1605
1606 28. When duplicate subpattern names are present (necessarily with different
1607 numbers, as required by 27 above), and a test is made by name in a
1608 conditional pattern, either for a subpattern having been matched, or for
1609 recursion in such a pattern, all the associated numbered subpatterns are
1610 tested, and the overall condition is true if the condition is true for any
1611 one of them. This is the way Perl works, and is also more like the way
1612 testing by number works.
1613
1614
1615 Version 7.9 11-Apr-09
1616 ---------------------
1617
1618 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1619 (pcretest), all targets were linked against these libraries. This included
1620 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1621 libraries. This caused unwanted dependencies to be created. This problem
1622 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1623 pcretest is linked with readline.
1624
1625 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1626 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1627 moved outside it again, because FALSE and TRUE are already defined in AIX,
1628 but BOOL is not.
1629
1630 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1631 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1632
1633 4. The pcregrep documentation said spaces were inserted as well as colons (or
1634 hyphens) following file names and line numbers when outputting matching
1635 lines. This is not true; no spaces are inserted. I have also clarified the
1636 wording for the --colour (or --color) option.
1637
1638 5. In pcregrep, when --colour was used with -o, the list of matching strings
1639 was not coloured; this is different to GNU grep, so I have changed it to be
1640 the same.
1641
1642 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1643 each matching line was coloured. Now it goes on to look for further matches
1644 of any of the test patterns, which is the same behaviour as GNU grep.
1645
1646 7. A pattern that could match an empty string could cause pcregrep to loop; it
1647 doesn't make sense to accept an empty string match in pcregrep, so I have
1648 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1649 seems to be how GNU grep behaves.
1650
1651 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1652 start or after a newline", because the conditional assertion was not being
1653 correctly handled. The rule now is that both the assertion and what follows
1654 in the first alternative must satisfy the test.
1655
1656 9. If auto-callout was enabled in a pattern with a conditional group whose
1657 condition was an assertion, PCRE could crash during matching, both with
1658 pcre_exec() and pcre_dfa_exec().
1659
1660 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1661 used for matching.
1662
1663 11. Unicode property support in character classes was not working for
1664 characters (bytes) greater than 127 when not in UTF-8 mode.
1665
1666 12. Added the -M command line option to pcretest.
1667
1668 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1669
1670 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1671
1672 16. Added comments and documentation about mis-use of no_arg in the C++
1673 wrapper.
1674
1675 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1676 from Martin Jerabek that uses macro names for all relevant character and
1677 string constants.
1678
1679 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1680 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1681 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1682 these, but not everybody uses configure.
1683
1684 19. A conditional group that had only one branch was not being correctly
1685 recognized as an item that could match an empty string. This meant that an
1686 enclosing group might also not be so recognized, causing infinite looping
1687 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1688 with the subject "ab", where knowledge that the repeated group can match
1689 nothing is needed in order to break the loop.
1690
1691 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1692 exec(), but without supplying a callout function, matching went wrong.
1693
1694 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1695 leak if the size of the offset vector was greater than 30. When the vector
1696 is smaller, the saved offsets during recursion go onto a local stack
1697 vector, but for larger vectors malloc() is used. It was failing to free
1698 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1699 error, in fact).
1700
1701 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1702 heapframe that is used only when UTF-8 support is enabled. This caused no
1703 problem, but was untidy.
1704
1705 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1706 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1707 included within another project.
1708
1709 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1710 slightly modified by me:
1711
1712 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1713 not building pcregrep.
1714
1715 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1716 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1717
1718 25. Forward references, both numeric and by name, in patterns that made use of
1719 duplicate group numbers, could behave incorrectly or give incorrect errors,
1720 because when scanning forward to find the reference group, PCRE was not
1721 taking into account the duplicate group numbers. A pattern such as
1722 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1723
1724 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1725 the feature of a custom pointer more persuasive (as requested by a user).
1726
1727 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1728 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1729 pre-defined.
1730
1731 28. Added support for (*UTF8) at the start of a pattern.
1732
1733 29. Arrange for flags added by the "release type" setting in CMake to be shown
1734 in the configuration summary.
1735
1736
1737 Version 7.8 05-Sep-08
1738 ---------------------
1739
1740 1. Replaced UCP searching code with optimized version as implemented for Ad
1741 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1742 stage table and inline lookup instead of a function, giving speed ups of 2
1743 to 5 times on some simple patterns that I tested. Permission was given to
1744 distribute the MultiStage2.py script that generates the tables (it's not in
1745 the tarball, but is in the Subversion repository).
1746
1747 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1748 scripts.
1749
1750 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1751 a group with a zero qualifier. The result of the study could be incorrect,
1752 or the function might crash, depending on the pattern.
1753
1754 4. Caseless matching was not working for non-ASCII characters in back
1755 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1756 It now works when Unicode Property Support is available.
1757
1758 5. In pcretest, an escape such as \x{de} in the data was always generating
1759 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1760 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1761 truncation.
1762
1763 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1764
1765 7. Added two (int) casts to pcregrep when printing the difference of two
1766 pointers, in case they are 64-bit values.
1767
1768 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1769 test 2 if it fails.
1770
1771 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1772 and a #define of that name to empty if it is not externally set. This is to
1773 allow users of MSVC to set it if necessary.
1774
1775 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1776 the convenience functions in the pcre_get.c source file.
1777
1778 11. An option change at the start of a pattern that had top-level alternatives
1779 could cause overwriting and/or a crash. This command provoked a crash in
1780 some environments:
1781
1782 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1783
1784 This potential security problem was recorded as CVE-2008-2371.
1785
1786 12. For a pattern where the match had to start at the beginning or immediately
1787 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1788 pcre_dfa_exec() could read past the end of the passed subject if there was
1789 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1790 pcretest so that it places the subject at the end of its malloc-ed buffer.
1791
1792 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1793 exec() might read past the end of the data buffer in UTF-8 mode.
1794
1795 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1796 the data contained the byte 0x85 as part of a UTF-8 character within its
1797 first line. This applied both to normal and DFA matching.
1798
1799 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1800 /^[^d]*?$/8 failed to match "abc".
1801
1802 16. Added a missing copyright notice to pcrecpp_internal.h.
1803
1804 17. Make it more clear in the documentation that values returned from
1805 pcre_exec() in ovector are byte offsets, not character counts.
1806
1807 18. Tidied a few places to stop certain compilers from issuing warnings.
1808
1809 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1810 supplied by Stefan Weber. I made a further small update for 7.8 because
1811 there is a change of source arrangements: the pcre_searchfuncs.c module is
1812 replaced by pcre_ucd.c.
1813
1814
1815 Version 7.7 07-May-08
1816 ---------------------
1817
1818 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1819 a string to a long long, pretend we don't even have a long long." This is
1820 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1821
1822 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1823 pre-7.6 versions, which defined a global no_arg variable instead of putting
1824 it in the RE class. (See also #8 below.)
1825
1826 3. Remove a line of dead code, identified by coverity and reported by Nuno
1827 Lopes.
1828
1829 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1830
1831 (1) The include/exclude patterns were being applied to the whole pathnames
1832 of files, instead of just to the final components.
1833
1834 (2) If there was more than one level of directory, the subdirectories were
1835 skipped unless they satisfied the include/exclude conditions. This is
1836 inconsistent with GNU grep (and could even be seen as contrary to the
1837 pcregrep specification - which I improved to make it absolutely clear).
1838 The action now is always to scan all levels of directory, and just
1839 apply the include/exclude patterns to regular files.
1840
1841 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1842 --exclude_dir in the tests to avoid scanning .svn directories.
1843
1844 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1845 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1846 doesn't support NULs in patterns.
1847
1848 7. Added some missing "const"s to declarations of static tables in
1849 pcre_compile.c and pcre_dfa_exec.c.
1850
1851 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1852 caused by fix #2 above. (Subsequently also a second patch to fix the
1853 first patch. And a third patch - this was a messy problem.)
1854
1855 9. Applied Craig's patch to remove the use of push_back().
1856
1857 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1858 matching function regexec().
1859
1860 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1861 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1862 references. PCRE supports relative numbers with this syntax (I don't think
1863 Oniguruma does).
1864
1865 12. Previously, a group with a zero repeat such as (...){0} was completely
1866 omitted from the compiled regex. However, this means that if the group
1867 was called as a subroutine from elsewhere in the pattern, things went wrong
1868 (an internal error was given). Such groups are now left in the compiled
1869 pattern, with a new opcode that causes them to be skipped at execution
1870 time.
1871
1872 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1873 to the way PCRE behaves:
1874
1875 (a) A lone ] character is dis-allowed (Perl treats it as data).
1876
1877 (b) A back reference to an unmatched subpattern matches an empty string
1878 (Perl fails the current match path).
1879
1880 (c) A data ] in a character class must be notated as \] because if the
1881 first data character in a class is ], it defines an empty class. (In
1882 Perl it is not possible to have an empty class.) The empty class []
1883 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1884 The negative empty class [^] matches any one character, independently
1885 of the DOTALL setting.
1886
1887 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1888 non-existent subpattern following a character class starting with ']' and
1889 containing () gave an internal compiling error instead of "reference to
1890 non-existent subpattern". Fortunately, when the pattern did exist, the
1891 compiled code was correct. (When scanning forwards to check for the
1892 existencd of the subpattern, it was treating the data ']' as terminating
1893 the class, so got the count wrong. When actually compiling, the reference
1894 was subsequently set up correctly.)
1895
1896 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1897 it was being rejected as not supported by pcre_dfa_exec(), even though
1898 other assertions are supported. I have made pcre_dfa_exec() support
1899 (*FAIL).
1900
1901 16. The implementation of 13c above involved the invention of a new opcode,
1902 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1903 cannot be changed at match time, I realized I could make a small
1904 improvement to matching performance by compiling OP_ALLANY instead of
1905 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1906 on the OP_ANY path.
1907
1908 17. Compiling pcretest on Windows with readline support failed without the
1909 following two fixes: (1) Make the unistd.h include conditional on
1910 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1911
1912 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1913 ncurses library to be included for pcretest when ReadLine support is
1914 requested, but also to allow for it to be overridden. This patch came from
1915 Daniel Bergström.
1916
1917 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1918 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1919 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1920 spotting this.
1921
1922
1923 Version 7.6 28-Jan-08
1924 ---------------------
1925
1926 1. A character class containing a very large number of characters with
1927 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1928 overflow.
1929
1930 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1931 HAVE_LONG_LONG is not defined.
1932
1933 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1934 bring them up to date and include new features. This patch includes:
1935
1936 - Fixed PH's badly added libz and libbz2 support.
1937 - Fixed a problem with static linking.
1938 - Added pcredemo. [But later removed - see 7 below.]
1939 - Fixed dftables problem and added an option.
1940 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1941 HAVE_LONG_LONG.
1942 - Added readline support for pcretest.
1943 - Added an listing of the option settings after cmake has run.
1944
1945 4. A user submitted a patch to Makefile that makes it easy to create
1946 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1947 Makefile.am that cause it to include this special target, without
1948 affecting anything else. Note that the same mingw target plus all
1949 the other distribution libraries and programs are now supported
1950 when configuring with CMake (see 6 below) instead of with
1951 Configure/Make.
1952
1953 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1954 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1955 exported in the Windows port". It has not yet been confirmed that the patch
1956 solves the problem, but it does no harm.
1957
1958 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1959 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1960 with CMake, and also correct the comment about stack recursion.
1961
1962 7. Remove the automatic building of pcredemo from the ./configure system and
1963 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1964 of a program that users should build themselves after PCRE is installed, so
1965 building it automatically is not really right. What is more, it gave
1966 trouble in some build environments.
1967
1968 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1969
1970
1971 Version 7.5 10-Jan-08
1972 ---------------------
1973
1974 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1975 values in parens when parsing an RE using the C++ wrapper."
1976
1977 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1978 Characters greater than 255 were excluded from the class instead of being
1979 included.
1980
1981 3. The same bug as (2) above applied to negated POSIX classes such as
1982 [:^space:].
1983
1984 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1985 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1986 I have changed it.
1987
1988 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1989 first named subpattern) and a construct such as (?&a) would reference the
1990 first named subpattern whose name started with "a" (in other words, the
1991 length check was missing). Both these problems are fixed. "Subpattern name
1992 expected" is now given for (?&) (a zero-length name), and this patch also
1993 makes it give the same error for \k'' (previously it complained that that
1994 was a reference to a non-existent subpattern).
1995
1996 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1997 this is right because (?- can be followed by option settings as well as by
1998 digits. I have, however, made the messages clearer.
1999
2000 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2001 than the number used in the conditional) now cause a compile-time error.
2002 This is actually not compatible with Perl, which accepts such patterns, but
2003 treats the conditional as always being FALSE (as PCRE used to), but it
2004 seems to me that giving a diagnostic is better.
2005
2006 8. Change "alphameric" to the more common word "alphanumeric" in comments
2007 and messages.
2008
2009 9. Fix two occurrences of "backslash" in comments that should have been
2010 "backspace".
2011
2012 10. Remove two redundant lines of code that can never be obeyed (their function
2013 was moved elsewhere).
2014
2015 11. The program that makes PCRE's Unicode character property table had a bug
2016 which caused it to generate incorrect table entries for sequences of
2017 characters that have the same character type, but are in different scripts.
2018 It amalgamated them into a single range, with the script of the first of
2019 them. In other words, some characters were in the wrong script. There were
2020 thirteen such cases, affecting characters in the following ranges:
2021
2022 U+002b0 - U+002c1
2023 U+0060c - U+0060d
2024 U+0061e - U+00612
2025 U+0064b - U+0065e
2026 U+0074d - U+0076d
2027 U+01800 - U+01805
2028 U+01d00 - U+01d77
2029 U+01d9b - U+01dbf
2030 U+0200b - U+0200f
2031 U+030fc - U+030fe
2032 U+03260 - U+0327f
2033 U+0fb46 - U+0fbb1
2034 U+10450 - U+1049d
2035
2036 12. The -o option (show only the matching part of a line) for pcregrep was not
2037 compatible with GNU grep in that, if there was more than one match in a
2038 line, it showed only the first of them. It now behaves in the same way as
2039 GNU grep.
2040
2041 13. If the -o and -v options were combined for pcregrep, it printed a blank
2042 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2043 does the same. The return code can be used to tell if there were any
2044 non-matching lines.
2045
2046 14. Added --file-offsets and --line-offsets to pcregrep.
2047
2048 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2049 infinitely looping recursion. The bug was that positive lookaheads were not
2050 being skipped when checking for a possible empty match (negative lookaheads
2051 and both kinds of lookbehind were skipped).
2052
2053 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2054 inclusion of <windows.h> to before rather than after the definition of
2055 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2056
2057 17. Specifying a possessive quantifier with a specific limit for a Unicode
2058 character property caused pcre_compile() to compile bad code, which led at
2059 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2060 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2061 caused the error; without that there was no problem.
