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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1407 - (show annotations)
Sun Dec 1 16:31:40 2013 UTC (5 years, 8 months ago) by ph10
File size: 257500 byte(s)
Add Jeff Trawick's patch to CMakeLists.txt for MSVC features.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.34 19-November-2013
5 -----------------------------
6
7 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
8 executable memory. Patch inspired by Carsten Klein.
9
10 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
11 this macro is never tested and has no effect, because the work to support
12 coverage involves only compiling and linking options and special targets in
13 the Makefile. The comment in config.h implied that defining the macro would
14 enable coverage support, which is totally false. There was also support for
15 setting this macro in the CMake files (my fault, I just copied it from
16 configure). SUPPORT_GCOV has now been removed.
17
18 3. Make a small performance improvement in strlen16() and strlen32() in
19 pcretest.
20
21 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
22 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
23 these cases). There was also one in pcretest.c.
24
25 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
26
27 6. In UTF mode, the code for checking whether a group could match an empty
28 string (which is used for indefinitely repeated groups to allow for
29 breaking an infinite loop) was broken when the group contained a repeated
30 negated single-character class with a character that occupied more than one
31 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
32 UTF-8 mode). The effect was undefined: the group might or might not be
33 deemed as matching an empty string, or the program might have crashed.
34
35 7. The code for checking whether a group could match an empty string was not
36 recognizing that \h, \H, \v, \V, and \R must match a character.
37
38 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
39 an empty string. If it can, pcretest shows this in its information output.
40
41 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
42 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
43 matched by pcre_exec() without using JIT:
44
45 (a) If the rest of the pattern did not match after a maximal run of
46 grapheme clusters, the code for backing up to try with fewer of them
47 did not always back up over a full grapheme when characters that do not
48 have the modifier quality were involved, e.g. Hangul syllables.
49
50 (b) If the match point in a subject started with modifier character, and
51 there was no match, the code could incorrectly back up beyond the match
52 point, and potentially beyond the first character in the subject,
53 leading to a segfault or an incorrect match result.
54
55 10. A conditional group with an assertion condition could lead to PCRE
56 recording an incorrect first data item for a match if no other first data
57 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
58 first data item, and therefore matched "ca" after "c" instead of at the
59 start.
60
61 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
62 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
63
64 12. The source of pcregrep now includes z/OS-specific code so that it can be
65 compiled for z/OS as part of the special z/OS distribution.
66
67 13. Added the -T and -TM options to pcretest.
68
69 14. The code in pcre_compile.c for creating the table of named capturing groups
70 has been refactored. Instead of creating the table dynamically during the
71 actual compiling pass, the information is remembered during the pre-compile
72 pass (on the stack unless there are more than 20 named groups, in which
73 case malloc() is used) and the whole table is created before the actual
74 compile happens. This has simplified the code (it is now nearly 150 lines
75 shorter) and prepared the way for better handling of references to groups
76 with duplicate names.
77
78 15. A back reference to a named subpattern when there is more than one of the
79 same name now checks them in the order in which they appear in the pattern.
80 The first one that is set is used for the reference. Previously only the
81 first one was inspected. This change makes PCRE more compatible with Perl.
82
83 16. Unicode character properties were updated from Unicode 6.3.0.
84
85 17. The compile-time code for auto-possessification has been refactored, based
86 on a patch by Zoltan Herczeg. It now happens after instead of during
87 compilation. The code is cleaner, and more cases are handled. The option
88 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
89 options in pcretest are provided to set it. It can also be set by
90 (*NO_AUTO_POSSESS) at the start of a pattern.
91
92 18. The character VT has been added to the default ("C" locale) set of
93 characters that match \s and are generally treated as white space,
94 following this same change in Perl 5.18. There is now no difference between
95 "Perl space" and "POSIX space". Whether VT is treated as white space in
96 other locales depends on the locale.
97
98 19. The code for checking named groups as conditions, either for being set or
99 for being recursed, has been refactored (this is related to 14 and 15
100 above). Processing unduplicated named groups should now be as fast at
101 numerical groups, and processing duplicated groups should be faster than
102 before.
103
104 20. Two patches to the CMake build system, by Alexander Barkov:
105
106 (1) Replace the "source" command by "." in CMakeLists.txt because
107 "source" is a bash-ism.
108
109 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
110 without these the CMake build does not work on Solaris.
111
112 21. Perl has changed its handling of \8 and \9. If there is no previously
113 encountered capturing group of those numbers, they are treated as the
114 literal characters 8 and 9 instead of a binary zero followed by the
115 literals. PCRE now does the same.
116
117 22. Following Perl, added \o{} to specify codepoints in octal, making it
118 possible to specify values greater than 0777 and also making them
119 unambiguous.
120
121 23. Perl now gives an error for missing closing braces after \x{... instead of
122 treating the string as literal. PCRE now does the same.
123
124 24. RunTest used to grumble if an inappropriate test was selected explicitly,
125 but just skip it when running all tests. This make it awkward to run ranges
126 of tests when one of them was inappropriate. Now it just skips any
127 inappropriate tests, as it always did when running all tests.
128
129 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
130 character types such as \d or \w, too many callouts were inserted, and the
131 data that they returned was rubbish.
132
133 26. In UCP mode, \s was not matching two of the characters that Perl matches,
134 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
135 were matched by \h. The code has now been refactored so that the lists of
136 the horizontal and vertical whitespace characters used for \h and \v (which
137 are defined only in one place) are now also used for \s.
138
139 27. Add JIT support for the 64 bit TileGX architecture.
140 Patch by Jiong Wang (Tilera Corporation).
141
142 28. Possessive quantifiers for classes (both explicit and automatically
143 generated) now use special opcodes instead of wrapping in ONCE brackets.
144
145 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
146 (because it's meaningless), this was not happening when PCRE_CASELESS was
147 set. Not wrong, but inefficient.
148
149 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
150 properties for \w, \d, etc) is present in a test regex. Otherwise if the
151 test contains no characters greater than 255, Perl doesn't realise it
152 should be using Unicode semantics.
153
154 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
155 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
156 does in Unicode mode.
157
158 32. Added the "forbid" facility to pcretest so that putting tests into the
159 wrong test files can sometimes be quickly detected.
160
161 33. There is now a limit (default 250) on the depth of nesting of parentheses.
162 This limit is imposed to control the amount of system stack used at compile
163 time. It can be changed at build time by --with-parens-nest-limit=xxx or
164 the equivalent in CMake.
165
166 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
167 errors. Perl warns for these when in warning mode, but PCRE has no facility
168 for giving warnings.
169
170 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
171 However, this was not working for (?!) because it is optimized to (*FAIL),
172 for which PCRE does not allow quantifiers. The optimization is now disabled
173 when a quantifier follows (?!). I can't see any use for this, but it makes
174 things uniform.
175
176 36. Perl no longer allows group names to start with digits, so I have made this
177 change also in PCRE. It simplifies the code a bit.
178
179 37. In extended mode, Perl ignores spaces before a + that indicates a
180 possessive quantifier. PCRE allowed a space before the quantifier, but not
181 before the possessive +. It now does.
182
183 38. The use of \K (reset reported match start) within a repeated possessive
184 group such as (a\Kb)*+ was not working.
185
186 40. Document that the same character tables must be used at compile time and
187 run time, and that the facility to pass tables to pcre_exec() and
188 pcre_dfa_exec() is for use only with saved/restored patterns.
189
190 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
191 features for Builds with MSVC:
192
193 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
194 builds). The .rc files can be used to set FileDescription and many other
195 attributes.
196
197 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
198 This allows higher-level build scripts which want .pdb files to avoid
199 hard-coding the exact files needed."
200
201
202 Version 8.33 28-May-2013
203 ------------------------
204
205 1. Added 'U' to some constants that are compared to unsigned integers, to
206 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
207 variables that are added to signed variables, to ensure the result is
208 signed and can be negated.
209
210 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
211 CMake config files.
212
213 3. Revise the creation of config.h.generic so that all boolean macros are
214 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
215 overriding via -D on the command line possible.
216
217 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
218 to unsigned int is reported to make a quite noticeable speed difference in
219 a specific Windows environment. Testing on Linux did also appear to show
220 some benefit (and it is clearly not harmful). Also fixed the definition of
221 Xop which should be unsigned.
222
223 5. Related to (4), changing the definition of the intermediate variable cc
224 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
225 improvements.
226
227 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
228 unnecessary spaces.
229
230 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
231 and later.
232
233 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
234
235 9. Optimizing fast_forward_start_bits in JIT.
236
237 10. Adding support for callouts in JIT, and fixing some issues revealed
238 during this work. Namely:
239
240 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
241
242 (b) Minimum length was not checked before the matching is started.
243
244 11. The value of capture_last that is passed to callouts was incorrect in some
245 cases when there was a capture on one path that was subsequently abandoned
246 after a backtrack. Also, the capture_last value is now reset after a
247 recursion, since all captures are also reset in this case.
248
249 12. The interpreter no longer returns the "too many substrings" error in the
250 case when an overflowing capture is in a branch that is subsequently
251 abandoned after a backtrack.
252
253 13. In the pathological case when an offset vector of size 2 is used, pcretest
254 now prints out the matched string after a yield of 0 or 1.
255
256 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
257 Only supported by the JIT compiler at the moment.
258
259 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
260
261 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
262 offset of the starting point of the matching process, provided the offsets
263 vector is large enough.
264
265 17. The \A escape now records a lookbehind value of 1, though its execution
266 does not actually inspect the previous character. This is to ensure that,
267 in partial multi-segment matching, at least one character from the old
268 segment is retained when a new segment is processed. Otherwise, if there
269 are no lookbehinds in the pattern, \A might match incorrectly at the start
270 of a new segment.
271
272 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
273
274 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
275 gives some modest performance improvement in 8-bit mode.
276
277 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
278 be expressed in certain programming languages using Universal Character
279 Names.
280
281 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
282 which points out that "non characters" are not "characters that may not
283 appear in Unicode strings" but rather "characters that are reserved for
284 internal use and have only local meaning".
285
286 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
287 there was a conditional group that depended on an assertion, if the
288 assertion was false, the callout that immediately followed the alternation
289 in the condition was skipped when pcre_exec() was used for matching.
290
291 23. Allow an explicit callout to be inserted before an assertion that is the
292 condition for a conditional group, for compatibility with automatic
293 callouts, which always insert a callout at this point.
294
295 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
296 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
297
298 25. (*PRUNE) is now supported by the JIT compiler.
299
300 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
301
302 27. Fix the case where there are two or more SKIPs with arguments that may be
303 ignored.
304
305 28. (*SKIP) is now supported by the JIT compiler.
306
307 29. (*THEN) is now supported by the JIT compiler.
308
309 30. Update RunTest with additional test selector options.
310
311 31. The way PCRE handles backtracking verbs has been changed in two ways.
312
313 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
314 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
315 backtracking. In some cases this makes it more Perl-compatible, but Perl's
316 rather obscure rules do not always do the same thing.
317
318 (2) Previously, backtracking verbs were confined within assertions. This is
319 no longer the case for positive assertions, except for (*ACCEPT). Again,
320 this sometimes improves Perl compatibility, and sometimes does not.
321
322 32. A number of tests that were in test 2 because Perl did things differently
323 have been moved to test 1, because either Perl or PCRE has changed, and
324 these tests are now compatible.
325
326 32. Backtracking control verbs are now handled in the same way in JIT and
327 interpreter.