2062
2063 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2064
2065 19. Added --enable-pcretest-libreadline.
2066
2067 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2068 RE::GlobalReplace(). As a result, the number of replacements returned was
2069 double what it should be. I removed one of the increments, but Craig sent a
2070 later patch that removed the other one (the right fix) and added unit tests
2071 that check the return values (which was not done before).
2072
2073 21. Several CMake things:
2074
2075 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2076 the names libpcre and libpcreposix, not just pcre and pcreposix.
2077
2078 (2) The above change means that pcretest and pcregrep are now correctly
2079 linked with the newly-built libraries, not previously installed ones.
2080
2081 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2082
2083 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2084 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2085 UTF-8 newline character). The key issue is that the pattern starts .*;
2086 this means that the match must be either at the beginning, or after a
2087 newline. The bug was in the code for advancing after a failed match and
2088 checking that the new position followed a newline. It was not taking
2089 account of UTF-8 characters correctly.
2090
2091 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2092 character classes. PCRE was not treating the sequence [:...:] as a
2093 character class unless the ... were all letters. Perl, however, seems to
2094 allow any characters between [: and :], though of course it rejects as
2095 unknown any "names" that contain non-letters, because all the known class
2096 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2097 for example, whereas PCRE did not - it did not recognize a POSIX character
2098 class. This seemed a bit dangerous, so the code has been changed to be
2099 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2100 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2101 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2102 Perl does, and where it didn't before.
2103
2104 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2105 Windows environments %n is disabled by default.
2106
2107
2108 Version 7.4 21-Sep-07
2109 ---------------------
2110
2111 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2112 means that a class such as [\s] counted as "explicit reference to CR or
2113 LF". That isn't really right - the whole point of the change was to try to
2114 help when there was an actual mention of one of the two characters. So now
2115 the change happens only if \r or \n (or a literal CR or LF) character is
2116 encountered.
2117
2118 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2119 of both had grown to the point where there were only 3 bits left.
2120 Fortunately, there was spare space in the data structure, and so I have
2121 moved the internal flags into a new 16-bit field to free up more option
2122 bits.
2123
2124 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2125 but did not set the internal JCHANGED flag - either of these is enough to
2126 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2127 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2128 start sets both bits.
2129
2130 4. Added options (at build time, compile time, exec time) to change \R from
2131 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2132
2133 5. doc/pcresyntax.html was missing from the distribution.
2134
2135 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2136 compatibility, even though it is no longer used.
2137
2138 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2139 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2140 windows.h file is present (where different names are used). [This was
2141 reversed later after testing - see 16 below.]
2142
2143 8. Changed all #include <config.h> to #include "config.h". There were also
2144 some further <pcre.h> cases that I changed to "pcre.h".
2145
2146 9. When pcregrep was used with the --colour option, it missed the line ending
2147 sequence off the lines that it output.
2148
2149 10. It was pointed out to me that arrays of string pointers cause lots of
2150 relocations when a shared library is dynamically loaded. A technique of
2151 using a single long string with a table of offsets can drastically reduce
2152 these. I have refactored PCRE in four places to do this. The result is
2153 dramatic:
2154
2155 Originally: 290
2156 After changing UCP table: 187
2157 After changing error message table: 43
2158 After changing table of "verbs" 36
2159 After changing table of Posix names 22
2160
2161 Thanks to the folks working on Gregex for glib for this insight.
2162
2163 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2164 unicode-properties was also set.
2165
2166 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2167
2168 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2169 checked only for CRLF.
2170
2171 14. Added casts to pcretest.c to avoid compiler warnings.
2172
2173 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2174
2175 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2176 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2177 entirely. This removes changes made in 7 above.
2178
2179 17. The CMake files have been updated, and there is now more information about
2180 building with CMake in the NON-UNIX-USE document.
2181
2182
2183 Version 7.3 28-Aug-07
2184 ---------------------
2185
2186 1. In the rejigging of the build system that eventually resulted in 7.1, the
2187 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2188 brackets there is not right, since it causes compilers to look for an
2189 installed pcre.h, not the version that is in the source that is being
2190 compiled (which of course may be different). I have changed it back to:
2191
2192 #include "pcre.h"
2193
2194 I have a vague recollection that the change was concerned with compiling in
2195 different directories, but in the new build system, that is taken care of
2196 by the VPATH setting the Makefile.
2197
2198 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2199 when the subject happened to end in the byte 0x85 (e.g. if the last
2200 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2201 characters but of course it shouldn't be taken as a newline when it is part
2202 of another character. The bug was that, for an unlimited repeat of . in
2203 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2204 characters when looking for a newline.
2205
2206 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2207
2208 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2209 in debug output.
2210
2211 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2212 long printing in the pcrecpp unittest when running under MinGW.
2213
2214 6. ESC_K was left out of the EBCDIC table.
2215
2216 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2217 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2218 limit also applies to "virtual nesting" when a pattern is recursive, and in
2219 this case 1000 isn't so big. I have been able to remove this limit at the
2220 expense of backing off one optimization in certain circumstances. Normally,
2221 when pcre_exec() would call its internal match() function recursively and
2222 immediately return the result unconditionally, it uses a "tail recursion"
2223 feature to save stack. However, when a subpattern that can match an empty
2224 string has an unlimited repetition quantifier, it no longer makes this
2225 optimization. That gives it a stack frame in which to save the data for
2226 checking that an empty string has been matched. Previously this was taken
2227 from the 1000-entry workspace that had been reserved. So now there is no
2228 explicit limit, but more stack is used.
2229
2230 8. Applied Daniel's patches to solve problems with the import/export magic
2231 syntax that is required for Windows, and which was going wrong for the
2232 pcreposix and pcrecpp parts of the library. These were overlooked when this
2233 problem was solved for the main library.
2234
2235 9. There were some crude static tests to avoid integer overflow when computing
2236 the size of patterns that contain repeated groups with explicit upper
2237 limits. As the maximum quantifier is 65535, the maximum group length was
2238 set at 30,000 so that the product of these two numbers did not overflow a
2239 32-bit integer. However, it turns out that people want to use groups that
2240 are longer than 30,000 bytes (though not repeat them that many times).
2241 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2242 made it possible to implement the integer overflow checks in a much more
2243 dynamic way, which I have now done. The artificial limitation on group
2244 length has been removed - we now have only the limit on the total length of
2245 the compiled pattern, which depends on the LINK_SIZE setting.
2246
2247 10. Fixed a bug in the documentation for get/copy named substring when
2248 duplicate names are permitted. If none of the named substrings are set, the
2249 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2250 empty string.
2251
2252 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2253 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2254 because the ] is interpreted as the first data character and the
2255 terminating ] is not found. PCRE has been made compatible with Perl in this
2256 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2257 cause memory overwriting.
2258
2259 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2260 string has been matched (to stop an infinite loop). It was not recognizing
2261 a conditional subpattern that could match an empty string if that
2262 subpattern was within another subpattern. For example, it looped when
2263 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2264 condition was not nested. This bug has been fixed.
2265
2266 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2267 past the start of the subject in the presence of bytes with the top bit
2268 set, for example "\x8aBCD".
2269
2270 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2271 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2272
2273 14. Optimized (?!) to (*FAIL).
2274
2275 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2276 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2277 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2278 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2279 does: it's just the validity check that is more restrictive.
2280
2281 16. Inserted checks for integer overflows during escape sequence (backslash)
2282 processing, and also fixed erroneous offset values for syntax errors during
2283 backslash processing.
2284
2285 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2286 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2287
2288 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2289 caused an overrun.
2290
2291 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2292 something other than just ASCII characters) inside a group that had an
2293 unlimited repeat caused a loop at compile time (while checking to see
2294 whether the group could match an empty string).
2295
2296 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2297 [\P{Any}] did so. (Error in the code for printing property names.)
2298
2299 21. An orphan \E inside a character class could cause a crash.
2300
2301 22. A repeated capturing bracket such as (A)? could cause a wild memory
2302 reference during compilation.
2303
2304 23. There are several functions in pcre_compile() that scan along a compiled
2305 expression for various reasons (e.g. to see if it's fixed length for look
2306 behind). There were bugs in these functions when a repeated \p or \P was
2307 present in the pattern. These operators have additional parameters compared
2308 with \d, etc, and these were not being taken into account when moving along
2309 the compiled data. Specifically:
2310
2311 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2312 length.
2313
2314 (b) An item such as \pL+ within a repeated group could cause crashes or
2315 loops.
2316
2317 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2318 "reference to non-existent subpattern" error.
2319
2320 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2321
2322 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2323 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2324
2325 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2326
2327 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2328 character were causing crashes (broken optimization).
2329
2330 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2331 \p or \P) caused a compile-time loop.
2332
2333 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2334 break. For example, the unstudied pattern [\r\n]A does not match the string
2335 "\r\nA" because change 7.0/46 below moves the current point on by two
2336 characters after failing to match at the start. However, the pattern \nA
2337 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2338 the same is true. There doesn't seem any very clean way out of this, but
2339 what I have chosen to do makes the common cases work: PCRE now takes note
2340 of whether there can be an explicit match for \r or \n anywhere in the
2341 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2342 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2343 pattern has explicit CR or LF references.
2344
2345 29. Added (*CR) etc for changing newline setting at start of pattern.
2346
2347
2348 Version 7.2 19-Jun-07
2349 ---------------------
2350
2351 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2352 which is apparently normally available under Windows.
2353
2354 2. Re-jig the pcregrep tests with different newline settings in an attempt
2355 to make them independent of the local environment's newline setting.
2356
2357 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2358
2359 4. Some of the "internals" tests were previously cut out when the link size
2360 was not 2, because the output contained actual offsets. The recent new
2361 "Z" feature of pcretest means that these can be cut out, making the tests
2362 usable with all link sizes.
2363
2364 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2365 stack recursion. This gives a massive performance boost under BSD, but just
2366 a small improvement under Linux. However, it saves one field in the frame
2367 in all cases.
2368
2369 6. Added more features from the forthcoming Perl 5.10:
2370
2371 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2372 recursion call. It refers to the nth most recently opened parentheses.
2373
2374 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2375 to be opened parentheses.
2376
2377 (c) Conditions that refer to capturing parentheses can be specified
2378 relatively, for example, (?(-2)... or (?(+3)...
2379
2380 (d) \K resets the start of the current match so that everything before
2381 is not part of it.
2382
2383 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2384
2385 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2386 reference syntax.
2387
2388 (g) (?| introduces a group in which the numbering of parentheses in each
2389 alternative starts with the same number.
2390
2391 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2392
2393 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2394 PCRE_INFO_JCHANGED.
2395
2396 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2397 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2398 for detecting groups that can match an empty string.
2399
2400 9. A pattern with a very large number of alternatives (more than several
2401 hundred) was running out of internal workspace during the pre-compile
2402 phase, where pcre_compile() figures out how much memory will be needed. A
2403 bit of new cunning has reduced the workspace needed for groups with
2404 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2405 workspace instead of running out of the 4096 that are available.
2406
2407 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2408
2409 11. Applied patch from Google to remove an optimization that didn't quite work.
2410 The report of the bug said:
2411
2412 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2413 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2414 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2415
2416 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2417 it matched the wrong number of bytes.
2418
2419
2420 Version 7.1 24-Apr-07
2421 ---------------------
2422
2423 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2424 that is more "standard", making use of automake and other Autotools. There
2425 is some re-arrangement of the files and adjustment of comments consequent
2426 on this.
2427
2428 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2429 for recursive directory scanning broke on some systems because the files
2430 are not scanned in any specific order and on different systems the order
2431 was different. A call to "sort" has been inserted into RunGrepTest for the
2432 approprate test as a short-term fix. In the longer term there may be an
2433 alternative.
2434
2435 3. I had an email from Eric Raymond about problems translating some of PCRE's
2436 man pages to HTML (despite the fact that I distribute HTML pages, some
2437 people do their own conversions for various reasons). The problems
2438 concerned the use of low-level troff macros .br and .in. I have therefore
2439 removed all such uses from the man pages (some were redundant, some could
2440 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2441 HTML has been updated to handle .nf/.fi and to complain if it encounters
2442 .br or .in.
2443
2444 4. Updated comments in configure.ac that get placed in config.h.in and also
2445 arranged for config.h to be included in the distribution, with the name
2446 config.h.generic, for the benefit of those who have to compile without
2447 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2448
2449 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2450 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2451 makevp.bat for the current PCRE, using the additional files
2452 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2453
2454 6. A Windows user reported a minor discrepancy with test 2, which turned out
2455 to be caused by a trailing space on an input line that had got lost in his
2456 copy. The trailing space was an accident, so I've just removed it.
2457
2458 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2459 that is needed.
2460
2461 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2462 as "const" (a) because they are and (b) because it helps the PHP
2463 maintainers who have recently made a script to detect big data structures
2464 in the php code that should be moved to the .rodata section. I remembered
2465 to update Builducptable as well, so it won't revert if ucptable.h is ever
2466 re-created.
2467
2468 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2469 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2470 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2471 support is not required. This saves 1.5-2K of code, which is important in
2472 some applications.
2473
2474 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2475 so as not to refer to the tables, even though these functions will never be
2476 called when UTF-8 support is disabled. Otherwise there are problems with a
2477 shared library.
2478
2479 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2480
2481 (a) It was defining its arguments as char * instead of void *.
2482
2483 (b) It was assuming that all moves were upwards in memory; this was true
2484 a long time ago when I wrote it, but is no longer the case.
2485
2486 The emulated memove() is provided for those environments that have neither
2487 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2488 is clearly not the case, as these two bugs were recently reported.
2489
2490 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2491 and Detrail to create the HTML documentation, the .txt form of the man
2492 pages, and it removes trailing spaces from listed files. It also creates
2493 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2494 case, it wraps all the #defines with #ifndefs. This script should be run
2495 before "make dist".
2496
2497 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2498 with Unicode property support.
2499
2500 (a) For a maximizing quantifier, if the two different cases of the
2501 character were of different lengths in their UTF-8 codings (there are
2502 some cases like this - I found 11), and the matching function had to
2503 back up over a mixture of the two cases, it incorrectly assumed they
2504 were both the same length.
2505
2506 (b) When PCRE was configured to use the heap rather than the stack for
2507 recursion during matching, it was not correctly preserving the data for
2508 the other case of a UTF-8 character when checking ahead for a match
2509 while processing a minimizing repeat. If the check also involved
2510 matching a wide character, but failed, corruption could cause an
2511 erroneous result when trying to check for a repeat of the original
2512 character.
2513
2514 13. Some tidying changes to the testing mechanism:
2515
2516 (a) The RunTest script now detects the internal link size and whether there
2517 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2518 values substituted by "configure". (The RunGrepTest script already did
2519 this for UTF-8.) The configure.ac script no longer substitutes the
2520 relevant variables.