328
329 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
330 contained a forward subroutine reference caused a compile error.
331
332 34. Auto-detect and optimize limited repetitions in JIT.
333
334 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
335 blocking (*UTF) etc.
336
337 36. In the interpreter, maximizing pattern repetitions for characters and
338 character types now use tail recursion, which reduces stack usage.
339
340 37. The value of the max lookbehind was not correctly preserved if a compiled
341 and saved regex was reloaded on a host of different endianness.
342
343 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
344 of the compiled pattern block, expand the flags field from 16 to 32 bits
345 because it was almost full.
346
347 39. Try madvise first before posix_madvise.
348
349 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
350 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
351 original change locked out all matches of empty strings. This has been
352 changed so that one match of an empty string per line is recognized.
353 Subsequent searches on the same line (for colouring or for --only-matching,
354 for example) do not recognize empty strings.
355
356 41. Applied a user patch to fix a number of spelling mistakes in comments.
357
358 42. Data lines longer than 65536 caused pcretest to crash.
359
360 43. Clarified the data type for length and startoffset arguments for pcre_exec
361 and pcre_dfa_exec in the function-specific man pages, where they were
362 explicitly stated to be in bytes, never having been updated. I also added
363 some clarification to the pcreapi man page.
364
365 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
366 a segmentation fault.
367
368
369 Version 8.32 30-November-2012
370 -----------------------------
371
372 1. Improved JIT compiler optimizations for first character search and single
373 character iterators.
374
375 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
376 Patch by Daniel Richard G.
377
378 3. Single character iterator optimizations in the JIT compiler.
379
380 4. Improved JIT compiler optimizations for character ranges.
381
382 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
383 Reported by Giuseppe D'Angelo.
384
385 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
386 of a line, was being set incorrectly in cases where .* appeared inside
387 atomic brackets at the start of a pattern, or where there was a subsequent
388 *PRUNE or *SKIP.
389
390 7. Improved instruction cache flush for POWER/PowerPC.
391 Patch by Daniel Richard G.
392
393 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
394 grep:
395
396 (a) There is now no limit to the number of patterns to be matched.
397
398 (b) An error is given if a pattern is too long.
399
400 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
401 are now supported.
402
403 (d) --exclude-from and --include-from (multiple use) have been added.
404
405 (e) Exclusions and inclusions now apply to all files and directories, not
406 just to those obtained from scanning a directory recursively.
407
408 (f) Multiple uses of -f and --file-list are now supported.
409
410 (g) In a Windows environment, the default for -d has been changed from
411 "read" (the GNU grep default) to "skip", because otherwise the presence
412 of a directory in the file list provokes an error.
413
414 (h) The documentation has been revised and clarified in places.
415
416 9. Improve the matching speed of capturing brackets.
417
418 10. Changed the meaning of \X so that it now matches a Unicode extended
419 grapheme cluster.
420
421 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
422 out POSIX threads when JIT support is configured.
423
424 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
425
426 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
427 structure could go wrong in environments where size_t is not the same size
428 as int.
429
430 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
431 set.
432
433 15. The EBCDIC support had decayed; later updates to the code had included
434 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
435 general tidy up of EBCDIC-related issues, and the documentation was also
436 not quite right. There is now a test that can be run on ASCII systems to
437 check some of the EBCDIC-related things (but is it not a full test).
438
439 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
440 in a small tidy to the code.
441
442 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
443
444 18. If the --only-matching (-o) option in pcregrep is specified multiple
445 times, each one causes appropriate output. For example, -o1 -o2 outputs the
446 substrings matched by the 1st and 2nd capturing parentheses. A separating
447 string can be specified by --om-separator (default empty).
448
449 19. Improving the first n character searches.
450
451 20. Turn case lists for horizontal and vertical white space into macros so that
452 they are defined only once.
453
454 21. This set of changes together give more compatible Unicode case-folding
455 behaviour for characters that have more than one other case when UCP
456 support is available.
457
458 (a) The Unicode property table now has offsets into a new table of sets of
459 three or more characters that are case-equivalent. The MultiStage2.py
460 script that generates these tables (the pcre_ucd.c file) now scans
461 CaseFolding.txt instead of UnicodeData.txt for character case
462 information.
463
464 (b) The code for adding characters or ranges of characters to a character
465 class has been abstracted into a generalized function that also handles
466 case-independence. In UTF-mode with UCP support, this uses the new data
467 to handle characters with more than one other case.
468
469 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
470 whose other case is greater than 256 are now correctly matched
471 caselessly. Previously, the high codepoint matched the low one, but not
472 vice versa.
473
474 (d) The processing of \h, \H, \v, and \ in character classes now makes use
475 of the new class addition function, using character lists defined as
476 macros alongside the case definitions of 20 above.
477
478 (e) Caseless back references now work with characters that have more than
479 one other case.
480
481 (f) General caseless matching of characters with more than one other case
482 is supported.
483
484 22. Unicode character properties were updated from Unicode 6.2.0
485
486 23. Improved CMake support under Windows. Patch by Daniel Richard G.
487
488 24. Add support for 32-bit character strings, and UTF-32
489
490 25. Major JIT compiler update (code refactoring and bugfixing).
491 Experimental Sparc 32 support is added.
492
493 26. Applied a modified version of Daniel Richard G's patch to create
494 pcre.h.generic and config.h.generic by "make" instead of in the
495 PrepareRelease script.
496
497 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
498 pcre_compile.c when checking for a zero character.
499
500 28. Introducing a native interface for JIT. Through this interface, the compiled
501 machine code can be directly executed. The purpose of this interface is to
502 provide fast pattern matching, so several sanity checks are not performed.
503 However, feature tests are still performed. The new interface provides
504 1.4x speedup compared to the old one.
505
506 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
507 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
508 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
509
510 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
511 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
512 the "old" RFC 2279). Instead, it ended up passing a negative length to
513 pcre_exec().
514
515 31. Add support for GCC's visibility feature to hide internal functions.
516
517 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
518 "unknown -C option" after outputting 0 or 1.
519
520 33. There is now support for generating a code coverage report for the test
521 suite in environments where gcc is the compiler and lcov is installed. This
522 is mainly for the benefit of the developers.
523
524 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
525 unaddressable using valgrind annotations, allowing valgrind to detect
526 invalid memory accesses. This is mainly for the benefit of the developers.
527
528 25. (*UTF) can now be used to start a pattern in any of the three libraries.
529
530 26. Give configure error if --enable-cpp but no C++ compiler found.
531
532
533 Version 8.31 06-July-2012
534 -------------------------
535
536 1. Fixing a wrong JIT test case and some compiler warnings.
537
538 2. Removed a bashism from the RunTest script.
539
540 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
541 to unsigned type, result still unsigned" that was given by an MS compiler
542 on encountering the code "-sizeof(xxx)".
543
544 4. Partial matching support is added to the JIT compiler.
545
546 5. Fixed several bugs concerned with partial matching of items that consist
547 of more than one character:
548
549 (a) /^(..)\1/ did not partially match "aba" because checking references was
550 done on an "all or nothing" basis. This also applied to repeated
551 references.
552
553 (b) \R did not give a hard partial match if \r was found at the end of the
554 subject.
555
556 (c) \X did not give a hard partial match after matching one or more
557 characters at the end of the subject.
558
559 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
560 a partial match for the string "\r".
561
562 (e) When newline was set to CRLF, the metacharacter "." did not recognize
563 a partial match for a CR character at the end of the subject string.
564
565 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
566 running pcretest, the text "(JIT)" added to the output whenever JIT is
567 actually used to run the match.
568
569 7. Individual JIT compile options can be set in pcretest by following -s+[+]
570 or /S+[+] with a digit between 1 and 7.
571
572 8. OP_NOT now supports any UTF character not just single-byte ones.
573
574 9. (*MARK) control verb is now supported by the JIT compiler.
575
576 10. The command "./RunTest list" lists the available tests without actually
577 running any of them. (Because I keep forgetting what they all are.)
578
579 11. Add PCRE_INFO_MAXLOOKBEHIND.
580
581 12. Applied a (slightly modified) user-supplied patch that improves performance
582 when the heap is used for recursion (compiled with --disable-stack-for-
583 recursion). Instead of malloc and free for each heap frame each time a
584 logical recursion happens, frames are retained on a chain and re-used where
585 possible. This sometimes gives as much as 30% improvement.
586
587 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
588 call.
589
590 14. As documented, (*COMMIT) is now confined to within a positive assertion.
591
592 15. It is now possible to link pcretest with libedit as an alternative to
593 libreadline.
594
595 16. (*COMMIT) control verb is now supported by the JIT compiler.
596
597 17. The Unicode data tables have been updated to Unicode 6.1.0.
598
599 18. Added --file-list option to pcregrep.
600
601 19. Added binary file support to pcregrep, including the -a, --binary-files,
602 -I, and --text options.
603
604 20. The madvise function is renamed for posix_madvise for QNX compatibility
605 reasons. Fixed by Giuseppe D'Angelo.
606
607 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
608
609 22. Changed the option for creating symbolic links for 16-bit man pages from
610 -s to -sf so that re-installing does not cause issues.
611
612 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
613
614 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
615 matches in certain environments (the workspace was not being correctly
616 retained). Also added to pcre_dfa_exec() a simple plausibility check on
617 some of the workspace data at the beginning of a restart.
618
619 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
620 was not doing so when it should - probably a typo introduced by SVN 528
621 (change 8.10/14).
622
623 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
624 \w+ when the character tables indicated that \x{c4} was a word character.
625 There were several related cases, all because the tests for doing a table
626 lookup were testing for characters less than 127 instead of 255.
627
628 27. If a pattern contains capturing parentheses that are not used in a match,
629 their slots in the ovector are set to -1. For those that are higher than
630 any matched groups, this happens at the end of processing. In the case when
631 there were back references that the ovector was too small to contain
632 (causing temporary malloc'd memory to be used during matching), and the
633 highest capturing number was not used, memory off the end of the ovector
634 was incorrectly being set to -1. (It was using the size of the temporary
635 memory instead of the true size.)
636
637 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
638 ovector size, it uses memory at the end of the block that it has got.
639
640 29. Check for an overlong MARK name and give an error at compile time. The
641 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
642
643 30. JIT compiler update.
644
645 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
646 Rill for the patch.
647
648 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
649
650 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
651
652 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
653 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
654 was enabled.
655
656 35. Improve JIT code generation for greedy plus quantifier.
657
658 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
659 1 to "aa" instead of to an empty string. The bug affected repeated groups
660 that could potentially match an empty string.
661
662 37. Optimizing single character iterators in JIT.
663
664 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
665 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
666 codepoints that are too big for the mode are faulted, and in a UTF mode,
667 disallowed codepoints are also faulted.
668
669 39. If PCRE was compiled with UTF support, in three places in the DFA
670 matcher there was code that should only have been obeyed in UTF mode, but
671 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
672 processing when bytes with values greater than 127 were present. In 16-bit
673 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
674 both cases the values are those that cannot be the first data item in a UTF
675 character. The three items that might have provoked this were recursions,
676 possessively repeated groups, and atomic groups.
677
678 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
679 and pcregrep, because some OS require shared objects to be explicitly
680 passed to ld, causing the link step to fail if they are not.
681
682 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
683 mode, patterns that started with \h* or \R* might be incorrectly matched.
684
685
686 Version 8.30 04-February-2012
687 -----------------------------
688
689 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
690 name is defined in ctype.h.
691
692 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
693 only in quite long subpatterns.
694
695 3. Removed the function pcre_info(), which has been obsolete and deprecated
696 since it was replaced by pcre_fullinfo() in February 2000.
697
698 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
699 match a (*MARK), and the match failed at the start of the subject, a
700 reference to memory before the start of the subject could occur. This bug
701 was introduced by fix 17 of release 8.21.
702
703 5. A reference to an unset group with zero minimum repetition was giving
704 totally wrong answers (in non-JavaScript-compatibility mode). For example,
705 /(another)?(\1?)test/ matched against "hello world test". This bug was
706 introduced in release 8.13.
707
708 6. Add support for 16-bit character strings (a large amount of work involving
709 many changes and refactorings).
710
711 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
712 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
713 from a file.
714
715 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
716 rounding is not applied in this particular case).
717
718 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
719 if they appear, or are escaped, in patterns.
720
721 10. Get rid of a number of -Wunused-but-set-variable warnings.
722
723 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
724 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
725 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
726 also returns the mark "x". This bug applied to capturing parentheses,
727 non-capturing parentheses, and atomic parentheses. It also applied to some
728 assertions.
729
730 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
731 information out of configure.ac instead of relying on pcre.h.generic, which
732 is not stored in the repository.
733
734 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
735 -lreadline.
736
737 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
738
739 15. Applied Graycode's patch to put the top-level frame on the stack rather
740 than the heap when not using the stack for recursion. This gives a
741 performance improvement in many cases when recursion is not deep.
742
743 16. Experimental code added to "pcretest -C" to output the stack frame size.
744
745
746 Version 8.21 12-Dec-2011
747 ------------------------
748
749 1. Updating the JIT compiler.
750
751 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
752 are added as well.
753
754 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
755 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
756 calling _pcre_jit_exec. Some extra comments are added.
757
758 4. (*MARK) settings inside atomic groups that do not contain any capturing
759 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
760 was introduced by change 18 for 8.20.
761
762 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
763 ECMA-262 standard.
764
765 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
766 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
767 This bug was probably introduced by change 9 of 8.13.
768
769 7. While fixing 6 above, I noticed that a number of other items were being
770 incorrectly rejected as "not fixed length". This arose partly because newer
771 opcodes had not been added to the fixed-length checking code. I have (a)
772 corrected the bug and added tests for these items, and (b) arranged for an
773 error to occur if an unknown opcode is encountered while checking for fixed
774 length instead of just assuming "not fixed length". The items that were
775 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
776 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
777 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
778
779 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
780 being incorrectly compiled and would have given unpredicatble results.
781
782 9. A possessively repeated subpattern with minimum repeat count greater than
783 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
784 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
785 the first (A) could occur when it should not.
786
787 10. Add a cast and remove a redundant test from the code.
788
789 11. JIT should use pcre_malloc/pcre_free for allocation.
790
791 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
792 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
793 is anything other than /usr, -L is still shown).
794
795 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
796
797 14. Perl does not support \N without a following name in a [] class; PCRE now
798 also gives an error.
799
800 15. If a forward reference was repeated with an upper limit of around 2000,
801 it caused the error "internal error: overran compiling workspace". The
802 maximum number of forward references (including repeats) was limited by the
803 internal workspace, and dependent on the LINK_SIZE. The code has been
804 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
805 the default depends on LINK_SIZE. There is a new upper limit (for safety)
806 of around 200,000 forward references. While doing this, I also speeded up
807 the filling in of repeated forward references.
808
809 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
810 incorrectly expecting the subject to contain another "a" after the start.
811
812 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
813 in the match, the SKIP should be ignored. This was not happening; instead
814 the SKIP was being treated as NOMATCH. For patterns such as
815 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
816 tested.
817
818 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
819 now much more compatible with Perl, in particular in cases where the result
820 is a non-match for a non-anchored pattern. For example, if
821 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
822 "m", where previously it did not return a name. A side effect of this
823 change is that for partial matches, the last encountered mark name is
824 returned, as for non matches. A number of tests that were previously not
825 Perl-compatible have been moved into the Perl-compatible test files. The
826 refactoring has had the pleasing side effect of removing one argument from
827 the match() function, thus reducing its stack requirements.
828
829 19. If the /S+ option was used in pcretest to study a pattern using JIT,
830 subsequent uses of /S (without +) incorrectly behaved like /S+.
831
832 21. Retrieve executable code size support for the JIT compiler and fixing
833 some warnings.
834
835 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
836 not work when the shorter character appeared right at the end of the
837 subject string.
838
839 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
840 systems.
841
842 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
843 output it when the /M option is used in pcretest.
844
845 25. The CheckMan script was not being included in the distribution. Also, added
846 an explicit "perl" to run Perl scripts from the PrepareRelease script
847 because this is reportedly needed in Windows.
848
849 26. If study data was being save in a file and studying had not found a set of
850 "starts with" bytes for the pattern, the data written to the file (though
851 never used) was taken from uninitialized memory and so caused valgrind to
852 complain.
853
854 27. Updated RunTest.bat as provided by Sheri Pierce.
855
856 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
857
858 29. Computation of memory usage for the table of capturing group names was
859 giving an unnecessarily large value.
860
861
862 Version 8.20 21-Oct-2011
863 ------------------------
864
865 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
866 a POSIX class. After further experiments with Perl, which convinced me that
867 Perl has bugs and confusions, a closing square bracket is no longer allowed
868 in a POSIX name. This bug also affected patterns with classes that started
869 with full stops.
870
871 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
872 captured substring, but while checking the failing first alternative,
873 substring 1 is temporarily captured. If the output vector supplied to
874 pcre_exec() was not big enough for this capture, the yield of the function
875 was still zero ("insufficient space for captured substrings"). This cannot
876 be totally fixed without adding another stack variable, which seems a lot
877 of expense for a edge case. However, I have improved the situation in cases
878 such as /(a)(b)x|abc/ matched against "abc", where the return code
879 indicates that fewer than the maximum number of slots in the ovector have
880 been set.
881
882 3. Related to (2) above: when there are more back references in a pattern than
883 slots in the output vector, pcre_exec() uses temporary memory during
884 matching, and copies in the captures as far as possible afterwards. It was
885 using the entire output vector, but this conflicts with the specification
886 that only 2/3 is used for passing back captured substrings. Now it uses
887 only the first 2/3, for compatibility. This is, of course, another edge
888 case.
889
890 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
891 main code base, and can be used by building with --enable-jit. When this is
892 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
893 runtime --no-jit option is given.
894
895 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
896 ovector, the return from the function was zero, implying that there were
897 other matches that did not fit. The correct "exactly full" value is now
898 returned.
899
900 6. If a subpattern that was called recursively or as a subroutine contained
901 (*PRUNE) or any other control that caused it to give a non-standard return,
902 invalid errors such as "Error -26 (nested recursion at the same subject
903 position)" or even infinite loops could occur.
904
905 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
906 computing the minimum length on reaching *ACCEPT, and so ended up with the
907 wrong value of 1 rather than 0. Further investigation indicates that
908 computing a minimum subject length in the presence of *ACCEPT is difficult
909 (think back references, subroutine calls), and so I have changed the code
910 so that no minimum is registered for a pattern that contains *ACCEPT.
911
912 8. If (*THEN) was present in the first (true) branch of a conditional group,
913 it was not handled as intended. [But see 16 below.]
914
915 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
916 Sheri Pierce.
917
918 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
919 the first byte in a match must be "a".
920
921 11. Change 17 for 8.13 increased the recursion depth for patterns like
922 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
923 pattern contains any instances of (*THEN). If it does not, the old
924 optimizations are restored. It would be nice to do this on a per-group
925 basis, but at the moment that is not feasible.
926
927 12. In some environments, the output of pcretest -C is CRLF terminated. This
928 broke RunTest's code that checks for the link size. A single white space
929 character after the value is now allowed for.
930
931 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
932 For "fr", it uses the Windows-specific input and output files.
933
934 14. If (*THEN) appeared in a group that was called recursively or as a
935 subroutine, it did not work as intended. [But see next item.]
936
937 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
938 pattern fragments (but not containing any | characters). If A and B are
939 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
940 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
941 D. In other words, Perl considers parentheses that do not contain any |
942 characters to be part of a surrounding alternative, whereas PCRE was
943 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
944 differently. PCRE now behaves in the same way as Perl, except in the case
945 of subroutine/recursion calls such as (?1) which have in any case always
946 been different (but PCRE had them first :-).
947
948 16. Related to 15 above: Perl does not treat the | in a conditional group as
949 creating alternatives. Such a group is treated in the same way as an
950 ordinary group without any | characters when processing (*THEN). PCRE has
951 been changed to match Perl's behaviour.
952
953 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
954 RunGrepTest script failed.
955
956 18. Change 22 for version 13 caused atomic groups to use more stack. This is
957 inevitable for groups that contain captures, but it can lead to a lot of
958 stack use in large patterns. The old behaviour has been restored for atomic
959 groups that do not contain any capturing parentheses.
960
961 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
962 suppress the check for a minimum subject length at run time. (If it was
963 given to pcre_exec() or pcre_dfa_exec() it did work.)
964
965 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
966 fail to work when decoding hex characters in data strings in EBCDIC
967 environments.
968
969 21. It appears that in at least one Mac OS environment, the isxdigit() function
970 is implemented as a macro that evaluates to its argument more than once,
971 contravening the C 90 Standard (I haven't checked a later standard). There
972 was an instance in pcretest which caused it to go wrong when processing
973 \x{...} escapes in subject strings. The has been rewritten to avoid using
974 things like p++ in the argument of isxdigit().
975
976
977 Version 8.13 16-Aug-2011
978 ------------------------
979
980 1. The Unicode data tables have been updated to Unicode 6.0.0.
981
982 2. Two minor typos in pcre_internal.h have been fixed.
983
984 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
985 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
986 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
987
988 4. There were a number of related bugs in the code for matching backrefences
989 caselessly in UTF-8 mode when codes for the characters concerned were
990 different numbers of bytes. For example, U+023A and U+2C65 are an upper
991 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
992 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
993 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
994 2-byte code at the end of the subject (it thought there wasn't enough data
995 left).
996
997 5. Comprehensive information about what went wrong is now returned by
998 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
999 as the output vector has at least 2 elements. The offset of the start of
1000 the failing character and a reason code are placed in the vector.
1001
1002 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1003 now returned is for the first byte of the failing character, instead of the
1004 last byte inspected. This is an incompatible change, but I hope it is small
1005 enough not to be a problem. It makes the returned offset consistent with
1006 pcre_exec() and pcre_dfa_exec().
1007
1008 7. pcretest now gives a text phrase as well as the error number when
1009 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1010 failure, the offset and reason code are output.
1011
1012 8. When \R was used with a maximizing quantifier it failed to skip backwards
1013 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1014 back over a single character (\n). This seems wrong (because it treated the
1015 two characters as a single entity when going forwards), conflicts with the
1016 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1017 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1018 has been changed.
1019
1020 9. Some internal refactoring has changed the processing so that the handling
1021 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1022 time (the PCRE_DOTALL option was changed this way some time ago: version
1023 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1024 which was always a bit of a fudge. It also means that there is one less
1025 argument for the match() function, which reduces its stack requirements
1026 slightly. This change also fixes an incompatibility with Perl: the pattern
1027 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1028
1029 10. More internal refactoring has drastically reduced the number of recursive
1030 calls to match() for possessively repeated groups such as (abc)++ when
1031 using pcre_exec().