2521
2522 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2523 with length and offset values. This means that the output is different
2524 for different internal link sizes. Test 2 is skipped for link sizes
2525 other than 2 because of this, bypassing the problem. Unfortunately,
2526 there was also a test in test 3 (the locale tests) that used /B and
2527 failed for link sizes other than 2. Rather than cut the whole test out,
2528 I have added a new /Z option to pcretest that replaces the length and
2529 offset values with spaces. This is now used to make test 3 independent
2530 of link size. (Test 2 will be tidied up later.)
2531
2532 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2533 segmentation fault instead of returning the appropriate error message.
2534
2535 15. In multiline mode when the newline sequence was set to "any", the pattern
2536 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2537 This doesn't seem right; it now treats the CRLF combination as the line
2538 ending, and so does not match in that case. It's only a pattern such as ^$
2539 that would hit this one: something like ^ABC$ would have failed after \r
2540 and then tried again after \r\n.
2541
2542 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2543 in an attempt to make files that differ only in their line terminators
2544 compare equal. This works on Linux.
2545
2546 17. Under certain error circumstances pcregrep might try to free random memory
2547 as it exited. This is now fixed, thanks to valgrind.
2548
2549 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2550 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2551 was because its rules for how to advance for /g after matching an empty
2552 string at the end of a line did not allow for this case. They now check for
2553 it specially.
2554
2555 20. pcretest is supposed to handle patterns and data of any length, by
2556 extending its buffers when necessary. It was getting this wrong when the
2557 buffer for a data line had to be extended.
2558
2559 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2560 CRLF as a newline sequence.
2561
2562 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2563 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2564 I have nevertheless tidied it up.
2565
2566 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2567
2568 24. Added a man page for pcre-config.
2569
2570
2571 Version 7.0 19-Dec-06
2572 ---------------------
2573
2574 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2575 moving to gcc 4.1.1.
2576
2577 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2578 sys/time.h, which is documented as needed for this function. It doesn't
2579 seem to matter on Linux, but it showed up on some releases of OS X.
2580
2581 3. It seems that there are systems where bytes whose values are greater than
2582 127 match isprint() in the "C" locale. The "C" locale should be the
2583 default when a C program starts up. In most systems, only ASCII printing
2584 characters match isprint(). This difference caused the output from pcretest
2585 to vary, making some of the tests fail. I have changed pcretest so that:
2586
2587 (a) When it is outputting text in the compiled version of a pattern, bytes
2588 other than 32-126 are always shown as hex escapes.
2589
2590 (b) When it is outputting text that is a matched part of a subject string,
2591 it does the same, unless a different locale has been set for the match
2592 (using the /L modifier). In this case, it uses isprint() to decide.
2593
2594 4. Fixed a major bug that caused incorrect computation of the amount of memory
2595 required for a compiled pattern when options that changed within the
2596 pattern affected the logic of the preliminary scan that determines the
2597 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2598 that the computed length was too small. The symptoms of this bug were
2599 either the PCRE error "internal error: code overflow" from pcre_compile(),
2600 or a glibc crash with a message such as "pcretest: free(): invalid next
2601 size (fast)". Examples of patterns that provoked this bug (shown in
2602 pcretest format) are:
2603
2604 /(?-x: )/x
2605 /(?x)(?-x: \s*#\s*)/
2606 /((?i)[\x{c0}])/8
2607 /(?i:[\x{c0}])/8
2608
2609 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2610 is now done differently.
2611
2612 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2613 wrapper classes; (b) implement a new function in the C++ scanner that is
2614 more efficient than the old way of doing things because it avoids levels of
2615 recursion in the regex matching; (c) add a paragraph to the documentation
2616 for the FullMatch() function.
2617
2618 6. The escape sequence \n was being treated as whatever was defined as
2619 "newline". Not only was this contrary to the documentation, which states
2620 that \n is character 10 (hex 0A), but it also went horribly wrong when
2621 "newline" was defined as CRLF. This has been fixed.
2622
2623 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2624 was being set to -1 for the "end of line" case (supposedly a value that no
2625 character can have). Though this value is never used (the check for end of
2626 line is "zero bytes in current character"), it caused compiler complaints.
2627 I've changed it to 0xffffffff.
2628
2629 8. In pcre_version.c, the version string was being built by a sequence of
2630 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2631 string (as it is for production releases) called a macro with an empty
2632 argument. The C standard says the result of this is undefined. The gcc
2633 compiler treats it as an empty string (which was what was wanted) but it is
2634 reported that Visual C gives an error. The source has been hacked around to
2635 avoid this problem.
2636
2637 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2638 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2639 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2640 of them did).
2641
2642 10. Originally, pcretest opened its input and output without "b"; then I was
2643 told that "b" was needed in some environments, so it was added for release
2644 5.0 to both the input and output. (It makes no difference on Unix-like
2645 systems.) Later I was told that it is wrong for the input on Windows. I've
2646 now abstracted the modes into two macros, to make it easier to fiddle with
2647 them, and removed "b" from the input mode under Windows.
2648
2649 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2650
2651 12. Added -help and --help to pcretest as an official way of being reminded
2652 of the options.
2653
2654 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2655 and pcrecpp.cc because they annoy compilers at high warning levels.
2656
2657 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2658
2659 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2660 scripts are not C programs :-) and which was not noticed because it works
2661 on Linux.
2662
2663 16. pcretest is supposed to handle any length of pattern and data line (as one
2664 line or as a continued sequence of lines) by extending its input buffer if
2665 necessary. This feature was broken for very long pattern lines, leading to
2666 a string of junk being passed to pcre_compile() if the pattern was longer
2667 than about 50K.
2668
2669 17. I have done a major re-factoring of the way pcre_compile() computes the
2670 amount of memory needed for a compiled pattern. Previously, there was code
2671 that made a preliminary scan of the pattern in order to do this. That was
2672 OK when PCRE was new, but as the facilities have expanded, it has become
2673 harder and harder to keep it in step with the real compile phase, and there
2674 have been a number of bugs (see for example, 4 above). I have now found a
2675 cunning way of running the real compile function in a "fake" mode that
2676 enables it to compute how much memory it would need, while actually only
2677 ever using a few hundred bytes of working memory and without too many
2678 tests of the mode. This should make future maintenance and development
2679 easier. A side effect of this work is that the limit of 200 on the nesting
2680 depth of parentheses has been removed (though this was never a serious
2681 limitation, I suspect). However, there is a downside: pcre_compile() now
2682 runs more slowly than before (30% or more, depending on the pattern). I
2683 hope this isn't a big issue. There is no effect on runtime performance.
2684
2685 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2686 newline (only possible for the last line of a file) and it was a
2687 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2688
2689 19. Added additional timing features to pcretest. (1) The -tm option now times
2690 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2691 separate command line item, by a number that specifies the number of
2692 repeats to use when timing. The default is 50000; this gives better
2693 precision, but takes uncomfortably long for very large patterns.
2694
2695 20. Extended pcre_study() to be more clever in cases where a branch of a
2696 subpattern has no definite first character. For example, (a*|b*)[cd] would
2697 previously give no result from pcre_study(). Now it recognizes that the
2698 first character must be a, b, c, or d.
2699
2700 21. There was an incorrect error "recursive call could loop indefinitely" if
2701 a subpattern (or the entire pattern) that was being tested for matching an
2702 empty string contained only one non-empty item after a nested subpattern.
2703 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2704 incorrectly, because the \d was being skipped in the check.
2705
2706 22. The pcretest program now has a new pattern option /B and a command line
2707 option -b, which is equivalent to adding /B to every pattern. This causes
2708 it to show the compiled bytecode, without the additional information that
2709 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2710 is the same as /B/I).
2711
2712 23. A new optimization is now able automatically to treat some sequences such
2713 as a*b as a*+b. More specifically, if something simple (such as a character
2714 or a simple class like \d) has an unlimited quantifier, and is followed by
2715 something that cannot possibly match the quantified thing, the quantifier
2716 is automatically "possessified".
2717
2718 24. A recursive reference to a subpattern whose number was greater than 39
2719 went wrong under certain circumstances in UTF-8 mode. This bug could also
2720 have affected the operation of pcre_study().
2721
2722 25. Realized that a little bit of performance could be had by replacing
2723 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2724
2725 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2726
2727 27. Possessive quantifiers such as a++ were previously implemented by turning
2728 them into atomic groups such as ($>a+). Now they have their own opcodes,
2729 which improves performance. This includes the automatically created ones
2730 from 23 above.
2731
2732 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2733 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2734 the first matched character to be a colon. This applied both to named and
2735 numbered groups.
2736
2737 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2738
2739 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2740 building PCRE on Windows easier, so I have included it in the distribution.
2741
2742 31. There is now a check in pcretest against a ridiculously large number being
2743 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2744 loop, the loop is abandoned.
2745
2746 32. Forward references to subpatterns in conditions such as (?(2)...) where
2747 subpattern 2 is defined later cause pcre_compile() to search forwards in
2748 the pattern for the relevant set of parentheses. This search went wrong
2749 when there were unescaped parentheses in a character class, parentheses
2750 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2751
2752 33. "Subroutine" calls and backreferences were previously restricted to
2753 referencing subpatterns earlier in the regex. This restriction has now
2754 been removed.
2755
2756 34. Added a number of extra features that are going to be in Perl 5.10. On the
2757 whole, these are just syntactic alternatives for features that PCRE had
2758 previously implemented using the Python syntax or my own invention. The
2759 other formats are all retained for compatibility.
2760
2761 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2762 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2763 also .NET compatible.
2764
2765 (b) A recursion or subroutine call to a named group can now be defined as
2766 (?&name) as well as (?P>name).
2767
2768 (c) A backreference to a named group can now be defined as \k<name> or
2769 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2770 5.10, are also .NET compatible.
2771
2772 (d) A conditional reference to a named group can now use the syntax
2773 (?(<name>) or (?('name') as well as (?(name).
2774
2775 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2776 groups (named and numbered) that are never evaluated inline, but can be
2777 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2778 is always false. There may be only one alternative in such a group.
2779
2780 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2781 as the simple (?(R). The condition is true only if the most recent
2782 recursion is that of the given number or name. It does not search out
2783 through the entire recursion stack.
2784
2785 (g) The escape \gN or \g{N} has been added, where N is a positive or
2786 negative number, specifying an absolute or relative reference.
2787
2788 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2789 some "unreachable code" warnings.
2790
2791 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2792 things, this adds five new scripts.
2793
2794 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2795 There were also incompatibilities regarding the handling of \Q..\E inside
2796 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2797 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2798
2799 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2800 matches an empty string, and forcibly breaks the loop. There were bugs in
2801 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2802 against aaaa the result was just "a" rather than "aaaa", for example. Two
2803 separate and independent bugs (that affected different cases) have been
2804 fixed.
2805
2806 39. Refactored the code to abolish the use of different opcodes for small
2807 capturing bracket numbers. This is a tidy that I avoided doing when I
2808 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2809 The new approach is not only tidier, it makes it possible to reduce the
2810 memory needed to fix the previous bug (38).
2811
2812 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2813 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2814 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2815 mode.
2816
2817 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2818 report.
2819
2820 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2821 copy construction and assignment in the C++ wrapper.
2822
2823 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2824 couple of bugs that could have given wrong results in the "--newline=crlf"
2825 case.
2826
2827 44. Added a number of casts and did some reorganization of signed/unsigned int
2828 variables following suggestions from Dair Grant. Also renamed the variable
2829 "this" as "item" because it is a C++ keyword.
2830
2831 45. Arranged for dftables to add
2832
2833 #include "pcre_internal.h"
2834
2835 to pcre_chartables.c because without it, gcc 4.x may remove the array
2836 definition from the final binary if PCRE is built into a static library and
2837 dead code stripping is activated.
2838
2839 46. For an unanchored pattern, if a match attempt fails at the start of a
2840 newline sequence, and the newline setting is CRLF or ANY, and the next two
2841 characters are CRLF, advance by two characters instead of one.
2842
2843
2844 Version 6.7 04-Jul-06
2845 ---------------------
2846
2847 1. In order to handle tests when input lines are enormously long, pcretest has
2848 been re-factored so that it automatically extends its buffers when
2849 necessary. The code is crude, but this _is_ just a test program. The
2850 default size has been increased from 32K to 50K.
2851
2852 2. The code in pcre_study() was using the value of the re argument before
2853 testing it for NULL. (Of course, in any sensible call of the function, it
2854 won't be NULL.)
2855
2856 3. The memmove() emulation function in pcre_internal.h, which is used on
2857 systems that lack both memmove() and bcopy() - that is, hardly ever -
2858 was missing a "static" storage class specifier.
2859
2860 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2861 containing an extended class (one that cannot be represented by a bitmap
2862 because it contains high-valued characters or Unicode property items, e.g.
2863 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2864 pattern, but PCRE should not loop if you do not (it no longer does).
2865 [Detail: two cases were found: (a) a repeated subpattern containing an
2866 extended class; (b) a recursive reference to a subpattern that followed a
2867 previous extended class. It wasn't skipping over the extended class
2868 correctly when UTF-8 mode was not set.]
2869
2870 5. A negated single-character class was not being recognized as fixed-length
2871 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2872 compile error "lookbehind assertion is not fixed length".
2873
2874 6. The RunPerlTest auxiliary script was showing an unexpected difference
2875 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2876 write a Perl script that can interpret lines of an input file either as
2877 byte characters or as UTF-8, which is what "perltest" was being required to
2878 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2879 can't do is switch easily at run time between having the "use utf8;" pragma
2880 or not. In the end, I fudged it by using the RunPerlTest script to insert
2881 "use utf8;" explicitly for the UTF-8 tests.
2882
2883 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2884 the end of the subject string, contrary to the documentation and to what
2885 Perl does. This was true of both matching functions. Now it matches only at
2886 the start of the subject and immediately after *internal* newlines.
2887
2888 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2889 a pointer to an int instead of a pointer to an unsigned long int. This
2890 caused problems on 64-bit systems.
2891
2892 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2893 instance of the 'standard' template library not being so standard".
2894
2895 10. There was no check on the number of named subpatterns nor the maximum
2896 length of a subpattern name. The product of these values is used to compute
2897 the size of the memory block for a compiled pattern. By supplying a very
2898 long subpattern name and a large number of named subpatterns, the size
2899 computation could be caused to overflow. This is now prevented by limiting
2900 the length of names to 32 characters, and the number of named subpatterns
2901 to 10,000.
2902
2903 11. Subpatterns that are repeated with specific counts have to be replicated in
2904 the compiled pattern. The size of memory for this was computed from the
2905 length of the subpattern and the repeat count. The latter is limited to
2906 65535, but there was no limit on the former, meaning that integer overflow
2907 could in principle occur. The compiled length of a repeated subpattern is
2908 now limited to 30,000 bytes in order to prevent this.
2909
2910 12. Added the optional facility to have named substrings with the same name.