1032
1033 11. While implementing 10, a number of bugs in the handling of groups were
1034 discovered and fixed:
1035
1036 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1037 (a|)*(?1) gave a compile-time internal error.
1038 ((a|)+)+ did not notice that the outer group could match an empty string.
1039 (^a|^)+ was not marked as anchored.
1040 (.*a|.*)+ was not marked as matching at start or after a newline.
1041
1042 12. Yet more internal refactoring has removed another argument from the match()
1043 function. Special calls to this function are now indicated by setting a
1044 value in a variable in the "match data" data block.
1045
1046 13. Be more explicit in pcre_study() instead of relying on "default" for
1047 opcodes that mean there is no starting character; this means that when new
1048 ones are added and accidentally left out of pcre_study(), testing should
1049 pick them up.
1050
1051 14. The -s option of pcretest has been documented for ages as being an old
1052 synonym of -m (show memory usage). I have changed it to mean "force study
1053 for every regex", that is, assume /S for every regex. This is similar to -i
1054 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1055 using it. It makes it easier to run collections of tests with and without
1056 study enabled, and thereby test pcre_study() more easily. All the standard
1057 tests are now run with and without -s (but some patterns can be marked as
1058 "never study" - see 20 below).
1059
1060 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1061 restoration of the capturing data to the outer values was not happening
1062 correctly.
1063
1064 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1065 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1066 pattern had matched an empty string, and so incorrectly returned a no
1067 match.
1068
1069 17. There was optimizing code for the last branch of non-capturing parentheses,
1070 and also for the obeyed branch of a conditional subexpression, which used
1071 tail recursion to cut down on stack usage. Unfortunately, now that there is
1072 the possibility of (*THEN) occurring in these branches, tail recursion is
1073 no longer possible because the return has to be checked for (*THEN). These
1074 two optimizations have therefore been removed. [But see 8.20/11 above.]
1075
1076 18. If a pattern containing \R was studied, it was assumed that \R always
1077 matched two bytes, thus causing the minimum subject length to be
1078 incorrectly computed because \R can also match just one byte.
1079
1080 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1081 was incorrectly computed.
1082
1083 20. If /S is present twice on a test pattern in pcretest input, it now
1084 *disables* studying, thereby overriding the use of -s on the command line
1085 (see 14 above). This is necessary for one or two tests to keep the output
1086 identical in both cases.
1087
1088 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1089 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1090
1091 22. When an atomic group that contained a capturing parenthesis was
1092 successfully matched, but the branch in which it appeared failed, the
1093 capturing was not being forgotten if a higher numbered group was later
1094 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1095 group 1 to "a", when in fact it should be unset. This applied to multi-
1096 branched capturing and non-capturing groups, repeated or not, and also to
1097 positive assertions (capturing in negative assertions does not happen
1098 in PCRE) and also to nested atomic groups.
1099
1100 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1101 subject after a captured substring, to make it easier to tell which of a
1102 number of identical substrings has been captured.
1103
1104 24. The way atomic groups are processed by pcre_exec() has been changed so that
1105 if they are repeated, backtracking one repetition now resets captured
1106 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1107 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1108 "aaa". Previously, it would have been "a". As part of this code
1109 refactoring, the way recursive calls are handled has also been changed.
1110
1111 25. If an assertion condition captured any substrings, they were not passed
1112 back unless some other capturing happened later. For example, if
1113 (?(?=(a))a) was matched against "a", no capturing was returned.
1114
1115 26. When studying a pattern that contained subroutine calls or assertions,
1116 the code for finding the minimum length of a possible match was handling
1117 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1118 group 1 called group 2 while simultaneously a separate group 2 called group
1119 1). A stack overflow occurred in this case. I have fixed this by limiting
1120 the recursion depth to 10.
1121
1122 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1123 Fortmann. This supports explicit test numbers on the command line, and has
1124 argument validation and error reporting.
1125
1126 28. An instance of \X with an unlimited repeat could fail if at any point the
1127 first character it looked at was a mark character.
1128
1129 29. Some minor code refactoring concerning Unicode properties and scripts
1130 should reduce the stack requirement of match() slightly.
1131
1132 30. Added the '=' option to pcretest to check the setting of unused capturing
1133 slots at the end of the pattern, which are documented as being -1, but are
1134 not included in the return count.
1135
1136 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1137 compiled something random. Now it gives a compile-time error (as does
1138 Perl).
1139
1140 32. A *MARK encountered during the processing of a positive assertion is now
1141 recorded and passed back (compatible with Perl).
1142
1143 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1144 had alternative anchored branches, the search for a second match in a line
1145 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1146 matched the line "0102" twice. The same bug affected patterns that started
1147 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1148 twice.
1149
1150 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1151 does, and because of capturing effects, quantifying parenthesized
1152 assertions may at times be useful. Quantifiers are now allowed for
1153 parenthesized assertions.
1154
1155 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1156
1157 36. \g was being checked for fancy things in a character class, when it should
1158 just be a literal "g".
1159
1160 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1161 appearance of a nested POSIX class supersedes an apparent external class.
1162 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1163 unescaped square brackets may also appear as part of class names. For
1164 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1165 more like Perl. (But see 8.20/1 above.)
1166
1167 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1168 was because it thought it was \N{name}, which is not supported).
1169
1170 39. Add minix to OS list not supporting the -S option in pcretest.
1171
1172 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1173 cannot analyze patterns in sufficient detail to catch mutual recursions
1174 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1175 subgroup is called recursively as a subpattern for a second time at the
1176 same position in the subject string. In previous releases this might have
1177 been caught by the recursion limit, or it might have run out of stack.
1178
1179 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1180 happen only once. PCRE was, however incorrectly giving a compile time error
1181 "recursive call could loop indefinitely" because it cannot analyze the
1182 pattern in sufficient detail. The compile time test no longer happens when
1183 PCRE is compiling a conditional subpattern, but actual runaway loops are
1184 now caught at runtime (see 40 above).
1185
1186 42. It seems that Perl allows any characters other than a closing parenthesis
1187 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1188 has been changed to be the same.
1189
1190 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1191 as not to get warnings when autogen.sh is called. Also changed
1192 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1193
1194 44. To help people who use pcregrep to scan files containing exceedingly long
1195 lines, the following changes have been made:
1196
1197 (a) The default value of the buffer size parameter has been increased from
1198 8K to 20K. (The actual buffer used is three times this size.)
1199
1200 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1201 PCRE is built.
1202
1203 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1204 to be set at run time.
1205
1206 (d) Numerical values in pcregrep options can be followed by K or M, for
1207 example --buffer-size=50K.
1208
1209 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1210 given and the return code is set to 2.
1211
1212 45. Add a pointer to the latest mark to the callout data block.
1213
1214 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1215 partial match of an empty string instead of no match. This was specific to
1216 the use of ".".
1217
1218 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1219 complete match instead of a partial match. This bug was dependent on both
1220 the PCRE_UTF8 and PCRE_DOTALL options being set.
1221
1222 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1223 starting byte set, because \b was not being ignored.
1224
1225
1226 Version 8.12 15-Jan-2011
1227 ------------------------
1228
1229 1. Fixed some typos in the markup of the man pages, and wrote a script that
1230 checks for such things as part of the documentation building process.
1231
1232 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1233 --match-limit and --recursion-limit options (added for 8.11). In
1234 particular, this made one of the standard tests fail. (The integer value
1235 went into the wrong half of a long int.)
1236
1237 3. If the --colour option was given to pcregrep with -v (invert match), it
1238 did strange things, either producing crazy output, or crashing. It should,
1239 of course, ignore a request for colour when reporting lines that do not
1240 match.
1241
1242 4. Another pcregrep bug caused similar problems if --colour was specified with
1243 -M (multiline) and the pattern match finished with a line ending.
1244
1245 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1246 matched in multiline mode, the following line was shown as part of the
1247 match. This seems wrong, so I have changed it.
1248
1249 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1250 the check for further matches in the same line (so they could be coloured)
1251 to overrun the end of the current line. If another match was found, it was
1252 incorrectly shown (and then shown again when found in the next line).
1253
1254 7. If pcregrep was compiled under Windows, there was a reference to the
1255 function pcregrep_exit() before it was defined. I am assuming this was
1256 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1257 reported by a user. I've moved the definition above the reference.
1258
1259
1260 Version 8.11 10-Dec-2010
1261 ------------------------
1262
1263 1. (*THEN) was not working properly if there were untried alternatives prior
1264 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1265 backtracked to try for "b" instead of moving to the next alternative branch
1266 at the same level (in this case, to look for "c"). The Perl documentation
1267 is clear that when (*THEN) is backtracked onto, it goes to the "next
1268 alternative in the innermost enclosing group".
1269
1270 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1271 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1272 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1273 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1274 (*THEN).
1275
1276 3. If \s appeared in a character class, it removed the VT character from
1277 the class, even if it had been included by some previous item, for example
1278 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1279 of \s, but is part of the POSIX "space" class.)
1280
1281 4. A partial match never returns an empty string (because you can always
1282 match an empty string at the end of the subject); however the checking for
1283 an empty string was starting at the "start of match" point. This has been
1284 changed to the "earliest inspected character" point, because the returned
1285 data for a partial match starts at this character. This means that, for
1286 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1287 (previously it gave "no match").
1288
1289 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1290 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1291 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1292 has an implication that the given string is incomplete (because a partial
1293 match is preferred over a full match). For this reason, these items now
1294 give a partial match in this situation. [Aside: previously, the one case
1295 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1296 match rather than a full match, which was wrong by the old rules, but is
1297 now correct.]
1298
1299 6. There was a bug in the handling of #-introduced comments, recognized when
1300 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1301 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1302 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1303 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1304 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1305 places in pcre_compile().
1306
1307 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1308 comments when looking ahead for named forward references to subpatterns,
1309 the only newline sequence it recognized was NL. It now handles newlines
1310 according to the set newline convention.
1311
1312 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1313 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1314 cater for a lack of strerror(). These oversights have been fixed.
1315
1316 9. Added --match-limit and --recursion-limit to pcregrep.
1317
1318 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1319
1320 11. When the -o option was used, pcregrep was setting a return code of 1, even
1321 when matches were found, and --line-buffered was not being honoured.
1322
1323 12. Added an optional parentheses number to the -o and --only-matching options
1324 of pcregrep.
1325
1326 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1327 can match an empty string. The code to do it in pcretest and pcredemo
1328 needed fixing:
1329
1330 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1331 only one byte after an empty string match just before CRLF (this case
1332 just got forgotten; "any" and "anycrlf" were OK).
1333
1334 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1335 mode when an empty string match preceded an ASCII character followed by
1336 a non-ASCII character. (The code for advancing by one character rather
1337 than one byte was nonsense.)
1338
1339 (c) The pcredemo.c sample program did not have any code at all to handle
1340 the cases when CRLF is a valid newline sequence.
1341
1342 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1343 as a starting offset was within the subject string. There is now a new
1344 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1345 negative or greater than the length of the string. In order to test this,
1346 pcretest is extended to allow the setting of negative starting offsets.
1347
1348 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1349 starting offset points to the beginning of a UTF-8 character was
1350 unnecessarily clumsy. I tidied it up.
1351
1352 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1353 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1354
1355 17. Nobody had reported that the --include_dir option, which was added in
1356 release 7.7 should have been called --include-dir (hyphen, not underscore)
1357 for compatibility with GNU grep. I have changed it to --include-dir, but
1358 left --include_dir as an undocumented synonym, and the same for
1359 --exclude-dir, though that is not available in GNU grep, at least as of
1360 release 2.5.4.