2911
2912 13. Added the ability to use a named substring as a condition, using the
2913 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2914 are numbers (not recommended). Forward references are permitted.
2915
2916 14. Added forward references in named backreferences (if you see what I mean).
2917
2918 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2919 pattern could run off the end of the subject. For example, the pattern
2920 "(?s)(.{1,5})"8 did this with the subject "ab".
2921
2922 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2923 PCRE_CASELESS was set when matching characters that were quantified with ?
2924 or *.
2925
2926 17. A character class other than a single negated character that had a minimum
2927 but no maximum quantifier - for example [ab]{6,} - was not handled
2928 correctly by pce_dfa_exec(). It would match only one character.
2929
2930 18. A valid (though odd) pattern that looked like a POSIX character
2931 class but used an invalid character after [ (for example [[,abc,]]) caused
2932 pcre_compile() to give the error "Failed: internal error: code overflow" or
2933 in some cases to crash with a glibc free() error. This could even happen if
2934 the pattern terminated after [[ but there just happened to be a sequence of
2935 letters, a binary zero, and a closing ] in the memory that followed.
2936
2937 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2938 over the years. Originally (before any Unicode support), just the bottom 8
2939 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2940 output from "man perlunicode" includes this:
2941
2942 The regular expression compiler produces polymorphic opcodes. That
2943 is, the pattern adapts to the data and automatically switches to
2944 the Unicode character scheme when presented with Unicode data--or
2945 instead uses a traditional byte scheme when presented with byte
2946 data.
2947
2948 Sadly, a wide octal escape does not cause a switch, and in a string with
2949 no other multibyte characters, these octal escapes are treated as before.
2950 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2951 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2952 Unicode string.
2953
2954 I have not perpetrated such confusion in PCRE. Up till now, it took just
2955 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2956 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2957 translate to the appropriate multibyte character.
2958
2959 29. Applied some refactoring to reduce the number of warnings from Microsoft
2960 and Borland compilers. This has included removing the fudge introduced
2961 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2962 a warning about an unused variable.
2963
2964 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2965 characters since release 4.0, because Perl (from release 5.004) does not.
2966 [Or at least, is documented not to: some releases seem to be in conflict
2967 with the documentation.] However, when a pattern was studied with
2968 pcre_study() and all its branches started with \s, PCRE still included VT
2969 as a possible starting character. Of course, this did no harm; it just
2970 caused an unnecessary match attempt.
2971
2972 22. Removed a now-redundant internal flag bit that recorded the fact that case
2973 dependency changed within the pattern. This was once needed for "required
2974 byte" processing, but is no longer used. This recovers a now-scarce options
2975 bit. Also moved the least significant internal flag bit to the most-
2976 significant bit of the word, which was not previously used (hangover from
2977 the days when it was an int rather than a uint) to free up another bit for
2978 the future.
2979
2980 23. Added support for CRLF line endings as well as CR and LF. As well as the
2981 default being selectable at build time, it can now be changed at runtime
2982 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2983 specify that it is scanning data with non-default line endings.
2984
2985 24. Changed the definition of CXXLINK to make it agree with the definition of
2986 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2987
2988 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2989 recursions. This makes a big different to stack usage for some patterns.
2990
2991 26. If a subpattern containing a named recursion or subroutine reference such
2992 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2993 the space required for the compiled pattern went wrong and gave too small a
2994 value. Depending on the environment, this could lead to "Failed: internal
2995 error: code overflow at offset 49" or "glibc detected double free or
2996 corruption" errors.
2997
2998 27. Applied patches from Google (a) to support the new newline modes and (b) to
2999 advance over multibyte UTF-8 characters in GlobalReplace.
3000
3001 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3002 difference for some implementation of PCRE in some Windows version.
3003
3004 29. Added some extra testing facilities to pcretest:
3005
3006 \q<number> in a data line sets the "match limit" value
3007 \Q<number> in a data line sets the "match recursion limt" value
3008 -S <number> sets the stack size, where <number> is in megabytes
3009
3010 The -S option isn't available for Windows.
3011
3012
3013 Version 6.6 06-Feb-06
3014 ---------------------
3015
3016 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3017 in pcreposix.h. I have copied the definition from pcre.h.
3018
3019 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3020 because pcre.h is no longer a built file.
3021
3022 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3023 not normally included in the compiled code.
3024
3025
3026 Version 6.5 01-Feb-06
3027 ---------------------
3028
3029 1. When using the partial match feature with pcre_dfa_exec(), it was not
3030 anchoring the second and subsequent partial matches at the new starting
3031 point. This could lead to incorrect results. For example, with the pattern
3032 /1234/, partially matching against "123" and then "a4" gave a match.
3033
3034 2. Changes to pcregrep:
3035
3036 (a) All non-match returns from pcre_exec() were being treated as failures
3037 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3038 error message is output. Some extra information is given for the
3039 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3040 probably the only errors that are likely to be caused by users (by
3041 specifying a regex that has nested indefinite repeats, for instance).
3042 If there are more than 20 of these errors, pcregrep is abandoned.
3043
3044 (b) A binary zero was treated as data while matching, but terminated the
3045 output line if it was written out. This has been fixed: binary zeroes
3046 are now no different to any other data bytes.
3047
3048 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3049 used to set a locale for matching. The --locale=xxxx long option has
3050 been added (no short equivalent) to specify a locale explicitly on the
3051 pcregrep command, overriding the environment variables.
3052
3053 (d) When -B was used with -n, some line numbers in the output were one less
3054 than they should have been.
3055
3056 (e) Added the -o (--only-matching) option.
3057
3058 (f) If -A or -C was used with -c (count only), some lines of context were
3059 accidentally printed for the final match.
3060
3061 (g) Added the -H (--with-filename) option.
3062
3063 (h) The combination of options -rh failed to suppress file names for files
3064 that were found from directory arguments.
3065
3066 (i) Added the -D (--devices) and -d (--directories) options.
3067
3068 (j) Added the -F (--fixed-strings) option.
3069
3070 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3071
3072 (l) Added the --colo(u)r option.
3073
3074 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3075 is not present by default.
3076
3077 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3078 items such as (?R) or (?1), when the recursion could match a number of
3079 alternatives. If it matched one of the alternatives, but subsequently,
3080 outside the recursion, there was a failure, the code tried to back up into
3081 the recursion. However, because of the way PCRE is implemented, this is not
3082 possible, and the result was an incorrect result from the match.
3083
3084 In order to prevent this happening, the specification of recursion has
3085 been changed so that all such subpatterns are automatically treated as
3086 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3087
3088 4. I had overlooked the fact that, in some locales, there are characters for
3089 which isalpha() is true but neither isupper() nor islower() are true. In
3090 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3091 and ordfeminine) are like this. This affected the treatment of \w and \W
3092 when they appeared in character classes, but not when they appeared outside
3093 a character class. The bit map for "word" characters is now created
3094 separately from the results of isalnum() instead of just taking it from the
3095 upper, lower, and digit maps. (Plus the underscore character, of course.)
3096
3097 5. The above bug also affected the handling of POSIX character classes such as
3098 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3099 permanent tables. Instead, the bit maps for such a class were previously
3100 created as the appropriate unions of the upper, lower, and digit bitmaps.
3101 Now they are created by subtraction from the [[:word:]] class, which has
3102 its own bitmap.
3103
3104 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3105 It is created by subtracting the vertical space characters (\x09, \x0a,
3106 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3107 subtraction was done in the overall bitmap for a character class, meaning
3108 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3109 be recognized. This bug has been fixed.
3110
3111 7. Patches from the folks at Google:
3112
3113 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3114 real life, but is still worth protecting against".
3115
3116 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3117 regular expressions".
3118
3119 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3120 have it.
3121
3122 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3123 "configure" and the latter not, in order to fix a problem somebody had
3124 with compiling the Arg class on HP-UX.
3125
3126 (e) Improve the error-handling of the C++ wrapper a little bit.
3127
3128 (f) New tests for checking recursion limiting.
3129
3130 8. The pcre_memmove() function, which is used only if the environment does not
3131 have a standard memmove() function (and is therefore rarely compiled),
3132 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3133 returning a result (though PCRE never actually uses the result).
3134
3135 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3136 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3137 returned instead of calling malloc() with an overflowing number that would
3138 most likely cause subsequent chaos.
3139
3140 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3141
3142 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3143 with this option is matched, the nmatch and pmatch options of regexec() are
3144 ignored.
3145
3146 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3147 provided in case anyone wants to the the POSIX interface with UTF-8
3148 strings.
3149
3150 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3151 C++ linking (needed for some HP-UX environments).
3152
3153 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3154 (unused parameter) and in the pcre_printint() function (omitted "default"
3155 switch label when the default is to do nothing).
3156
3157 15. Added some code to make it possible, when PCRE is compiled as a C++
3158 library, to replace subject pointers for pcre_exec() with a smart pointer
3159 class, thus making it possible to process discontinuous strings.
3160
3161 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3162 much the same function. They were added by different people who were trying
3163 to make PCRE easy to compile on non-Unix systems. It has been suggested
3164 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3165 for compiling on Windows systems. I have therefore replaced it with
3166 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3167 defaults to "extern" for C or "extern C" for C++, which works fine on
3168 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3169 SCOPE with something explicit in config.h. In addition:
3170
3171 (a) pcreposix.h still had just "extern" instead of either of these macros;
3172 I have replaced it with PCRE_DATA_SCOPE.
3173
3174 (b) Functions such as _pcre_xclass(), which are internal to the library,
3175 but external in the C sense, all had PCRE_EXPORT in their definitions.
3176 This is apparently wrong for the Windows case, so I have removed it.
3177 (It makes no difference on Unix-like systems.)
3178
3179 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3180 of recursive calls to match(). This is different to MATCH_LIMIT because
3181 that limits the total number of calls to match(), not all of which increase
3182 the depth of recursion. Limiting the recursion depth limits the amount of
3183 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3184 when PCRE is compiled, and changed at run time. A patch from Google adds
3185 this functionality to the C++ interface.
3186
3187 18. Changes to the handling of Unicode character properties:
3188
3189 (a) Updated the table to Unicode 4.1.0.
3190
3191 (b) Recognize characters that are not in the table as "Cn" (undefined).
3192
3193 (c) I revised the way the table is implemented to a much improved format
3194 which includes recognition of ranges. It now supports the ranges that
3195 are defined in UnicodeData.txt, and it also amalgamates other
3196 characters into ranges. This has reduced the number of entries in the
3197 table from around 16,000 to around 3,000, thus reducing its size
3198 considerably. I realized I did not need to use a tree structure after
3199 all - a binary chop search is just as efficient. Having reduced the
3200 number of entries, I extended their size from 6 bytes to 8 bytes to
3201 allow for more data.
3202
3203 (d) Added support for Unicode script names via properties such as \p{Han}.
3204
3205 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3206 matching that character.
3207
3208 20. When matching a repeated Unicode property with a minimum greater than zero,
3209 (for example \pL{2,}), PCRE could look past the end of the subject if it
3210 reached it while seeking the minimum number of characters. This could
3211 happen only if some of the characters were more than one byte long, because
3212 there is a check for at least the minimum number of bytes.
3213
3214 21. Refactored the implementation of \p and \P so as to be more general, to
3215 allow for more different types of property in future. This has changed the
3216 compiled form incompatibly. Anybody with saved compiled patterns that use
3217 \p or \P will have to recompile them.
3218
3219 22. Added "Any" and "L&" to the supported property types.
3220
3221 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3222 but give a compile time error if the value is greater than 0xff.
3223
3224 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3225 accidentally not being installed or uninstalled.
3226
3227 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3228 made were to insert the current release number. This seemed silly, because
3229 it made things harder for people building PCRE on systems that don't run
3230 "configure". I have turned pcre.h into a distributed file, no longer built
3231 by "configure", with the version identification directly included. There is
3232 no longer a pcre.h.in file.
3233
3234 However, this change necessitated a change to the pcre-config script as
3235 well. It is built from pcre-config.in, and one of the substitutions was the
3236 release number. I have updated configure.ac so that ./configure now finds
3237 the release number by grepping pcre.h.
3238
3239 26. Added the ability to run the tests under valgrind.
3240
3241
3242 Version 6.4 05-Sep-05
3243 ---------------------
3244
3245 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3246 "--" to be printed when multiple files were scanned, even when none of the
3247 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3248 consider it to be a bug, and have restored the previous behaviour.
3249
3250 2. A couple of code tidies to get rid of compiler warnings.
3251
3252 3. The pcretest program used to cheat by referring to symbols in the library
3253 whose names begin with _pcre_. These are internal symbols that are not
3254 really supposed to be visible externally, and in some environments it is
3255 possible to suppress them. The cheating is now confined to including
3256 certain files from the library's source, which is a bit cleaner.
3257
3258 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3259 file's purpose clearer.
3260
3261 5. Reorganized pcre_ucp_findchar().
3262
3263
3264 Version 6.3 15-Aug-05
3265 ---------------------
3266
3267 1. The file libpcre.pc.in did not have general read permission in the tarball.
3268
3269 2. There were some problems when building without C++ support:
3270
3271 (a) If C++ support was not built, "make install" and "make test" still
3272 tried to test it.
3273
3274 (b) There were problems when the value of CXX was explicitly set. Some
3275 changes have been made to try to fix these, and ...
3276
3277 (c) --disable-cpp can now be used to explicitly disable C++ support.
3278
3279 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3280 backslash in a target when C++ was disabled. This confuses some
3281 versions of "make", apparently. Using an intermediate variable solves
3282 this. (Same for CPP_LOBJ.)
3283
3284 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3285 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3286 necessary on certain architectures.
3287
3288 4. Added a setting of -export-symbols-regex to the link command to remove
3289 those symbols that are exported in the C sense, but actually are local
3290 within the library, and not documented. Their names all begin with
3291 "_pcre_". This is not a perfect job, because (a) we have to except some
3292 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3293 available (and never for static libraries). I have made a note to try to
3294 find a way round (a) in the future.
3295
3296
3297 Version 6.2 01-Aug-05
3298 ---------------------
3299
3300 1. There was no test for integer overflow of quantifier values. A construction
3301 such as {1111111111111111} would give undefined results. What is worse, if
3302 a minimum quantifier for a parenthesized subpattern overflowed and became
3303 negative, the calculation of the memory size went wrong. This could have
3304 led to memory overwriting.
3305
3306 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3307
3308 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3309 operating environments where this matters.
3310
3311 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3312 PCRE options from within the C++ wrapper.
3313
3314 5. Named capturing subpatterns were not being correctly counted when a pattern
3315 was compiled. This caused two problems: (a) If there were more than 100
3316 such subpatterns, the calculation of the memory needed for the whole
3317 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3318 back references of the form \12, where the number was greater than 9, were
3319 not recognized as back references, even though there were sufficient
3320 previous subpatterns.
3321
3322 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3323 versions of gcc, e.g. 2.95.4.