1361
1362 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1363 characters from a string of bytes) have been redefined so as not to use
1364 loops, in order to improve performance in some environments. At the same
1365 time, I abstracted some of the common code into auxiliary macros to save
1366 repetition (this should not affect the compiled code).
1367
1368 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1369 compile-time error is now given if \c is not followed by an ASCII
1370 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1371 different, and any byte value is allowed.)
1372
1373 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1374 START_OPTIMIZE option, which is now allowed at compile time - but just
1375 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1376 to pcregrep and other applications that have no direct access to PCRE
1377 options. The new /Y option in pcretest sets this option when calling
1378 pcre_compile().
1379
1380 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1381 back references. Groups containing recursive back references were forced to
1382 be atomic by that change, but in the case of named groups, the amount of
1383 memory required was incorrectly computed, leading to "Failed: internal
1384 error: code overflow". This has been fixed.
1385
1386 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1387 pcretest.c, to avoid build problems in some Borland environments.
1388
1389
1390 Version 8.10 25-Jun-2010
1391 ------------------------
1392
1393 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1394 THEN.
1395
1396 2. (*ACCEPT) was not working when inside an atomic group.
1397
1398 3. Inside a character class, \B is treated as a literal by default, but
1399 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1400 causes the error). The code is unchanged, but I tidied the documentation.
1401
1402 4. Inside a character class, PCRE always treated \R and \X as literals,
1403 whereas Perl faults them if its -w option is set. I have changed PCRE so
1404 that it faults them when PCRE_EXTRA is set.
1405
1406 5. Added support for \N, which always matches any character other than
1407 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1408
1409 6. When compiling pcregrep with newer versions of gcc which may have
1410 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1411 declared with attribute warn_unused_result" were given. Just casting the
1412 result to (void) does not stop the warnings; a more elaborate fudge is
1413 needed. I've used a macro to implement this.
1414
1415 7. Minor change to pcretest.c to avoid a compiler warning.
1416
1417 8. Added four artifical Unicode properties to help with an option to make
1418 \s etc use properties (see next item). The new properties are: Xan
1419 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1420
1421 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1422 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1423 this option. Modified pcretest to add /W to test this facility. Added
1424 REG_UCP to make it available via the POSIX interface.
1425
1426 10. Added --line-buffered to pcregrep.
1427
1428 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1429 studied, and the match started with a letter with a code point greater than
1430 127 whose first byte was different to the first byte of the other case of
1431 the letter, the other case of this starting letter was not recognized
1432 (#976).
1433
1434 12. If a pattern that was studied started with a repeated Unicode property
1435 test, for example, \p{Nd}+, there was the theoretical possibility of
1436 setting up an incorrect bitmap of starting bytes, but fortunately it could
1437 not have actually happened in practice until change 8 above was made (it
1438 added property types that matched character-matching opcodes).
1439
1440 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1441 possible starting bytes for non-anchored patterns.
1442
1443 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1444 \R, and also a number of cases that involve Unicode properties, both
1445 explicit and implicit when PCRE_UCP is set.
1446
1447 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1448 input, it could crash or give wrong results if characters with values
1449 greater than 0xc0 were present in the subject string. (Detail: it assumed
1450 UTF-8 input when processing these items.)
1451
1452 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1453 size_t is 64-bit (#991).
1454
1455 17. Added a check for running out of memory when PCRE is compiled with
1456 --disable-stack-for-recursion (#990).
1457
1458 18. If the last data line in a file for pcretest does not have a newline on
1459 the end, a newline was missing in the output.
1460
1461 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1462 less than 128) in its various bitmaps. However, there is a facility for
1463 generating tables according to the current locale when PCRE is compiled. It
1464 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1465 characters, are recognized by isspace() and therefore were getting set in
1466 these tables, and indeed these tables seem to approximate to ISO 8859. This
1467 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1468 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1469 which of course cannot start UTF-8 characters. I have changed the code so
1470 that only real ASCII characters (less than 128) and the correct starting
1471 bytes for UTF-8 encodings are set for characters greater than 127 when in
1472 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1473 altogether.)
1474
1475 20. Added the /T option to pcretest so as to be able to run tests with non-
1476 standard character tables, thus making it possible to include the tests
1477 used for 19 above in the standard set of tests.
1478
1479 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1480 reference to a subpattern the other side of a comment that contains an
1481 opening parenthesis caused either an internal compiling error, or a
1482 reference to the wrong subpattern.
1483
1484
1485 Version 8.02 19-Mar-2010
1486 ------------------------
1487
1488 1. The Unicode data tables have been updated to Unicode 5.2.0.
1489
1490 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1491 configured.
1492
1493 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1494 original author of that file, following a query about its status.
1495
1496 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1497 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1498
1499 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1500 quantifier applied to a forward-referencing subroutine call, could compile
1501 incorrect code or give the error "internal error: previously-checked
1502 referenced subpattern not found".
1503
1504 6. Both MS Visual Studio and Symbian OS have problems with initializing
1505 variables to point to external functions. For these systems, therefore,
1506 pcre_malloc etc. are now initialized to local functions that call the
1507 relevant global functions.
1508
1509 7. There were two entries missing in the vectors called coptable and poptable
1510 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1511 I've fixed the data, and added a kludgy way of testing at compile time that
1512 the lengths are correct (equal to the number of opcodes).
1513
1514 8. Following on from 7, I added a similar kludge to check the length of the
1515 eint vector in pcreposix.c.
1516
1517 9. Error texts for pcre_compile() are held as one long string to avoid too
1518 much relocation at load time. To find a text, the string is searched,
1519 counting zeros. There was no check for running off the end of the string,
1520 which could happen if a new error number was added without updating the
1521 string.
1522
1523 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1524
1525 11. \K was not working if it appeared in an atomic group or in a group that
1526 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1527 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1528 the assertion is positive, but not if it is negative.
1529
1530 12. Change 11 fortuitously reduced the size of the stack frame used in the
1531 "match()" function of pcre_exec.c by one pointer. Forthcoming
1532 implementation of support for (*MARK) will need an extra pointer on the
1533 stack; I have reserved it now, so that the stack frame size does not
1534 decrease.
1535
1536 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1537 item in branch that calls a recursion is a subroutine call - as in the
1538 second branch in the above example - was incorrectly given the compile-
1539 time error "recursive call could loop indefinitely" because pcre_compile()
1540 was not correctly checking the subroutine for matching a non-empty string.
1541
1542 14. The checks for overrunning compiling workspace could trigger after an
1543 overrun had occurred. This is a "should never occur" error, but it can be
1544 triggered by pathological patterns such as hundreds of nested parentheses.
1545 The checks now trigger 100 bytes before the end of the workspace.
1546
1547 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1548
1549
1550 Version 8.01 19-Jan-2010
1551 ------------------------
1552
1553 1. If a pattern contained a conditional subpattern with only one branch (in
1554 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1555 computed the wrong minimum data length (which is of course zero for such
1556 subpatterns). This could cause incorrect "no match" results.
1557
1558 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1559 the pattern is reset in the first branch, pcre_compile() failed with
1560 "internal error: code overflow at offset...". This happened only when
1561 the reset was to the original external option setting. (An optimization
1562 abstracts leading options settings into an external setting, which was the
1563 cause of this.)
1564
1565 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1566 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1567 assertion pattern did not match (meaning that the assertion was true), it
1568 was incorrectly treated as false if the SKIP had been reached during the
1569 matching. This also applied to assertions used as conditions.
1570
1571 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1572 assertion subpattern, including such a pattern used as a condition,
1573 unpredictable results occurred, instead of the error return
1574 PCRE_ERROR_DFA_UITEM.
1575
1576 5. The C++ GlobalReplace function was not working like Perl for the special
1577 situation when an empty string is matched. It now does the fancy magic
1578 stuff that is necessary.
1579
1580 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1581 removed. (These were left over from very, very early versions of PCRE.)
1582
1583 7. Some cosmetic changes to the code to make life easier when compiling it
1584 as part of something else:
1585
1586 (a) Change DEBUG to PCRE_DEBUG.
1587
1588 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1589 called "current" as "current_branch", to prevent a collision with the
1590 Linux macro when compiled as a kernel module.
1591
1592 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1593 prevent a collision with the Linux macro when compiled as a kernel
1594 module.
1595
1596 8. In pcre_compile() there are some checks for integer overflows that used to
1597 cast potentially large values to (double). This has been changed to that
1598 when building, a check for int64_t is made, and if it is found, it is used
1599 instead, thus avoiding the use of floating point arithmetic. (There is no
1600 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1601 double.
1602
1603 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1604 2005 (difference between two addresses compared to an unsigned value).
1605
1606 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1607 custom one, because of the following reported problem in Windows:
1608
1609 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1610 under Win32.
1611 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1612 therefore missing the function definition.
1613 - The compiler thus generates a "C" signature for the test function.
1614 - The linker fails to find the "C" function.
1615 - PCRE fails to configure if asked to do so against libbz2.
1616
1617 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1618 messages were output:
1619
1620 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1621 rerunning libtoolize, to keep the correct libtool macros in-tree.
1622 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1623
1624 I have done both of these things.
1625
1626 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1627 most of the time, it *can* run out if it is given a pattern that contains a
1628 runaway infinite recursion. I updated the discussion in the pcrestack man
1629 page.
1630
1631 13. Now that we have gone to the x.xx style of version numbers, the minor
1632 version may start with zero. Using 08 or 09 is a bad idea because users
1633 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1634 interpreted as invalid octal numbers. I've updated the previous comment in
1635 configure.ac, and also added a check that gives an error if 08 or 09 are
1636 used.
1637
1638 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1639 causing partial matching to fail when the end of the subject matched \W
1640 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1641
1642 15. There were some discrepancies between the declarations in pcre_internal.h
1643 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1644 their definitions. The declarations used "const uschar *" and the
1645 definitions used USPTR. Even though USPTR is normally defined as "const
1646 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1647 reported that: "This difference in casting confuses some C++ compilers, for
1648 example, SunCC recognizes above declarations as different functions and
1649 generates broken code for hbpcre." I have changed the declarations to use
1650 USPTR.
1651
1652 16. GNU libtool is named differently on some systems. The autogen.sh script now
1653 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1654 (FreeBSD).
1655
1656 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1657 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1658 comment: "Figure out how to create a longlong from a string: strtoll and
1659 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1660 instance, but it only takes 2 args instead of 3!"
1661
1662 18. A subtle bug concerned with back references has been fixed by a change of
1663 specification, with a corresponding code fix. A pattern such as
1664 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1665 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1666 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1667 same bug. Such groups have to be quantified to be useful, or contained
1668 inside another quantified group. (If there's no repetition, the reference
1669 can never match.) The problem arises because, having left the group and
1670 moved on to the rest of the pattern, a later failure that backtracks into
1671 the group uses the captured value from the final iteration of the group
1672 rather than the correct earlier one. I have fixed this in PCRE by forcing
1673 any group that contains a reference to itself to be an atomic group; that
1674 is, there cannot be any backtracking into it once it has completed. This is
1675 similar to recursive and subroutine calls.
1676
1677
1678 Version 8.00 19-Oct-09
1679 ----------------------
1680
1681 1. The table for translating pcre_compile() error codes into POSIX error codes
1682 was out-of-date, and there was no check on the pcre_compile() error code
1683 being within the table. This could lead to an OK return being given in
1684 error.