3324
3325
3326 Version 6.1 21-Jun-05
3327 ---------------------
3328
3329 1. There was one reference to the variable "posix" in pcretest.c that was not
3330 surrounded by "#if !defined NOPOSIX".
3331
3332 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3333 the cross-check on the old pcre_info() function, for the benefit of the
3334 cut-down version of PCRE that is currently imported into Exim.
3335
3336 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3337 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3338 patterns that start (?i) but I don't think that matters. The use of (?i) is
3339 just an example; this all applies to the other options as well.
3340
3341 4. Since libtool seems to echo the compile commands it is issuing, the output
3342 from "make" can be reduced a bit by putting "@" in front of each libtool
3343 compile command.
3344
3345 5. Patch from the folks at Google for configure.in to be a bit more thorough
3346 in checking for a suitable C++ installation before trying to compile the
3347 C++ stuff. This should fix a reported problem when a compiler was present,
3348 but no suitable headers.
3349
3350 6. The man pages all had just "PCRE" as their title. I have changed them to
3351 be the relevant file name. I have also arranged that these names are
3352 retained in the file doc/pcre.txt, which is a concatenation in text format
3353 of all the man pages except the little individual ones for each function.
3354
3355 7. The NON-UNIX-USE file had not been updated for the different set of source
3356 files that come with release 6. I also added a few comments about the C++
3357 wrapper.
3358
3359
3360 Version 6.0 07-Jun-05
3361 ---------------------
3362
3363 1. Some minor internal re-organization to help with my DFA experiments.
3364
3365 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3366 didn't matter for the library itself when fully configured, but did matter
3367 when compiling without UCP support, or within Exim, where the ucp files are
3368 not imported.
3369
3370 3. Refactoring of the library code to split up the various functions into
3371 different source modules. The addition of the new DFA matching code (see
3372 below) to a single monolithic source would have made it really too
3373 unwieldy, quite apart from causing all the code to be include in a
3374 statically linked application, when only some functions are used. This is
3375 relevant even without the DFA addition now that patterns can be compiled in
3376 one application and matched in another.
3377
3378 The downside of splitting up is that there have to be some external
3379 functions and data tables that are used internally in different modules of
3380 the library but which are not part of the API. These have all had their
3381 names changed to start with "_pcre_" so that they are unlikely to clash
3382 with other external names.
3383
3384 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3385 a different (DFA) algorithm. Although it is slower than the original
3386 function, it does have some advantages for certain types of matching
3387 problem.
3388
3389 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3390 including restarting after a partial match.
3391
3392 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3393 defined when compiling for Windows was sent to me. I have put it into the
3394 code, though I have no means of testing or verifying it.
3395
3396 7. Added the pcre_refcount() auxiliary function.
3397
3398 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3399 match before or at the first newline in the subject string. In pcretest,
3400 the /f option on a pattern can be used to set this.
3401
3402 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3403 would behave wrongly. This has been present in PCRE since release 4.0.
3404
3405 10. A number of changes to the pcregrep command:
3406
3407 (a) Refactored how -x works; insert ^(...)$ instead of setting
3408 PCRE_ANCHORED and checking the length, in preparation for adding
3409 something similar for -w.
3410
3411 (b) Added the -w (match as a word) option.
3412
3413 (c) Refactored the way lines are read and buffered so as to have more
3414 than one at a time available.
3415
3416 (d) Implemented a pcregrep test script.
3417
3418 (e) Added the -M (multiline match) option. This allows patterns to match
3419 over several lines of the subject. The buffering ensures that at least
3420 8K, or the rest of the document (whichever is the shorter) is available
3421 for matching (and similarly the previous 8K for lookbehind assertions).
3422
3423 (f) Changed the --help output so that it now says
3424
3425 -w, --word-regex(p)
3426
3427 instead of two lines, one with "regex" and the other with "regexp"
3428 because that confused at least one person since the short forms are the
3429 same. (This required a bit of code, as the output is generated
3430 automatically from a table. It wasn't just a text change.)
3431
3432 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3433 option but starts with a hyphen. Could be a pattern or a path name
3434 starting with a hyphen, for instance.
3435
3436 (h) "-" can be given as a file name to represent stdin.
3437
3438 (i) When file names are being printed, "(standard input)" is used for
3439 the standard input, for compatibility with GNU grep. Previously
3440 "<stdin>" was used.
3441
3442 (j) The option --label=xxx can be used to supply a name to be used for
3443 stdin when file names are being printed. There is no short form.
3444
3445 (k) Re-factored the options decoding logic because we are going to add
3446 two more options that take data. Such options can now be given in four
3447 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3448
3449 (l) Added the -A, -B, and -C options for requesting that lines of context
3450 around matches be printed.
3451
3452 (m) Added the -L option to print the names of files that do not contain
3453 any matching lines, that is, the complement of -l.
3454
3455 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3456 continue to scan other files.
3457
3458 (o) The -s option was incorrectly implemented. For compatibility with other
3459 greps, it now suppresses the error message for a non-existent or non-
3460 accessible file (but not the return code). There is a new option called
3461 -q that suppresses the output of matching lines, which was what -s was
3462 previously doing.
3463
3464 (p) Added --include and --exclude options to specify files for inclusion
3465 and exclusion when recursing.
3466
3467 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3468 Hopefully, it now does.
3469
3470 12. Missing cast in pcre_study().
3471
3472 13. Added an "uninstall" target to the makefile.
3473
3474 14. Replaced "extern" in the function prototypes in Makefile.in with
3475 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3476 world, but is set differently for Windows.
3477
3478 15. Added a second compiling function called pcre_compile2(). The only
3479 difference is that it has an extra argument, which is a pointer to an
3480 integer error code. When there is a compile-time failure, this is set
3481 non-zero, in addition to the error test pointer being set to point to an
3482 error message. The new argument may be NULL if no error number is required
3483 (but then you may as well call pcre_compile(), which is now just a
3484 wrapper). This facility is provided because some applications need a
3485 numeric error indication, but it has also enabled me to tidy up the way
3486 compile-time errors are handled in the POSIX wrapper.
3487
3488 16. Added VPATH=.libs to the makefile; this should help when building with one
3489 prefix path and installing with another. (Or so I'm told by someone who
3490 knows more about this stuff than I do.)
3491
3492 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3493 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3494 match everything, including newlines. This is not POSIX-compatible, but
3495 somebody wanted the feature. From pcretest it can be activated by using
3496 both the P and the s flags.
3497
3498 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3499
3500 19. libpcre.pc was being incorrectly installed as executable.
3501
3502 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3503 it now also looks for '\r' so that it will work unmodified on Windows.
3504
3505 21. Added Google's contributed C++ wrapper to the distribution.
3506
3507 22. Added some untidy missing memory free() calls in pcretest, to keep
3508 Electric Fence happy when testing.
3509
3510
3511
3512 Version 5.0 13-Sep-04
3513 ---------------------
3514
3515 1. Internal change: literal characters are no longer packed up into items
3516 containing multiple characters in a single byte-string. Each character
3517 is now matched using a separate opcode. However, there may be more than one
3518 byte in the character in UTF-8 mode.
3519
3520 2. The pcre_callout_block structure has two new fields: pattern_position and
3521 next_item_length. These contain the offset in the pattern to the next match
3522 item, and its length, respectively.
3523
3524 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3525 insertion of callouts before each pattern item. Added the /C option to
3526 pcretest to make use of this.
3527
3528 4. On the advice of a Windows user, the lines
3529
3530 #if defined(_WIN32) || defined(WIN32)
3531 _setmode( _fileno( stdout ), 0x8000 );
3532 #endif /* defined(_WIN32) || defined(WIN32) */
3533
3534 have been added to the source of pcretest. This apparently does useful
3535 magic in relation to line terminators.
3536
3537 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3538 for the benefit of those environments where the "b" makes a difference.
3539
3540 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3541 to know about it. I have put a hack into configure.in that adds in code
3542 to set GCC=yes if CC=icc. This seems to end up at a point in the
3543 generated configure script that is early enough to affect the setting of
3544 compiler options, which is what is needed, but I have no means of testing
3545 whether it really works. (The user who reported this had patched the
3546 generated configure script, which of course I cannot do.)
3547
3548 LATER: After change 22 below (new libtool files), the configure script
3549 seems to know about icc (and also ecc). Therefore, I have commented out
3550 this hack in configure.in.
3551
3552 7. Added support for pkg-config (2 patches were sent in).
3553
3554 8. Negated POSIX character classes that used a combination of internal tables
3555 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3556 [[:^ascii]]. Typically, they would match almost any characters. The other
3557 POSIX classes were not broken in this way.
3558
3559 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3560 to find the match, as PCRE was deluded into thinking that the match had to
3561 start at the start point or following a newline. The same bug applied to
3562 patterns with negative forward assertions or any backward assertions
3563 preceding ".*" at the start, unless the pattern required a fixed first
3564 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3565
3566 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3567 starting at the last subject character, bytes beyond the end of the subject
3568 string were read.
3569
3570 11. Renamed the variable "class" as "classbits" to make life easier for C++
3571 users. (Previously there was a macro definition, but it apparently wasn't
3572 enough.)
3573
3574 12. Added the new field "tables" to the extra data so that tables can be passed
3575 in at exec time, or the internal tables can be re-selected. This allows
3576 a compiled regex to be saved and re-used at a later time by a different
3577 program that might have everything at different addresses.
3578
3579 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3580 -R library as well as a -L library.
3581
3582 14. The debugging options of pcretest (-d on the command line or D on a
3583 pattern) showed incorrect output for anything following an extended class
3584 that contained multibyte characters and which was followed by a quantifier.
3585
3586 15. Added optional support for general category Unicode character properties
3587 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3588 support. It adds about 90K to the size of the library. The meanings of the
3589 inbuilt class escapes such as \d and \s have NOT been changed.
3590
3591 16. Updated pcredemo.c to include calls to free() to release the memory for the
3592 compiled pattern.
3593
3594 17. The generated file chartables.c was being created in the source directory
3595 instead of in the building directory. This caused the build to fail if the
3596 source directory was different from the building directory, and was
3597 read-only.
3598
3599 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3600 file. No doubt somebody will tell me if they don't make sense... Also added
3601 Dan Mooney's comments about building on OpenVMS.
3602
3603 19. Added support for partial matching via the PCRE_PARTIAL option for
3604 pcre_exec() and the \P data escape in pcretest.
3605
3606 20. Extended pcretest with 3 new pattern features:
3607
3608 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3609 write the compiled pattern to the file whose name is "rest-of-line".
3610 This is a straight binary dump of the data, with the saved pointer to
3611 the character tables forced to be NULL. The study data, if any, is
3612 written too. After writing, pcretest reads a new pattern.
3613
3614 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3615 compiled pattern from the given file. There must not be any
3616 occurrences of "<" in the file name (pretty unlikely); if there are,
3617 pcretest will instead treat the initial "<" as a pattern delimiter.
3618 After reading in the pattern, pcretest goes on to read data lines as
3619 usual.
3620
3621 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3622 and 16-bit fields in a compiled pattern, to simulate a pattern that
3623 was compiled on a host of opposite endianness.
3624
3625 21. The pcre-exec() function can now cope with patterns that were compiled on
3626 hosts of opposite endianness, with this restriction:
3627
3628 As for any compiled expression that is saved and used later, the tables
3629 pointer field cannot be preserved; the extra_data field in the arguments
3630 to pcre_exec() should be used to pass in a tables address if a value
3631 other than the default internal tables were used at compile time.
3632
3633 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3634 now diagnosed as an error. Previously, most of the time, a negative number
3635 would have been treated as zero, but if in addition "ovector" was passed as
3636 NULL, a crash could occur.
3637
3638 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3639 new versions from the libtool 1.5 distribution (the last one is a copy of
3640 a file called libtool.m4). This seems to have fixed the need to patch
3641 "configure" to support Darwin 1.3 (which I used to do). However, I still
3642 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3643 workstation).
3644
3645 24. Changed the PCRE licence to be the more standard "BSD" licence.
3646
3647
3648 Version 4.5 01-Dec-03
3649 ---------------------
3650
3651 1. There has been some re-arrangement of the code for the match() function so
3652 that it can be compiled in a version that does not call itself recursively.
3653 Instead, it keeps those local variables that need separate instances for
3654 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3655 needs to "recurse". Keeping track of where control must go is done by means
3656 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3657 hide most of the details from the main code, and operates only if
3658 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3659 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3660 operating.
3661
3662 To make it easier for callers to provide specially tailored get/free
3663 functions for this usage, two new functions, pcre_stack_malloc, and
3664 pcre_stack_free, are used. They are always called in strict stacking order,
3665 and the size of block requested is always the same.
3666
3667 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3668 PCRE has been compiled to use the stack or the heap for recursion. The
3669 -C option of pcretest uses this to show which version is compiled.
3670
3671 A new data escape \S, is added to pcretest; it causes the amounts of store
3672 obtained and freed by both kinds of malloc/free at match time to be added
3673 to the output.
3674
3675 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3676 what's available on my current Linux desktop machine.
3677
3678 3. When matching a UTF-8 string, the test for a valid string at the start has
3679 been extended. If start_offset is not zero, PCRE now checks that it points
3680 to a byte that is the start of a UTF-8 character. If not, it returns
3681 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3682 this is necessary because there may be backward assertions in the pattern.
3683 When matching the same subject several times, it may save resources to use
3684 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3685
3686 4. The code for checking the validity of UTF-8 strings has been tightened so
3687 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3688 containing "overlong sequences".
3689
3690 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3691 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3692 should have been "&", but it just so happened that all the cases this let
3693 through by mistake were picked up later in the function.
3694
3695 6. I had used a variable called "isblank" - this is a C99 function, causing
3696 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3697
3698 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3699 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3700 so the version is shown; (c) stop "make test" echoing "./RunTest".
3701
3702 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3703
3704 9. The prototype for memmove() for systems that don't have it was using
3705 size_t, but the inclusion of the header that defines size_t was later. I've
3706 moved the #includes for the C headers earlier to avoid this.
3707
3708 10. Added some adjustments to the code to make it easier to compiler on certain
3709 special systems:
3710
3711 (a) Some "const" qualifiers were missing.
3712 (b) Added the macro EXPORT before all exported functions; by default this
3713 is defined to be empty.
3714 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3715 that it reads its output file name as an argument instead of writing
3716 to the standard output and assuming this can be redirected.
3717
3718 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3719 class containing characters with values greater than 255, PCRE compilation
3720 went into a loop.
3721
3722 12. A recursive reference to a subpattern that was within another subpattern
3723 that had a minimum quantifier of zero caused PCRE to crash. For example,
3724 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3725 recursion. If the recursively-called subpattern itself had a zero repeat,
3726 that was OK.
3727
3728 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3729 buffer into which it was copied (for escape processing) was still set at
3730 1024, so long lines caused crashes.