1685
1686 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1687 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1688 in a Windows environment.
1689
1690 3. The pcregrep --count option prints the count for each file even when it is
1691 zero, as does GNU grep. However, pcregrep was also printing all files when
1692 --files-with-matches was added. Now, when both options are given, it prints
1693 counts only for those files that have at least one match. (GNU grep just
1694 prints the file name in this circumstance, but including the count seems
1695 more useful - otherwise, why use --count?) Also ensured that the
1696 combination -clh just lists non-zero counts, with no names.
1697
1698 4. The long form of the pcregrep -F option was incorrectly implemented as
1699 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1700 but it seems right to fix it, and I didn't think it was worth preserving
1701 the old behaviour.
1702
1703 5. The command line items --regex=pattern and --regexp=pattern were not
1704 recognized by pcregrep, which required --regex pattern or --regexp pattern
1705 (with a space rather than an '='). The man page documented the '=' forms,
1706 which are compatible with GNU grep; these now work.
1707
1708 6. No libpcreposix.pc file was created for pkg-config; there was just
1709 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1710
1711 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1712 when UCP support is not needed, by modifying the Python script that
1713 generates it from Unicode data files. This should not matter if the module
1714 is correctly used as a library, but I received one complaint about 50K of
1715 unwanted data. My guess is that the person linked everything into his
1716 program rather than using a library. Anyway, it does no harm.
1717
1718 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1719 was a minimum greater than 1 for a wide character in a possessive
1720 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1721 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1722 character. Chaos in the form of incorrect output or a compiling loop could
1723 result.
1724
1725 9. The restrictions on what a pattern can contain when partial matching is
1726 requested for pcre_exec() have been removed. All patterns can now be
1727 partially matched by this function. In addition, if there are at least two
1728 slots in the offset vector, the offset of the earliest inspected character
1729 for the match and the offset of the end of the subject are set in them when
1730 PCRE_ERROR_PARTIAL is returned.
1731
1732 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1733 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1734 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1735 and may be more useful for multi-segment matching.
1736
1737 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1738 used to be given if ever the end of the subject was reached; now it is
1739 given only if matching could not proceed because another character was
1740 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1741 string "Z", which now yields "no match" instead of "partial match". In the
1742 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1743 final character ended with (*FAIL).
1744
1745 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1746 if the pattern had a "must contain" character that was already found in the
1747 earlier partial match, unless partial matching was again requested. For
1748 example, with the pattern /dog.(body)?/, the "must contain" character is
1749 "g". If the first part-match was for the string "dog", restarting with
1750 "sbody" failed. This bug has been fixed.
1751
1752 13. The string returned by pcre_dfa_exec() after a partial match has been
1753 changed so that it starts at the first inspected character rather than the
1754 first character of the match. This makes a difference only if the pattern
1755 starts with a lookbehind assertion or \b or \B (\K is not supported by
1756 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1757 matching functions compatible, and I think it's the right thing to do.
1758
1759 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1760 so that the demonstration program is easily available in environments where
1761 PCRE has not been installed from source.
1762
1763 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1764 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1765 library.
1766
1767 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1768 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1769 is not the first non-POSIX option to be added. Clearly some people find
1770 these options useful.
1771
1772 17. If a caller to the POSIX matching function regexec() passes a non-zero
1773 value for nmatch with a NULL value for pmatch, the value of
1774 nmatch is forced to zero.
1775
1776 18. RunGrepTest did not have a test for the availability of the -u option of
1777 the diff command, as RunTest does. It now checks in the same way as
1778 RunTest, and also checks for the -b option.
1779
1780 19. If an odd number of negated classes containing just a single character
1781 interposed, within parentheses, between a forward reference to a named
1782 subpattern and the definition of the subpattern, compilation crashed with
1783 an internal error, complaining that it could not find the referenced
1784 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1785 [The bug was that it was starting one character too far in when skipping
1786 over the character class, thus treating the ] as data rather than
1787 terminating the class. This meant it could skip too much.]
1788
1789 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1790 /g option in pcretest when the pattern contains \K, which makes it possible
1791 to have an empty string match not at the start, even when the pattern is
1792 anchored. Updated pcretest and pcredemo to use this option.
1793
1794 21. If the maximum number of capturing subpatterns in a recursion was greater
1795 than the maximum at the outer level, the higher number was returned, but
1796 with unset values at the outer level. The correct (outer level) value is
1797 now given.
1798
1799 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1800 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1801 make it do so. The string so far is captured, making this feature
1802 compatible with Perl.
1803
1804 23. The tests have been re-organized, adding tests 11 and 12, to make it
1805 possible to check the Perl 5.10 features against Perl 5.10.
1806
1807 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1808 pattern matches a fixed length string. PCRE did not allow this; now it
1809 does. Neither allows recursion.
1810
1811 25. I finally figured out how to implement a request to provide the minimum
1812 length of subject string that was needed in order to match a given pattern.
1813 (It was back references and recursion that I had previously got hung up
1814 on.) This code has now been added to pcre_study(); it finds a lower bound
1815 to the length of subject needed. It is not necessarily the greatest lower
1816 bound, but using it to avoid searching strings that are too short does give
1817 some useful speed-ups. The value is available to calling programs via
1818 pcre_fullinfo().
1819
1820 26. While implementing 25, I discovered to my embarrassment that pcretest had
1821 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1822 study optimizations had never been tested with that matching function.
1823 Oops. What is worse, even when it was passed study data, there was a bug in
1824 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1825 were also very few tests of studied patterns with pcre_dfa_exec().
1826
1827 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1828 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1829 on the other side of the coin, they are no longer allowed to have different
1830 names, because these cannot be distinguished in PCRE, and this has caused
1831 confusion. (This is a difference from Perl.)
1832
1833 28. When duplicate subpattern names are present (necessarily with different
1834 numbers, as required by 27 above), and a test is made by name in a
1835 conditional pattern, either for a subpattern having been matched, or for
1836 recursion in such a pattern, all the associated numbered subpatterns are
1837 tested, and the overall condition is true if the condition is true for any
1838 one of them. This is the way Perl works, and is also more like the way
1839 testing by number works.
1840
1841
1842 Version 7.9 11-Apr-09
1843 ---------------------
1844
1845 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1846 (pcretest), all targets were linked against these libraries. This included
1847 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1848 libraries. This caused unwanted dependencies to be created. This problem
1849 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1850 pcretest is linked with readline.
1851
1852 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1853 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1854 moved outside it again, because FALSE and TRUE are already defined in AIX,
1855 but BOOL is not.
1856
1857 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1858 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1859
1860 4. The pcregrep documentation said spaces were inserted as well as colons (or
1861 hyphens) following file names and line numbers when outputting matching
1862 lines. This is not true; no spaces are inserted. I have also clarified the
1863 wording for the --colour (or --color) option.
1864
1865 5. In pcregrep, when --colour was used with -o, the list of matching strings
1866 was not coloured; this is different to GNU grep, so I have changed it to be
1867 the same.
1868
1869 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1870 each matching line was coloured. Now it goes on to look for further matches
1871 of any of the test patterns, which is the same behaviour as GNU grep.
1872
1873 7. A pattern that could match an empty string could cause pcregrep to loop; it
1874 doesn't make sense to accept an empty string match in pcregrep, so I have
1875 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1876 seems to be how GNU grep behaves. [But see later change 40 for release
1877 8.33.]
1878
1879 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1880 start or after a newline", because the conditional assertion was not being
1881 correctly handled. The rule now is that both the assertion and what follows
1882 in the first alternative must satisfy the test.
1883
1884 9. If auto-callout was enabled in a pattern with a conditional group whose
1885 condition was an assertion, PCRE could crash during matching, both with
1886 pcre_exec() and pcre_dfa_exec().
1887
1888 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1889 used for matching.
1890
1891 11. Unicode property support in character classes was not working for
1892 characters (bytes) greater than 127 when not in UTF-8 mode.
1893
1894 12. Added the -M command line option to pcretest.
1895
1896 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1897
1898 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1899
1900 16. Added comments and documentation about mis-use of no_arg in the C++
1901 wrapper.
1902
1903 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1904 from Martin Jerabek that uses macro names for all relevant character and
1905 string constants.
1906
1907 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1908 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1909 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1910 these, but not everybody uses configure.
1911
1912 19. A conditional group that had only one branch was not being correctly
1913 recognized as an item that could match an empty string. This meant that an
1914 enclosing group might also not be so recognized, causing infinite looping
1915 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1916 with the subject "ab", where knowledge that the repeated group can match
1917 nothing is needed in order to break the loop.
1918
1919 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1920 exec(), but without supplying a callout function, matching went wrong.
1921
1922 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1923 leak if the size of the offset vector was greater than 30. When the vector
1924 is smaller, the saved offsets during recursion go onto a local stack
1925 vector, but for larger vectors malloc() is used. It was failing to free
1926 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1927 error, in fact).
1928
1929 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1930 heapframe that is used only when UTF-8 support is enabled. This caused no
1931 problem, but was untidy.
1932
1933 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1934 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1935 included within another project.
1936
1937 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1938 slightly modified by me:
1939
1940 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1941 not building pcregrep.
1942
1943 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1944 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1945
1946 25. Forward references, both numeric and by name, in patterns that made use of
1947 duplicate group numbers, could behave incorrectly or give incorrect errors,
1948 because when scanning forward to find the reference group, PCRE was not
1949 taking into account the duplicate group numbers. A pattern such as
1950 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1951
1952 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1953 the feature of a custom pointer more persuasive (as requested by a user).
1954
1955 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1956 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1957 pre-defined.
1958
1959 28. Added support for (*UTF8) at the start of a pattern.
1960
1961 29. Arrange for flags added by the "release type" setting in CMake to be shown
1962 in the configuration summary.
1963
1964
1965 Version 7.8 05-Sep-08
1966 ---------------------
1967
1968 1. Replaced UCP searching code with optimized version as implemented for Ad
1969 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1970 stage table and inline lookup instead of a function, giving speed ups of 2
1971 to 5 times on some simple patterns that I tested. Permission was given to
1972 distribute the MultiStage2.py script that generates the tables (it's not in
1973 the tarball, but is in the Subversion repository).
1974
1975 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1976 scripts.
1977
1978 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1979 a group with a zero qualifier. The result of the study could be incorrect,
1980 or the function might crash, depending on the pattern.
1981
1982 4. Caseless matching was not working for non-ASCII characters in back
1983 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1984 It now works when Unicode Property Support is available.
1985
1986 5. In pcretest, an escape such as \x{de} in the data was always generating
1987 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1988 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1989 truncation.
1990
1991 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1992
1993 7. Added two (int) casts to pcregrep when printing the difference of two
1994 pointers, in case they are 64-bit values.
1995
1996 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1997 test 2 if it fails.
1998
1999 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2000 and a #define of that name to empty if it is not externally set. This is to
2001 allow users of MSVC to set it if necessary.
2002
2003 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2004 the convenience functions in the pcre_get.c source file.
2005
2006 11. An option change at the start of a pattern that had top-level alternatives
2007 could cause overwriting and/or a crash. This command provoked a crash in
2008 some environments:
2009
2010 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2011
2012 This potential security problem was recorded as CVE-2008-2371.
2013
2014 12. For a pattern where the match had to start at the beginning or immediately
2015 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2016 pcre_dfa_exec() could read past the end of the passed subject if there was
2017 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2018 pcretest so that it places the subject at the end of its malloc-ed buffer.