3731
3732 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3733 "internal error: code overflow...". This applied to any character class
3734 that was followed by a possessive quantifier.
3735
3736 15. Modified the Makefile to add libpcre.la as a prerequisite for
3737 libpcreposix.la because I was told this is needed for a parallel build to
3738 work.
3739
3740 16. If a pattern that contained .* following optional items at the start was
3741 studied, the wrong optimizing data was generated, leading to matching
3742 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3743 matching string must start with a or b or c. The correct conclusion for
3744 this pattern is that a match can start with any character.
3745
3746
3747 Version 4.4 13-Aug-03
3748 ---------------------
3749
3750 1. In UTF-8 mode, a character class containing characters with values between
3751 127 and 255 was not handled correctly if the compiled pattern was studied.
3752 In fixing this, I have also improved the studying algorithm for such
3753 classes (slightly).
3754
3755 2. Three internal functions had redundant arguments passed to them. Removal
3756 might give a very teeny performance improvement.
3757
3758 3. Documentation bug: the value of the capture_top field in a callout is *one
3759 more than* the number of the hightest numbered captured substring.
3760
3761 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3762 in incorrectly linking with a previously installed version. They now link
3763 explicitly with libpcre.la.
3764
3765 5. configure.in no longer needs to recognize Cygwin specially.
3766
3767 6. A problem in pcre.in for Windows platforms is fixed.
3768
3769 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3770 pcretest, it used to include the size of the study block as part of its
3771 output. Unfortunately, the structure contains a field that has a different
3772 size on different hardware architectures. This meant that the tests that
3773 showed this size failed. As the block is currently always of a fixed size,
3774 this information isn't actually particularly useful in pcretest output, so
3775 I have just removed it.
3776
3777 8. Three pre-processor statements accidentally did not start in column 1.
3778 Sadly, there are *still* compilers around that complain, even though
3779 standard C has not required this for well over a decade. Sigh.
3780
3781 9. In pcretest, the code for checking callouts passed small integers in the
3782 callout_data field, which is a void * field. However, some picky compilers
3783 complained about the casts involved for this on 64-bit systems. Now
3784 pcretest passes the address of the small integer instead, which should get
3785 rid of the warnings.
3786
3787 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3788 both compile and run time, and gives an error if an invalid UTF-8 sequence
3789 is found. There is a option for disabling this check in cases where the
3790 string is known to be correct and/or the maximum performance is wanted.
3791
3792 11. In response to a bug report, I changed one line in Makefile.in from
3793
3794 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3795 to
3796 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3797
3798 to look similar to other lines, but I have no way of telling whether this
3799 is the right thing to do, as I do not use Windows. No doubt I'll get told
3800 if it's wrong...
3801
3802
3803 Version 4.3 21-May-03
3804 ---------------------
3805
3806 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3807 Makefile.
3808
3809 2. Some refactoring to improve the quality of the code:
3810
3811 (i) The utf8_table... variables are now declared "const".
3812
3813 (ii) The code for \cx, which used the "case flipping" table to upper case
3814 lower case letters, now just substracts 32. This is ASCII-specific,
3815 but the whole concept of \cx is ASCII-specific, so it seems
3816 reasonable.
3817
3818 (iii) PCRE was using its character types table to recognize decimal and
3819 hexadecimal digits in the pattern. This is silly, because it handles
3820 only 0-9, a-f, and A-F, but the character types table is locale-
3821 specific, which means strange things might happen. A private
3822 table is now used for this - though it costs 256 bytes, a table is
3823 much faster than multiple explicit tests. Of course, the standard
3824 character types table is still used for matching digits in subject
3825 strings against \d.
3826
3827 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3828 ending in _t are). So I've renamed it as ESC_tee.
3829
3830 3. The first argument for regexec() in the POSIX wrapper should have been
3831 defined as "const".
3832
3833 4. Changed pcretest to use malloc() for its buffers so that they can be
3834 Electric Fenced for debugging.
3835
3836 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3837 to read one or more bytes before the start of the subject string. Often this
3838 had no effect on PCRE's behaviour, but in some circumstances it could
3839 provoke a segmentation fault.
3840
3841 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3842 to try to read one or more bytes before the start of the subject string.
3843
3844 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3845 UTF-8 support could misbehave in various ways if the subject string
3846 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3847 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3848 back over UTF-8 characters.)
3849
3850
3851 Version 4.2 14-Apr-03
3852 ---------------------
3853
3854 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3855
3856 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3857 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3858 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3859 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3860 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3861 and BUILD_EXEEXT
3862 Note: automatic setting of the BUILD variables is not yet working
3863 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3864 compile-time but not at link-time
3865 [LINK]: use for linking executables only
3866 make different versions for Windows and non-Windows
3867 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3868 libraries
3869 [LINK_FOR_BUILD]: new variable
3870 [OBJEXT]: use throughout
3871 [EXEEXT]: use throughout
3872 <winshared>: new target
3873 <wininstall>: new target
3874 <dftables.o>: use native compiler
3875 <dftables>: use native linker
3876 <install>: handle Windows platform correctly
3877 <clean>: ditto
3878 <check>: ditto
3879 copy DLL to top builddir before testing
3880
3881 As part of these changes, -no-undefined was removed again. This was reported
3882 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3883 in any case.
3884
3885 3. Some tidies to get rid of compiler warnings:
3886
3887 . In the match_data structure, match_limit was an unsigned long int, whereas
3888 match_call_count was an int. I've made them both unsigned long ints.
3889
3890 . In pcretest the fact that a const uschar * doesn't automatically cast to
3891 a void * provoked a warning.
3892
3893 . Turning on some more compiler warnings threw up some "shadow" variables
3894 and a few more missing casts.
3895
3896 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3897 option, a class that contained a single character with a value between 128
3898 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3899
3900 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3901 option, a class that contained several characters, but with at least one
3902 whose value was between 128 and 255 caused PCRE to crash.
3903
3904
3905 Version 4.1 12-Mar-03
3906 ---------------------
3907
3908 1. Compiling with gcc -pedantic found a couple of places where casts were
3909 needed, and a string in dftables.c that was longer than standard compilers are
3910 required to support.
3911
3912 2. Compiling with Sun's compiler found a few more places where the code could
3913 be tidied up in order to avoid warnings.
3914
3915 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3916 first of these names is deprecated in the latest Autoconf in favour of the name
3917 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3918 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3919 analogy I have changed it to CFLAGS_FOR_BUILD.
3920
3921 4. Added -no-undefined to the linking command in the Makefile, because this is
3922 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3923 linking step for the pcreposix library.
3924
3925 5. PCRE was failing to diagnose the case of two named groups with the same
3926 name.
3927
3928 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3929 literal character that is needed in the subject for a match, and scans along to
3930 ensure that it is present before embarking on the full matching process. This
3931 saves time in cases of nested unlimited repeats that are never going to match.
3932 Problem: the scan can take a lot of time if the subject is very long (e.g.
3933 megabytes), thus penalizing straightforward matches. It is now done only if the
3934 amount of subject to be scanned is less than 1000 bytes.
3935
3936 7. A lesser problem with the same optimization is that it was recording the
3937 first character of an anchored pattern as "needed", thus provoking a search
3938 right along the subject, even when the first match of the pattern was going to
3939 fail. The "needed" character is now not set for anchored patterns, unless it
3940 follows something in the pattern that is of non-fixed length. Thus, it still
3941 fulfils its original purpose of finding quick non-matches in cases of nested
3942 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3943
3944
3945 Version 4.0 17-Feb-03
3946 ---------------------
3947
3948 1. If a comment in an extended regex that started immediately after a meta-item
3949 extended to the end of string, PCRE compiled incorrect data. This could lead to
3950 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3951
3952 2. Moved to autoconf 2.53 and libtool 1.4.2.
3953
3954 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3955 the special perltest8 script is no longer needed - all the tests can be run
3956 from a single perltest script.
3957
3958 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3959 by \s. It has now been removed in PCRE. This means it isn't recognized as
3960 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3961 class [:space:] *does* include VT, thereby creating a mess.
3962
3963 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3964 space and tab.
3965
3966 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3967 its new features into the main test script, reducing the number of scripts.
3968
3969 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3970 were backward compatible, and made the (?i) apply to the whole pattern, as if
3971 /i were given. Now it behaves more logically, and applies the option setting
3972 only to what follows. PCRE has been changed to follow suit. However, if it
3973 finds options settings right at the start of the pattern, it extracts them into
3974 the global options, as before. Thus, they show up in the info data.
3975
3976 8. Added support for the \Q...\E escape sequence. Characters in between are
3977 treated as literals. This is slightly different from Perl in that $ and @ are
3978 also handled as literals inside the quotes. In Perl, they will cause variable
3979 interpolation. Note the following examples:
3980
3981 Pattern PCRE matches Perl matches
3982
3983 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3984 \Qabc\$xyz\E abc\$xyz abc\$xyz
3985 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3986
3987 For compatibility with Perl, \Q...\E sequences are recognized inside character
3988 classes as well as outside them.
3989
3990 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3991 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3992 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3993 signed/unsigned warnings.
3994
3995 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3996 option for pcretest, so I've replaced it by a simple function that does just
3997 that job.
3998
3999 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4000 "pcregrep -".
4001
4002 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4003 Java package. This provides some syntactic sugar for simple cases of what my
4004 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4005 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4006 item, you can use this simplified notation. Note that only makes sense with
4007 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4008 greediness, whatever the setting of the PCRE_UNGREEDY option.
4009
4010 13. A change of greediness default within a pattern was not taking effect at
4011 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4012 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4013 was abstracted outside.
4014
4015 14. PCRE now supports the \G assertion. It is true when the current matching
4016 position is at the start point of the match. This differs from \A when the
4017 starting offset is non-zero. Used with the /g option of pcretest (or similar
4018 code), it works in the same way as it does for Perl's /g option. If all
4019 alternatives of a regex begin with \G, the expression is anchored to the start
4020 match position, and the "anchored" flag is set in the compiled expression.
4021
4022 15. Some bugs concerning the handling of certain option changes within patterns
4023 have been fixed. These applied to options other than (?ims). For example,
4024 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4025 been the other way round. Some of this was related to change 7 above.
4026
4027 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4028 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4029 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4030 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4031
4032 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4033 mode. Unlike ".", it always matches newline, whatever the setting of
4034 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4035 assertions. Perl allows it, but it doesn't (in general) work because it can't
4036 calculate the length of the lookbehind. At least, that's the case for Perl
4037 5.8.0 - I've been told they are going to document that it doesn't work in
4038 future.
4039
4040 18. Added an error diagnosis for escapes that PCRE does not support: these are
4041 \L, \l, \N, \P, \p, \U, \u, and \X.
4042
4043 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4044 reading past the end of the pattern in cases such as /[abcd/.
4045
4046 20. PCRE was getting more memory than necessary for patterns with classes that
4047 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4048
4049 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4050 compiling PCRE for use with Virtual Pascal.
4051
4052 22. Small fix to the Makefile to make it work properly if the build is done
4053 outside the source tree.
4054
4055 23. Added a new extension: a condition to go with recursion. If a conditional
4056 subpattern starts with (?(R) the "true" branch is used if recursion has
4057 happened, whereas the "false" branch is used only at the top level.
4058
4059 24. When there was a very long string of literal characters (over 255 bytes
4060 without UTF support, over 250 bytes with UTF support), the computation of how
4061 much memory was required could be incorrect, leading to segfaults or other
4062 strange effects.
4063
4064 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4065 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4066 there was a subsequent back reference to those brackets. This meant that, for
4067 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4068 possible to check for precisely this case. All we can do is abandon the
4069 optimization if .* occurs inside capturing brackets when there are any back
4070 references whatsoever. (See below for a better fix that came later.)
4071
4072 26. The handling of the optimization for finding the first character of a
4073 non-anchored pattern, and for finding a character that is required later in the
4074 match were failing in some cases. This didn't break the matching; it just
4075 failed to optimize when it could. The way this is done has been re-implemented.
4076
4077 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4078
4079 28. Added a new feature that provides some of the functionality that Perl
4080 provides with (?{...}). The facility is termed a "callout". The way it is done
4081 in PCRE is for the caller to provide an optional function, by setting
4082 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4083 global variable. By default it is unset, which disables all calling out. To get
4084 the function called, the regex must include (?C) at appropriate points. This
4085 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4086 This provides a means of identifying different callout points. When PCRE
4087 reaches such a point in the regex, if pcre_callout has been set, the external
4088 function is called. It is provided with data in a structure called
4089 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4090 matching continues; if it returns a non-zero value, the match at the current
4091 point fails. However, backtracking will occur if possible. [This was changed
4092 later and other features added - see item 49 below.]
4093
4094 29. pcretest is upgraded to test the callout functionality. It provides a
4095 callout function that displays information. By default, it shows the start of
4096 the match and the current position in the text. There are some new data escapes
4097 to vary what happens:
4098
4099 \C+ in addition, show current contents of captured substrings
4100 \C- do not supply a callout function
4101 \C!n return 1 when callout number n is reached
4102 \C!n!m return 1 when callout number n is reached for the mth time
4103
4104 30. If pcregrep was called with the -l option and just a single file name, it
4105 output "<stdin>" if a match was found, instead of the file name.
4106
4107 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4108 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4109 pcre_exec(). This saves a malloc/free per call. The default value of
4110 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4111 when configuring.
4112
4113 32. The default maximum size of a compiled pattern is 64K. There have been a
4114 few cases of people hitting this limit. The code now uses macros to handle the
4115 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4116 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4117 configuring. Tests 2 and 5 work only with 2-byte links because they output
4118 debugging information about compiled patterns.
4119
4120 33. Internal code re-arrangements:
4121
4122 (a) Moved the debugging function for printing out a compiled regex into
4123 its own source file (printint.c) and used #include to pull it into
4124 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4125 separate copies.
4126
4127 (b) Defined the list of op-code names for debugging as a macro in
4128 internal.h so that it is next to the definition of the opcodes.
4129
4130 (c) Defined a table of op-code lengths for simpler skipping along compiled
4131 code. This is again a macro in internal.h so that it is next to the
4132 definition of the opcodes.
4133
4134 34. Added support for recursive calls to individual subpatterns, along the
4135 lines of Robin Houston's patch (but implemented somewhat differently).
4136
4137 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4138 allow it to read and process whole directories in Win32. This code was
4139 contributed by Lionel Fourquaux; it has not been tested by me.
4140
4141 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4142 used to name a group. Names consist of alphanumerics and underscores, and must
4143 be unique. Back references use the syntax (?P=name) and recursive calls use
4144 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4145 numbers. The function pcre_fullinfo() can be used after compilation to extract
4146 a name/number map. There are three relevant calls:
4147
4148 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4149 PCRE_INFO_NAMECOUNT yields the number of entries
4150 PCRE_INFO_NAMETABLE yields a pointer to the map.
4151
4152 The map is a vector of fixed-size entries. The size of each entry depends on
4153 the length of the longest name used. The first two bytes of each entry are the
4154 group number, most significant byte first. There follows the corresponding
4155 name, zero terminated. The names are in alphabetical order.