2019
2020 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2021 exec() might read past the end of the data buffer in UTF-8 mode.
2022
2023 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2024 the data contained the byte 0x85 as part of a UTF-8 character within its
2025 first line. This applied both to normal and DFA matching.
2026
2027 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2028 /^[^d]*?$/8 failed to match "abc".
2029
2030 16. Added a missing copyright notice to pcrecpp_internal.h.
2031
2032 17. Make it more clear in the documentation that values returned from
2033 pcre_exec() in ovector are byte offsets, not character counts.
2034
2035 18. Tidied a few places to stop certain compilers from issuing warnings.
2036
2037 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2038 supplied by Stefan Weber. I made a further small update for 7.8 because
2039 there is a change of source arrangements: the pcre_searchfuncs.c module is
2040 replaced by pcre_ucd.c.
2041
2042
2043 Version 7.7 07-May-08
2044 ---------------------
2045
2046 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2047 a string to a long long, pretend we don't even have a long long." This is
2048 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2049
2050 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2051 pre-7.6 versions, which defined a global no_arg variable instead of putting
2052 it in the RE class. (See also #8 below.)
2053
2054 3. Remove a line of dead code, identified by coverity and reported by Nuno
2055 Lopes.
2056
2057 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2058
2059 (1) The include/exclude patterns were being applied to the whole pathnames
2060 of files, instead of just to the final components.
2061
2062 (2) If there was more than one level of directory, the subdirectories were
2063 skipped unless they satisfied the include/exclude conditions. This is
2064 inconsistent with GNU grep (and could even be seen as contrary to the
2065 pcregrep specification - which I improved to make it absolutely clear).
2066 The action now is always to scan all levels of directory, and just
2067 apply the include/exclude patterns to regular files.
2068
2069 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2070 --exclude_dir in the tests to avoid scanning .svn directories.
2071
2072 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2073 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2074 doesn't support NULs in patterns.
2075
2076 7. Added some missing "const"s to declarations of static tables in
2077 pcre_compile.c and pcre_dfa_exec.c.
2078
2079 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2080 caused by fix #2 above. (Subsequently also a second patch to fix the
2081 first patch. And a third patch - this was a messy problem.)
2082
2083 9. Applied Craig's patch to remove the use of push_back().
2084
2085 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2086 matching function regexec().
2087
2088 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2089 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2090 references. PCRE supports relative numbers with this syntax (I don't think
2091 Oniguruma does).
2092
2093 12. Previously, a group with a zero repeat such as (...){0} was completely
2094 omitted from the compiled regex. However, this means that if the group
2095 was called as a subroutine from elsewhere in the pattern, things went wrong
2096 (an internal error was given). Such groups are now left in the compiled
2097 pattern, with a new opcode that causes them to be skipped at execution
2098 time.
2099
2100 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2101 to the way PCRE behaves:
2102
2103 (a) A lone ] character is dis-allowed (Perl treats it as data).
2104
2105 (b) A back reference to an unmatched subpattern matches an empty string
2106 (Perl fails the current match path).
2107
2108 (c) A data ] in a character class must be notated as \] because if the
2109 first data character in a class is ], it defines an empty class. (In
2110 Perl it is not possible to have an empty class.) The empty class []
2111 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2112 The negative empty class [^] matches any one character, independently
2113 of the DOTALL setting.
2114
2115 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2116 non-existent subpattern following a character class starting with ']' and
2117 containing () gave an internal compiling error instead of "reference to
2118 non-existent subpattern". Fortunately, when the pattern did exist, the
2119 compiled code was correct. (When scanning forwards to check for the
2120 existence of the subpattern, it was treating the data ']' as terminating
2121 the class, so got the count wrong. When actually compiling, the reference
2122 was subsequently set up correctly.)
2123
2124 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2125 it was being rejected as not supported by pcre_dfa_exec(), even though
2126 other assertions are supported. I have made pcre_dfa_exec() support
2127 (*FAIL).
2128
2129 16. The implementation of 13c above involved the invention of a new opcode,
2130 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2131 cannot be changed at match time, I realized I could make a small
2132 improvement to matching performance by compiling OP_ALLANY instead of
2133 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2134 on the OP_ANY path.
2135
2136 17. Compiling pcretest on Windows with readline support failed without the
2137 following two fixes: (1) Make the unistd.h include conditional on
2138 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2139
2140 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2141 ncurses library to be included for pcretest when ReadLine support is
2142 requested, but also to allow for it to be overridden. This patch came from
2143 Daniel Bergström.
2144
2145 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2146 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2147 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2148 spotting this.
2149
2150
2151 Version 7.6 28-Jan-08
2152 ---------------------
2153
2154 1. A character class containing a very large number of characters with
2155 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2156 overflow.
2157
2158 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2159 HAVE_LONG_LONG is not defined.
2160
2161 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2162 bring them up to date and include new features. This patch includes:
2163
2164 - Fixed PH's badly added libz and libbz2 support.
2165 - Fixed a problem with static linking.
2166 - Added pcredemo. [But later removed - see 7 below.]
2167 - Fixed dftables problem and added an option.
2168 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2169 HAVE_LONG_LONG.
2170 - Added readline support for pcretest.
2171 - Added an listing of the option settings after cmake has run.
2172
2173 4. A user submitted a patch to Makefile that makes it easy to create
2174 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2175 Makefile.am that cause it to include this special target, without
2176 affecting anything else. Note that the same mingw target plus all
2177 the other distribution libraries and programs are now supported
2178 when configuring with CMake (see 6 below) instead of with
2179 Configure/Make.
2180
2181 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2182 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2183 exported in the Windows port". It has not yet been confirmed that the patch
2184 solves the problem, but it does no harm.
2185
2186 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2187 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2188 with CMake, and also correct the comment about stack recursion.
2189
2190 7. Remove the automatic building of pcredemo from the ./configure system and
2191 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2192 of a program that users should build themselves after PCRE is installed, so
2193 building it automatically is not really right. What is more, it gave
2194 trouble in some build environments.
2195
2196 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2197
2198
2199 Version 7.5 10-Jan-08
2200 ---------------------
2201
2202 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2203 values in parens when parsing an RE using the C++ wrapper."
2204
2205 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2206 Characters greater than 255 were excluded from the class instead of being
2207 included.
2208
2209 3. The same bug as (2) above applied to negated POSIX classes such as
2210 [:^space:].
2211
2212 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2213 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2214 I have changed it.
2215
2216 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2217 first named subpattern) and a construct such as (?&a) would reference the
2218 first named subpattern whose name started with "a" (in other words, the
2219 length check was missing). Both these problems are fixed. "Subpattern name
2220 expected" is now given for (?&) (a zero-length name), and this patch also
2221 makes it give the same error for \k'' (previously it complained that that
2222 was a reference to a non-existent subpattern).
2223
2224 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2225 this is right because (?- can be followed by option settings as well as by
2226 digits. I have, however, made the messages clearer.
2227
2228 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2229 than the number used in the conditional) now cause a compile-time error.
2230 This is actually not compatible with Perl, which accepts such patterns, but
2231 treats the conditional as always being FALSE (as PCRE used to), but it
2232 seems to me that giving a diagnostic is better.
2233
2234 8. Change "alphameric" to the more common word "alphanumeric" in comments
2235 and messages.
2236
2237 9. Fix two occurrences of "backslash" in comments that should have been
2238 "backspace".
2239
2240 10. Remove two redundant lines of code that can never be obeyed (their function
2241 was moved elsewhere).
2242
2243 11. The program that makes PCRE's Unicode character property table had a bug
2244 which caused it to generate incorrect table entries for sequences of
2245 characters that have the same character type, but are in different scripts.
2246 It amalgamated them into a single range, with the script of the first of
2247 them. In other words, some characters were in the wrong script. There were
2248 thirteen such cases, affecting characters in the following ranges:
2249
2250 U+002b0 - U+002c1
2251 U+0060c - U+0060d
2252 U+0061e - U+00612
2253 U+0064b - U+0065e
2254 U+0074d - U+0076d
2255 U+01800 - U+01805
2256 U+01d00 - U+01d77
2257 U+01d9b - U+01dbf
2258 U+0200b - U+0200f
2259 U+030fc - U+030fe
2260 U+03260 - U+0327f
2261 U+0fb46 - U+0fbb1
2262 U+10450 - U+1049d
2263
2264 12. The -o option (show only the matching part of a line) for pcregrep was not
2265 compatible with GNU grep in that, if there was more than one match in a
2266 line, it showed only the first of them. It now behaves in the same way as
2267 GNU grep.
2268
2269 13. If the -o and -v options were combined for pcregrep, it printed a blank
2270 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2271 does the same. The return code can be used to tell if there were any
2272 non-matching lines.
2273
2274 14. Added --file-offsets and --line-offsets to pcregrep.
2275
2276 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2277 infinitely looping recursion. The bug was that positive lookaheads were not
2278 being skipped when checking for a possible empty match (negative lookaheads
2279 and both kinds of lookbehind were skipped).
2280
2281 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2282 inclusion of <windows.h> to before rather than after the definition of
2283 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2284
2285 17. Specifying a possessive quantifier with a specific limit for a Unicode
2286 character property caused pcre_compile() to compile bad code, which led at
2287 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2288 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2289 caused the error; without that there was no problem.
2290
2291 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2292
2293 19. Added --enable-pcretest-libreadline.
2294
2295 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2296 RE::GlobalReplace(). As a result, the number of replacements returned was
2297 double what it should be. I removed one of the increments, but Craig sent a
2298 later patch that removed the other one (the right fix) and added unit tests
2299 that check the return values (which was not done before).
2300
2301 21. Several CMake things:
2302
2303 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2304 the names libpcre and libpcreposix, not just pcre and pcreposix.
2305
2306 (2) The above change means that pcretest and pcregrep are now correctly
2307 linked with the newly-built libraries, not previously installed ones.
2308
2309 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2310
2311 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2312 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2313 UTF-8 newline character). The key issue is that the pattern starts .*;
2314 this means that the match must be either at the beginning, or after a
2315 newline. The bug was in the code for advancing after a failed match and
2316 checking that the new position followed a newline. It was not taking
2317 account of UTF-8 characters correctly.
2318
2319 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2320 character classes. PCRE was not treating the sequence [:...:] as a
2321 character class unless the ... were all letters. Perl, however, seems to
2322 allow any characters between [: and :], though of course it rejects as
2323 unknown any "names" that contain non-letters, because all the known class
2324 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2325 for example, whereas PCRE did not - it did not recognize a POSIX character
2326 class. This seemed a bit dangerous, so the code has been changed to be
2327 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2328 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2329 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2330 Perl does, and where it didn't before.
2331
2332 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2333 Windows environments %n is disabled by default.
2334
2335
2336 Version 7.4 21-Sep-07
2337 ---------------------
2338
2339 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2340 means that a class such as [\s] counted as "explicit reference to CR or
2341 LF". That isn't really right - the whole point of the change was to try to
2342 help when there was an actual mention of one of the two characters. So now
2343 the change happens only if \r or \n (or a literal CR or LF) character is
2344 encountered.
2345
2346 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2347 of both had grown to the point where there were only 3 bits left.
2348 Fortunately, there was spare space in the data structure, and so I have
2349 moved the internal flags into a new 16-bit field to free up more option
2350 bits.
2351
2352 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2353 but did not set the internal JCHANGED flag - either of these is enough to
2354 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2355 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2356 start sets both bits.