4156
4157 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4158 case instead of 255. Making it the same both with and without UTF-8 support
4159 means that the same test output works with both.
4160
4161 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4162 calling malloc() with a zero argument.
4163
4164 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4165 optimization. I've improved things by keeping a bitmap of backreferences with
4166 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4167 fact referenced, the optimization can be applied. It is unlikely that a
4168 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4169 the match to follow \n) will appear inside brackets with a number greater than
4170 31, but if it does, any back reference > 31 suppresses the optimization.
4171
4172 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4173 of disabling numbered capturing parentheses. Any opening parenthesis that is
4174 not followed by ? behaves as if it were followed by ?: but named parentheses
4175 can still be used for capturing (and they will acquire numbers in the usual
4176 way).
4177
4178 41. Redesigned the return codes from the match() function into yes/no/error so
4179 that errors can be passed back from deep inside the nested calls. A malloc
4180 failure while inside a recursive subpattern call now causes the
4181 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4182
4183 42. It is now possible to set a limit on the number of times the match()
4184 function is called in a call to pcre_exec(). This facility makes it possible to
4185 limit the amount of recursion and backtracking, though not in a directly
4186 obvious way, because the match() function is used in a number of different
4187 circumstances. The count starts from zero for each position in the subject
4188 string (for non-anchored patterns). The default limit is, for compatibility, a
4189 large number, namely 10 000 000. You can change this in two ways:
4190
4191 (a) When configuring PCRE before making, you can use --with-match-limit=n
4192 to set a default value for the compiled library.
4193
4194 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4195 a different value is set. See 45 below.
4196
4197 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4198
4199 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4200 of things that can be changed at compile time. The first argument specifies
4201 what is wanted and the second points to where the information is to be placed.
4202 The current list of available information is:
4203
4204 PCRE_CONFIG_UTF8
4205
4206 The output is an integer that is set to one if UTF-8 support is available;
4207 otherwise it is set to zero.
4208
4209 PCRE_CONFIG_NEWLINE
4210
4211 The output is an integer that it set to the value of the code that is used for
4212 newline. It is either LF (10) or CR (13).
4213
4214 PCRE_CONFIG_LINK_SIZE
4215
4216 The output is an integer that contains the number of bytes used for internal
4217 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4218
4219 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4220
4221 The output is an integer that contains the threshold above which the POSIX
4222 interface uses malloc() for output vectors. See item 31 above.
4223
4224 PCRE_CONFIG_MATCH_LIMIT
4225
4226 The output is an unsigned integer that contains the default limit of the number
4227 of match() calls in a pcre_exec() execution. See 42 above.
4228
4229 44. pcretest has been upgraded by the addition of the -C option. This causes it
4230 to extract all the available output from the new pcre_config() function, and to
4231 output it. The program then exits immediately.
4232
4233 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4234 order to support additional features. One way would have been to define
4235 pcre_exec2() (for example) with extra arguments, but this would not have been
4236 extensible, and would also have required all calls to the original function to
4237 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4238 is used for passing in "extra" data from pcre_study().
4239
4240 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4241 contains the following fields:
4242
4243 flags a bitmap indicating which of the following fields are set
4244 study_data opaque data from pcre_study()
4245 match_limit a way of specifying a limit on match() calls for a specific
4246 call to pcre_exec()
4247 callout_data data for callouts (see 49 below)
4248
4249 The flag bits are also defined in pcre.h, and are
4250
4251 PCRE_EXTRA_STUDY_DATA
4252 PCRE_EXTRA_MATCH_LIMIT
4253 PCRE_EXTRA_CALLOUT_DATA
4254
4255 The pcre_study() function now returns one of these new pcre_extra blocks, with
4256 the actual study data pointed to by the study_data field, and the
4257 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4258 before. That is, this change is entirely upwards-compatible and requires no
4259 change to existing code.
4260
4261 If you want to pass in additional data to pcre_exec(), you can either place it
4262 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4263 block.
4264
4265 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4266 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4267 times with different match limits, until it finds the minimum value needed for
4268 pcre_exec() to complete. The value is then output. This can be instructive; for
4269 most simple matches the number is quite small, but for pathological cases it
4270 gets very large very quickly.
4271
4272 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4273 returns the size of the data block pointed to by the study_data field in a
4274 pcre_extra block, that is, the value that was passed as the argument to
4275 pcre_malloc() when PCRE was getting memory in which to place the information
4276 created by pcre_study(). The fourth argument should point to a size_t variable.
4277 pcretest has been extended so that this information is shown after a successful
4278 pcre_study() call when information about the compiled regex is being displayed.
4279
4280 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4281 because what follows is always an absolute path. (Later: it turns out that this
4282 is more than cosmetic for MinGW, because it doesn't like empty path
4283 components.)
4284
4285 49. Some changes have been made to the callout feature (see 28 above):
4286
4287 (i) A callout function now has three choices for what it returns:
4288
4289 0 => success, carry on matching
4290 > 0 => failure at this point, but backtrack if possible
4291 < 0 => serious error, return this value from pcre_exec()
4292
4293 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4294 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4295 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4296 use by callout functions. It will never be used by PCRE itself.
4297
4298 (ii) The pcre_extra structure (see 45 above) has a void * field called
4299 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4300 pcre_callout_block structure has a field of the same name. The contents of
4301 the field passed in the pcre_extra structure are passed to the callout
4302 function in the corresponding field in the callout block. This makes it
4303 easier to use the same callout-containing regex from multiple threads. For
4304 testing, the pcretest program has a new data escape
4305
4306 \C*n pass the number n (may be negative) as callout_data
4307
4308 If the callout function in pcretest receives a non-zero value as
4309 callout_data, it returns that value.
4310
4311 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4312 there were some redundant $(CFLAGS) in commands that are now specified as
4313 $(LINK), which already includes $(CFLAGS).
4314
4315 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4316 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4317 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4318 one-byte characters throughout. Note that case-insensitive matching applies
4319 only to characters whose values are less than 256. PCRE doesn't support the
4320 notion of cases for higher-valued characters.
4321
4322 (i) A character class whose characters are all within 0-255 is handled as
4323 a bit map, and the map is inverted for negative classes. Previously, a
4324 character > 255 always failed to match such a class; however it should
4325 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4326
4327 (ii) A negated character class with a single character < 255 is coded as
4328 "not this character" (OP_NOT). This wasn't working properly when the test
4329 character was multibyte, either singly or repeated.
4330
4331 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4332 mode, for example: \x{100}{2,3}.
4333
4334 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4335 singly or repeated) now correctly test multibyte characters. However,
4336 PCRE doesn't recognize any characters with values greater than 255 as
4337 digits, spaces, or word characters. Such characters always match \D, \S,
4338 and \W, and never match \d, \s, or \w.
4339
4340 (v) Classes may now contain characters and character ranges with values
4341 greater than 255. For example: [ab\x{100}-\x{400}].
4342
4343 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4344 PCRE in UTF-8 mode.
4345
4346 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4347 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4348 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4349 value.)
4350
4351 53. The single man page has become too large. I have therefore split it up into
4352 a number of separate man pages. These also give rise to individual HTML pages;
4353 these are now put in a separate directory, and there is an index.html page that
4354 lists them all. Some hyperlinking between the pages has been installed.
4355
4356 54. Added convenience functions for handling named capturing parentheses.
4357
4358 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4359 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4360 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4361 are faulted.
4362
4363 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4364 calling configure. These values are used when compiling the dftables.c program
4365 which is run to generate the source of the default character tables. They
4366 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4367 you will need to set these values.
4368
4369 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4370
4371
4372 Version 3.9 02-Jan-02
4373 ---------------------
4374
4375 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4376
4377 2. If --disable-static was given, the building process failed when trying to
4378 build pcretest and pcregrep. (For some reason it was using libtool to compile
4379 them, which is not right, as they aren't part of the library.)
4380
4381
4382 Version 3.8 18-Dec-01
4383 ---------------------
4384
4385 1. The experimental UTF-8 code was completely screwed up. It was packing the
4386 bytes in the wrong order. How dumb can you get?
4387
4388
4389 Version 3.7 29-Oct-01
4390 ---------------------
4391
4392 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4393 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4394 this didn't happen under Solaris 8, where I normally test things.
4395
4396 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4397 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4398 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4399 this fix an hour or so after the initial 3.7 release.)
4400
4401
4402 Version 3.6 23-Oct-01
4403 ---------------------
4404
4405 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4406 offsets passed as NULL with zero offset count.
4407
4408 2. The config.guess and config.sub files had not been updated when I moved to
4409 the latest autoconf.
4410
4411
4412 Version 3.5 15-Aug-01
4413 ---------------------
4414
4415 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4416 had been forgotten.
4417
4418 2. By using declared but undefined structures, we can avoid using "void"
4419 definitions in pcre.h while keeping the internal definitions of the structures
4420 private.
4421
4422 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4423 user point of view, this means that both static and shared libraries are built
4424 by default, but this can be individually controlled. More of the work of
4425 handling this static/shared cases is now inside libtool instead of PCRE's make
4426 file.
4427
4428 4. The pcretest utility is now installed along with pcregrep because it is
4429 useful for users (to test regexs) and by doing this, it automatically gets
4430 relinked by libtool. The documentation has been turned into a man page, so
4431 there are now .1, .txt, and .html versions in /doc.
4432
4433 5. Upgrades to pcregrep:
4434 (i) Added long-form option names like gnu grep.
4435 (ii) Added --help to list all options with an explanatory phrase.
4436 (iii) Added -r, --recursive to recurse into sub-directories.
4437 (iv) Added -f, --file to read patterns from a file.
4438
4439 6. pcre_exec() was referring to its "code" argument before testing that
4440 argument for NULL (and giving an error if it was NULL).
4441
4442 7. Upgraded Makefile.in to allow for compiling in a different directory from
4443 the source directory.
4444
4445 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4446 options bits, the pointer it was passed was to an int instead of to an unsigned
4447 long int. This mattered only on 64-bit systems.
4448
4449 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4450 generated) instead of pcre.in, which it its source. Also made the same change
4451 in several of the .c files.
4452
4453 10. A new release of gcc defines printf() as a macro, which broke pcretest
4454 because it had an ifdef in the middle of a string argument for printf(). Fixed
4455 by using separate calls to printf().
4456
4457 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4458 script, to force use of CR or LF instead of \n in the source. On non-Unix
4459 systems, the value can be set in config.h.
4460
4461 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4462 absolute limit. Changed the text of the error message to make this clear, and
4463 likewise updated the man page.
4464
4465 13. The limit of 99 on the number of capturing subpatterns has been removed.
4466 The new limit is 65535, which I hope will not be a "real" limit.
4467
4468
4469 Version 3.4 22-Aug-00
4470 ---------------------
4471
4472 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4473
4474 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4475
4476
4477 Version 3.3 01-Aug-00
4478 ---------------------
4479
4480 1. If an octal character was given, but the value was greater than \377, it
4481 was not getting masked to the least significant bits, as documented. This could
4482 lead to crashes in some systems.
4483
4484 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4485 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4486
4487 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4488 These just pass their arguments on to (pcre_free)(), but they are provided
4489 because some uses of PCRE bind it to non-C systems that can call its functions,
4490 but cannot call free() or pcre_free() directly.
4491
4492 4. Add "make test" as a synonym for "make check". Corrected some comments in
4493 the Makefile.
4494
4495 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4496 Makefile.
4497
4498 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4499 command called pgrep for grepping around the active processes.
4500
4501 7. Added the beginnings of support for UTF-8 character strings.
4502
4503 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4504 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4505 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4506 out for the ar command.)
4507
4508
4509 Version 3.2 12-May-00
4510 ---------------------
4511
4512 This is purely a bug fixing release.
4513
4514 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4515 of ZA. This was just one example of several cases that could provoke this bug,
4516 which was introduced by change 9 of version 2.00. The code for breaking
4517 infinite loops after an iteration that matches an empty string was't working
4518 correctly.
4519
4520 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4521 when matched against abbab (for example). After matching an empty string, it
4522 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4523 caused it to match further down the string than it should.
4524
4525 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4526 was there because it doesn't seem to be needed, and it causes trouble on some
4527 systems, as it is not a Standard C header. It has been removed.
4528
4529 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4530 were reported on the Macintosh. The changes were from
4531
4532 while ((c = *(++ptr)) != 0 && c != '\n');
4533 to
4534 while ((c = *(++ptr)) != 0 && c != '\n') ;
4535
4536 Totally extraordinary, but if that's what it takes...
4537
4538 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4539 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4540 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4541 assumes the way PCRE uses memmove() (always moving upwards).
4542
4543 6. PCRE is being used in one environment where strchr() is not available. There
4544 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4545 faster code anyway.
4546
4547
4548 Version 3.1 09-Feb-00
4549 ---------------------
4550
4551 The only change in this release is the fixing of some bugs in Makefile.in for
4552 the "install" target:
4553
4554 (1) It was failing to install pcreposix.h.
4555
4556 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4557
4558
4559 Version 3.0 01-Feb-00
4560 ---------------------
4561
4562 1. Add support for the /+ modifier to perltest (to output $` like it does in
4563 pcretest).
4564
4565 2. Add support for the /g modifier to perltest.
4566
4567 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4568 matches null strings.
4569
4570 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4571 pattern. Perl treats empty patterns specially - it reuses the most recent
4572 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4573 effect.
4574
4575 5. The POSIX interface was broken in that it was just handing over the POSIX
4576 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4577 required a bigger vector, with some working space on the end. This means that
4578 the POSIX wrapper now has to get and free some memory, and copy the results.
4579
4580 6. Added some simple autoconf support, placing the test data and the
4581 documentation in separate directories, re-organizing some of the
4582 information files, and making it build pcre-config (a GNU standard). Also added
4583 libtool support for building PCRE as a shared library, which is now the
4584 default.
4585
4586 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4587 09 are not valid octal constants. Single digits will be used for minor values
4588 less than 10.
4589
4590 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4591 existing programs that set these in the POSIX interface can use PCRE without
4592 modification.
4593
4594 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4595 return all that pcre_info() returns, plus additional data. The pcre_info()
4596 function is retained for compatibility, but is considered to be obsolete.
4597
4598 10. Added experimental recursion feature (?R) to handle one common case that
4599 Perl 5.6 will be able to do with (?p{...}).
4600
4601 11. Added support for POSIX character classes like [:alpha:], which Perl is
4602 adopting.
4603
4604
4605 Version 2.08 31-Aug-99
4606 ----------------------
4607
4608 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4609 trying to match at the startoffset position, but instead was moving forward to
4610 the next newline as if a previous match had failed.
4611
4612 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4613 and could get into a loop if a null string was matched other than at the start
4614 of the subject.
4615
4616 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4617 be distinguished at compile time, and for completeness also added PCRE_DATE.
4618
4619 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4620 in GnuWin32 environments.
4621
4622
4623 Version 2.07 29-Jul-99
4624 ----------------------
4625
4626 1. The documentation is now supplied in plain text form and HTML as well as in
4627 the form of man page sources.
4628
4629 2. C++ compilers don't like assigning (void *) values to other pointer types.
4630 In particular this affects malloc(). Although there is no problem in Standard
4631 C, I've put in casts to keep C++ compilers happy.
4632
4633 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4634 should be (const char *).
4635
4636 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4637 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4638 However, I haven't made this a standard facility. The documentation doesn't
4639 mention it, and the Makefile doesn't support it.
4640
4641 5. The Makefile now contains an "install" target, with editable destinations at
4642 the top of the file. The pcretest program is not installed.
4643
4644 6. pgrep -V now gives the PCRE version number and date.
4645
4646 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4647 causing the entire string to be ignored, instead of just the last character.
4648
4649 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4650 non-matching string, it can take a very, very long time, even for strings of
4651 quite modest length, because of the nested recursion. PCRE now does better in
4652 some of these cases. It does this by remembering the last required literal
4653 character in the pattern, and pre-searching the subject to ensure it is present
4654 before running the real match. In other words, it applies a heuristic to detect
4655 some types of certain failure quickly, and in the above example, if presented
4656 with a string that has no trailing " it gives "no match" very quickly.
4657
4658 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4659 other alternatives are tried instead.
4660
4661
4662 Version 2.06 09-Jun-99
4663 ----------------------
4664
4665 1. Change pcretest's output for amount of store used to show just the code
4666 space, because the remainder (the data block) varies in size between 32-bit and
4667 64-bit systems.
4668
4669 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4670 start matching at. This allows lookbehinds to work when searching for multiple
4671 occurrences in a string.
4672
4673 3. Added additional options to pcretest for testing multiple occurrences:
4674
4675 /+ outputs the rest of the string that follows a match
4676 /g loops for multiple occurrences, using the new startoffset argument
4677 /G loops for multiple occurrences by passing an incremented pointer
4678
4679 4. PCRE wasn't doing the "first character" optimization for patterns starting
4680 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4681 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4682 the letter 'x'. On long subject strings, this gives a significant speed-up.
4683
4684
4685 Version 2.05 21-Apr-99
4686 ----------------------
4687
4688 1. Changed the type of magic_number from int to long int so that it works
4689 properly on 16-bit systems.
4690
4691 2. Fixed a bug which caused patterns starting with .* not to work correctly
4692 when the subject string contained newline characters. PCRE was assuming
4693 anchoring for such patterns in all cases, which is not correct because .* will
4694 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4695 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4696 must be retried after every newline in the subject.
4697
4698
4699 Version 2.04 18-Feb-99
4700 ----------------------
4701
4702 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4703 computation of the store needed to hold the pattern was incorrect (too large).
4704 If such patterns were nested a few deep, this could multiply and become a real
4705 problem.
4706
4707 2. Added /M option to pcretest to show the memory requirement of a specific
4708 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4709
4710 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4711 compiled in such a way that the backtracking after subsequent failure was
4712 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4713 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4714
4715
4716 Version 2.03 02-Feb-99
4717 ----------------------
4718
4719 1. Fixed typo and small mistake in man page.
4720
4721 2. Added 4th condition (GPL supersedes if conflict) and created separate
4722 LICENCE file containing the conditions.
4723
4724 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4725 Perl, that is the internal \ allows the delimiter to be included in the
4726 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4727 the final delimiter, add \ to the end of the pattern (to test the error).
4728
4729 4. Added the convenience functions for extracting substrings after a successful
4730 match. Updated pcretest to make it able to test these functions.
4731
4732
4733 Version 2.02 14-Jan-99
4734 ----------------------
4735
4736 1. Initialized the working variables associated with each extraction so that
4737 their saving and restoring doesn't refer to uninitialized store.
4738
4739 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4740 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4741 fix the problem.
4742
4743 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4744 calls, and wasn't printing the correct value for compiling calls. Increased the
4745 default value of LOOPREPEAT, and the number of significant figures in the
4746 times.
4747
4748 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4749
4750 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4751 a building problem on Windows NT with a FAT file system.
4752
4753
4754 Version 2.01 21-Oct-98
4755 ----------------------
4756
4757 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4758 to character tables built by pcre_maketables() in the current locale. If NULL
4759 is passed, the default tables are used.
4760
4761
4762 Version 2.00 24-Sep-98
4763 ----------------------
4764
4765 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4766 it any more.
4767
4768 2. Allow quantification of (?>) groups, and make it work correctly.
4769
4770 3. The first character computation wasn't working for (?>) groups.
4771
4772 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4773 end of the subject) and add 5.005's \z, which really does match only at the
4774 very end of the subject.
4775
4776 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4777
4778 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4779 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4780 localized options. All options to pcre_study() were also removed.
4781
4782 7. Add other new features from 5.005:
4783
4784 $(?<= positive lookbehind
4785 $(?<! negative lookbehind
4786 (?imsx-imsx) added the unsetting capability
4787 such a setting is global if at outer level; local otherwise
4788 (?imsx-imsx:) non-capturing groups with option setting
4789 (?(cond)re|re) conditional pattern matching
4790
4791 A backreference to itself in a repeated group matches the previous
4792 captured string.
4793
4794 8. General tidying up of studying (both automatic and via "study")
4795 consequential on the addition of new assertions.
4796
4797 9. As in 5.005, unlimited repeated groups that could match an empty substring
4798 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4799 runtime if any iteration does actually match an empty substring.
4800
4801 10. Include the RunTest script in the distribution.
4802
4803 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4804 discrepancies, some of which were old and were also with respect to 5.004. They
4805 have now been fixed.
4806
4807
4808 Version 1.09 28-Apr-98
4809 ----------------------
4810
4811 1. A negated single character class followed by a quantifier with a minimum
4812 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4813 program crashes, or just wrong answers. This did not apply to negated classes
4814 containing more than one character, or to minima other than one.
4815
4816
4817 Version 1.08 27-Mar-98
4818 ----------------------
4819
4820 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4821
4822 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4823 latter must appear before anything that relies on it in the pattern.
4824
4825
4826 Version 1.07 16-Feb-98
4827 ----------------------
4828
4829 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4830 repeat of a potentially empty string).
4831
4832
4833 Version 1.06 23-Jan-98
4834 ----------------------
4835
4836 1. Added Markus Oberhumer's little patches for C++.
4837
4838 2. Literal strings longer than 255 characters were broken.
4839
4840
4841 Version 1.05 23-Dec-97
4842 ----------------------
4843
4844 1. Negated character classes containing more than one character were failing if
4845 PCRE_CASELESS was set at run time.
4846
4847
4848 Version 1.04 19-Dec-97
4849 ----------------------
4850
4851 1. Corrected the man page, where some "const" qualifiers had been omitted.
4852
4853 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4854 input syntax.
4855
4856 3. Fixed memory leak which occurred when a regex with back references was
4857 matched with an offsets vector that wasn't big enough. The temporary memory
4858 that is used in this case wasn't being freed if the match failed.
4859
4860 4. Tidied pcretest to ensure it frees memory that it gets.
4861
4862 5. Temporary memory was being obtained in the case where the passed offsets
4863 vector was exactly big enough.
4864
4865 6. Corrected definition of offsetof() from change 5 below.
4866
4867 7. I had screwed up change 6 below and broken the rules for the use of
4868 setjmp(). Now fixed.
4869
4870
4871 Version 1.03 18-Dec-97
4872 ----------------------
4873
4874 1. A erroneous regex with a missing opening parenthesis was correctly
4875 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4876 on some systems.
4877
4878 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4879 it was reported that one broken compiler failed on the former because "code" is
4880 also an independent variable.
4881
4882 3. The erroneous regex a[]b caused an array overrun reference.
4883
4884 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4885 fail on data ending with that character. (It was going on too far, and checking
4886 the next character, typically a binary zero.) This was specific to the
4887 optimized code for single-character negative classes.
4888
4889 5. Added a contributed patch from the TIN world which does the following:
4890
4891 + Add an undef for memmove, in case the the system defines a macro for it.
4892
4893 + Add a definition of offsetof(), in case there isn't one. (I don't know
4894 the reason behind this - offsetof() is part of the ANSI standard - but
4895 it does no harm).
4896
4897 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4898 most of the places where whitespace preceded '#'. I have given up and
4899 allowed the remaining 2 cases to be at the margin.
4900
4901 + Rename some variables in pcre to eliminate shadowing. This seems very
4902 pedantic, but does no harm, of course.
4903
4904 6. Moved the call to setjmp() into its own function, to get rid of warnings
4905 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4906
4907 7. Constructs such as \d{8,} were compiling into the equivalent of
4908 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4909 outcome, but in this particular case used more store than had been allocated,
4910 which caused the bug to be discovered because it threw up an internal error.
4911
4912 8. The debugging code in both pcre and pcretest for outputting the compiled
4913 form of a regex was going wrong in the case of back references followed by
4914 curly-bracketed repeats.
4915
4916
4917 Version 1.02 12-Dec-97
4918 ----------------------
4919
4920 1. Typos in pcre.3 and comments in the source fixed.
4921
4922 2. Applied a contributed patch to get rid of places where it used to remove
4923 'const' from variables, and fixed some signed/unsigned and uninitialized
4924 variable warnings.
4925
4926 3. Added the "runtest" target to Makefile.
4927
4928 4. Set default compiler flag to -O2 rather than just -O.
4929
4930
4931 Version 1.01 19-Nov-97
4932 ----------------------
4933
4934 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4935 like /([ab]*)*/, that is, for classes with more than one character in them.
4936
4937 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4938 as /((?>a*))*/ (a PCRE_EXTRA facility).
4939
4940
4941 Version 1.00 18-Nov-97
4942 ----------------------
4943
4944 1. Added compile-time macros to support systems such as SunOS4 which don't have
4945 memmove() or strerror() but have other things that can be used instead.
4946
4947 2. Arranged that "make clean" removes the executables.
4948
4949
4950 Version 0.99 27-Oct-97
4951 ----------------------
4952
4953 1. Fixed bug in code for optimizing classes with only one character. It was
4954 initializing a 32-byte map regardless, which could cause it to run off the end
4955 of the memory it had got.
4956
4957 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4958
4959
4960 Version 0.98 22-Oct-97
4961 ----------------------
4962
4963 1. Fixed bug in code for handling temporary memory usage when there are more
4964 back references than supplied space in the ovector. This could cause segfaults.
4965
4966
4967 Version 0.97 21-Oct-97
4968 ----------------------
4969
4970 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4971
4972 2. Optimized negated single characters not to use a bit map.
4973
4974 3. Brought error texts together as macro definitions; clarified some of them;
4975 fixed one that was wrong - it said "range out of order" when it meant "invalid
4976 escape sequence".
4977
4978 4. Changed some char * arguments to const char *.
4979
4980 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4981
4982 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4983 pcretest.
4984
4985
4986 Version 0.96 16-Oct-97
4987 ----------------------
4988
4989 1. Added a simple "pgrep" utility to the distribution.
4990
4991 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4992 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4993 where "ddd" means "one or more decimal digits".
4994
4995 3. Fixed serious bug. If a pattern had a back reference, but the call to
4996 pcre_exec() didn't supply a large enough ovector to record the related
4997 identifying subpattern, the match always failed. PCRE now remembers the number
4998 of the largest back reference, and gets some temporary memory in which to save
4999 the offsets during matching if necessary, in order to ensure that
5000 backreferences always work.
5001
5002 4. Increased the compatibility with Perl in a number of ways:
5003
5004 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5005 to request this handling. The option can be set at compile or exec time.
5006
5007 (b) $ matches before a terminating newline by default; an option
5008 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5009 mode). The option can be set at compile or exec time.
5010
5011 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5012 the same as Perl's. If the decimal number it represents is less than 10
5013 or there aren't that many previous left capturing parentheses, an octal
5014 escape is read. Inside a character class, it's always an octal escape,
5015 even if it is a single digit.
5016
5017 (d) An escaped but undefined alphabetic character is taken as a literal,
5018 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5019 escapes.
5020
5021 (e) {0} is now permitted. (The previous item is removed from the compiled
5022 pattern).
5023
5024 5. Changed all the names of code files so that the basic parts are no longer
5025 than 10 characters, and abolished the teeny "globals.c" file.
5026
5027 6. Changed the handling of character classes; they are now done with a 32-byte
5028 bit map always.
5029
5030 7. Added the -d and /D options to pcretest to make it possible to look at the
5031 internals of compilation without having to recompile pcre.
5032
5033
5034 Version 0.95 23-Sep-97
5035 ----------------------
5036
5037 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5038 \x20 at the start of a run of normal characters. These were being treated as
5039 real characters, instead of the source characters being re-checked.
5040
5041
5042 Version 0.94 18-Sep-97
5043 ----------------------
5044
5045 1. The functions are now thread-safe, with the caveat that the global variables
5046 containing pointers to malloc() and free() or alternative functions are the
5047 same for all threads.
5048
5049 2. Get pcre_study() to generate a bitmap of initial characters for non-
5050 anchored patterns when this is possible, and use it if passed to pcre_exec().
5051
5052
5053 Version 0.93 15-Sep-97
5054 ----------------------
5055
5056 1. /(b)|(:+)/ was computing an incorrect first character.
5057
5058 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5059 but not actually doing anything yet.
5060
5061 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5062 as Perl does (e.g. [-az] or [az-]).
5063
5064 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5065 all possible positions.
5066
5067 5. Split up into different modules to avoid including unneeded functions in a
5068 compiled binary. However, compile and exec are still in one module. The "study"
5069 function is split off.
5070
5071 6. The character tables are now in a separate module whose source is generated
5072 by an auxiliary program - but can then be edited by hand if required. There are
5073 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5074 toupper() in the code.
5075
5076 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5077 make them global. Abolish the function for setting them, as the caller can now
5078 set them directly.
5079
5080
5081 Version 0.92 11-Sep-97
5082 ----------------------
5083
5084 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5085 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5086
5087 2. Caseless matching was not working in character classes if the characters in
5088 the pattern were in upper case.
5089
5090 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5091
5092 4. Make PCRE_ANCHORED public and accept as a compile option.
5093
5094 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5095 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5096 pass them.
5097
5098 6. Give an error if bad option bits passed at compile or run time.
5099
5100 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5101 pcretest to cause it to pass that flag.
5102
5103 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5104 options, and the first character, if set.
5105
5106 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5107
5108
5109 Version 0.91 10-Sep-97
5110 ----------------------
5111
5112 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5113 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5114
5115 2. PCRE was looping on encountering an indefinitely repeated back reference to
5116 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5117 Perl does - treats the match as successful.
5118
5119 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5