2357
2358 4. Added options (at build time, compile time, exec time) to change \R from
2359 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2360
2361 5. doc/pcresyntax.html was missing from the distribution.
2362
2363 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2364 compatibility, even though it is no longer used.
2365
2366 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2367 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2368 windows.h file is present (where different names are used). [This was
2369 reversed later after testing - see 16 below.]
2370
2371 8. Changed all #include <config.h> to #include "config.h". There were also
2372 some further <pcre.h> cases that I changed to "pcre.h".
2373
2374 9. When pcregrep was used with the --colour option, it missed the line ending
2375 sequence off the lines that it output.
2376
2377 10. It was pointed out to me that arrays of string pointers cause lots of
2378 relocations when a shared library is dynamically loaded. A technique of
2379 using a single long string with a table of offsets can drastically reduce
2380 these. I have refactored PCRE in four places to do this. The result is
2381 dramatic:
2382
2383 Originally: 290
2384 After changing UCP table: 187
2385 After changing error message table: 43
2386 After changing table of "verbs" 36
2387 After changing table of Posix names 22
2388
2389 Thanks to the folks working on Gregex for glib for this insight.
2390
2391 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2392 unicode-properties was also set.
2393
2394 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2395
2396 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2397 checked only for CRLF.
2398
2399 14. Added casts to pcretest.c to avoid compiler warnings.
2400
2401 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2402
2403 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2404 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2405 entirely. This removes changes made in 7 above.
2406
2407 17. The CMake files have been updated, and there is now more information about
2408 building with CMake in the NON-UNIX-USE document.
2409
2410
2411 Version 7.3 28-Aug-07
2412 ---------------------
2413
2414 1. In the rejigging of the build system that eventually resulted in 7.1, the
2415 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2416 brackets there is not right, since it causes compilers to look for an
2417 installed pcre.h, not the version that is in the source that is being
2418 compiled (which of course may be different). I have changed it back to:
2419
2420 #include "pcre.h"
2421
2422 I have a vague recollection that the change was concerned with compiling in
2423 different directories, but in the new build system, that is taken care of
2424 by the VPATH setting the Makefile.
2425
2426 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2427 when the subject happened to end in the byte 0x85 (e.g. if the last
2428 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2429 characters but of course it shouldn't be taken as a newline when it is part
2430 of another character. The bug was that, for an unlimited repeat of . in
2431 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2432 characters when looking for a newline.
2433
2434 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2435
2436 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2437 in debug output.
2438
2439 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2440 long printing in the pcrecpp unittest when running under MinGW.
2441
2442 6. ESC_K was left out of the EBCDIC table.
2443
2444 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2445 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2446 limit also applies to "virtual nesting" when a pattern is recursive, and in
2447 this case 1000 isn't so big. I have been able to remove this limit at the
2448 expense of backing off one optimization in certain circumstances. Normally,
2449 when pcre_exec() would call its internal match() function recursively and
2450 immediately return the result unconditionally, it uses a "tail recursion"
2451 feature to save stack. However, when a subpattern that can match an empty
2452 string has an unlimited repetition quantifier, it no longer makes this
2453 optimization. That gives it a stack frame in which to save the data for
2454 checking that an empty string has been matched. Previously this was taken
2455 from the 1000-entry workspace that had been reserved. So now there is no
2456 explicit limit, but more stack is used.
2457
2458 8. Applied Daniel's patches to solve problems with the import/export magic
2459 syntax that is required for Windows, and which was going wrong for the
2460 pcreposix and pcrecpp parts of the library. These were overlooked when this
2461 problem was solved for the main library.
2462
2463 9. There were some crude static tests to avoid integer overflow when computing
2464 the size of patterns that contain repeated groups with explicit upper
2465 limits. As the maximum quantifier is 65535, the maximum group length was
2466 set at 30,000 so that the product of these two numbers did not overflow a
2467 32-bit integer. However, it turns out that people want to use groups that
2468 are longer than 30,000 bytes (though not repeat them that many times).
2469 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2470 made it possible to implement the integer overflow checks in a much more
2471 dynamic way, which I have now done. The artificial limitation on group
2472 length has been removed - we now have only the limit on the total length of
2473 the compiled pattern, which depends on the LINK_SIZE setting.
2474
2475 10. Fixed a bug in the documentation for get/copy named substring when
2476 duplicate names are permitted. If none of the named substrings are set, the
2477 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2478 empty string.
2479
2480 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2481 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2482 because the ] is interpreted as the first data character and the
2483 terminating ] is not found. PCRE has been made compatible with Perl in this
2484 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2485 cause memory overwriting.
2486
2487 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2488 string has been matched (to stop an infinite loop). It was not recognizing
2489 a conditional subpattern that could match an empty string if that
2490 subpattern was within another subpattern. For example, it looped when
2491 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2492 condition was not nested. This bug has been fixed.
2493
2494 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2495 past the start of the subject in the presence of bytes with the top bit
2496 set, for example "\x8aBCD".
2497
2498 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2499 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2500
2501 14. Optimized (?!) to (*FAIL).
2502
2503 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2504 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2505 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2506 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2507 does: it's just the validity check that is more restrictive.
2508
2509 16. Inserted checks for integer overflows during escape sequence (backslash)
2510 processing, and also fixed erroneous offset values for syntax errors during
2511 backslash processing.
2512
2513 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2514 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2515
2516 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2517 caused an overrun.
2518
2519 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2520 something other than just ASCII characters) inside a group that had an
2521 unlimited repeat caused a loop at compile time (while checking to see
2522 whether the group could match an empty string).
2523
2524 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2525 [\P{Any}] did so. (Error in the code for printing property names.)
2526
2527 21. An orphan \E inside a character class could cause a crash.
2528
2529 22. A repeated capturing bracket such as (A)? could cause a wild memory
2530 reference during compilation.
2531
2532 23. There are several functions in pcre_compile() that scan along a compiled
2533 expression for various reasons (e.g. to see if it's fixed length for look
2534 behind). There were bugs in these functions when a repeated \p or \P was
2535 present in the pattern. These operators have additional parameters compared
2536 with \d, etc, and these were not being taken into account when moving along
2537 the compiled data. Specifically:
2538
2539 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2540 length.
2541
2542 (b) An item such as \pL+ within a repeated group could cause crashes or
2543 loops.
2544
2545 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2546 "reference to non-existent subpattern" error.
2547
2548 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2549
2550 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2551 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2552
2553 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2554
2555 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2556 character were causing crashes (broken optimization).
2557
2558 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2559 \p or \P) caused a compile-time loop.
2560
2561 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2562 break. For example, the unstudied pattern [\r\n]A does not match the string
2563 "\r\nA" because change 7.0/46 below moves the current point on by two
2564 characters after failing to match at the start. However, the pattern \nA
2565 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2566 the same is true. There doesn't seem any very clean way out of this, but
2567 what I have chosen to do makes the common cases work: PCRE now takes note
2568 of whether there can be an explicit match for \r or \n anywhere in the
2569 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2570 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2571 pattern has explicit CR or LF references.
2572
2573 29. Added (*CR) etc for changing newline setting at start of pattern.
2574
2575
2576 Version 7.2 19-Jun-07
2577 ---------------------
2578
2579 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2580 which is apparently normally available under Windows.
2581
2582 2. Re-jig the pcregrep tests with different newline settings in an attempt
2583 to make them independent of the local environment's newline setting.
2584
2585 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2586
2587 4. Some of the "internals" tests were previously cut out when the link size
2588 was not 2, because the output contained actual offsets. The recent new
2589 "Z" feature of pcretest means that these can be cut out, making the tests
2590 usable with all link sizes.
2591
2592 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2593 stack recursion. This gives a massive performance boost under BSD, but just
2594 a small improvement under Linux. However, it saves one field in the frame
2595 in all cases.
2596
2597 6. Added more features from the forthcoming Perl 5.10:
2598
2599 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2600 recursion call. It refers to the nth most recently opened parentheses.
2601
2602 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2603 to be opened parentheses.
2604
2605 (c) Conditions that refer to capturing parentheses can be specified
2606 relatively, for example, (?(-2)... or (?(+3)...
2607
2608 (d) \K resets the start of the current match so that everything before
2609 is not part of it.
2610
2611 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2612
2613 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2614 reference syntax.
2615
2616 (g) (?| introduces a group in which the numbering of parentheses in each
2617 alternative starts with the same number.
2618
2619 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2620
2621 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2622 PCRE_INFO_JCHANGED.
2623
2624 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2625 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2626 for detecting groups that can match an empty string.
2627
2628 9. A pattern with a very large number of alternatives (more than several
2629 hundred) was running out of internal workspace during the pre-compile
2630 phase, where pcre_compile() figures out how much memory will be needed. A
2631 bit of new cunning has reduced the workspace needed for groups with
2632 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2633 workspace instead of running out of the 4096 that are available.
2634
2635 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2636
2637 11. Applied patch from Google to remove an optimization that didn't quite work.
2638 The report of the bug said:
2639
2640 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2641 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2642 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2643
2644 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2645 it matched the wrong number of bytes.
2646
2647
2648 Version 7.1 24-Apr-07
2649 ---------------------
2650
2651 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2652 that is more "standard", making use of automake and other Autotools. There
2653 is some re-arrangement of the files and adjustment of comments consequent
2654 on this.
2655
2656 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2657 for recursive directory scanning broke on some systems because the files
2658 are not scanned in any specific order and on different systems the order
2659 was different. A call to "sort" has been inserted into RunGrepTest for the
2660 approprate test as a short-term fix. In the longer term there may be an
2661 alternative.
2662
2663 3. I had an email from Eric Raymond about problems translating some of PCRE's
2664 man pages to HTML (despite the fact that I distribute HTML pages, some
2665 people do their own conversions for various reasons). The problems
2666 concerned the use of low-level troff macros .br and .in. I have therefore
2667 removed all such uses from the man pages (some were redundant, some could
2668 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2669 HTML has been updated to handle .nf/.fi and to complain if it encounters
2670 .br or .in.
2671
2672 4. Updated comments in configure.ac that get placed in config.h.in and also
2673 arranged for config.h to be included in the distribution, with the name
2674 config.h.generic, for the benefit of those who have to compile without
2675 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2676
2677 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2678 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2679 makevp.bat for the current PCRE, using the additional files
2680 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2681
2682 6. A Windows user reported a minor discrepancy with test 2, which turned out
2683 to be caused by a trailing space on an input line that had got lost in his
2684 copy. The trailing space was an accident, so I've just removed it.
2685
2686 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2687 that is needed.
2688
2689 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2690 as "const" (a) because they are and (b) because it helps the PHP
2691 maintainers who have recently made a script to detect big data structures
2692 in the php code that should be moved to the .rodata section. I remembered
2693 to update Builducptable as well, so it won't revert if ucptable.h is ever
2694 re-created.
2695
2696 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2697 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2698 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2699 support is not required. This saves 1.5-2K of code, which is important in
2700 some applications.
2701
2702 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2703 so as not to refer to the tables, even though these functions will never be
2704 called when UTF-8 support is disabled. Otherwise there are problems with a
2705 shared library.
2706
2707 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2708
2709 (a) It was defining its arguments as char * instead of void *.
2710
2711 (b) It was assuming that all moves were upwards in memory; this was true
2712 a long time ago when I wrote it, but is no longer the case.
2713
2714 The emulated memove() is provided for those environments that have neither
2715 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2716 is clearly not the case, as these two bugs were recently reported.
2717
2718 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2719 and Detrail to create the HTML documentation, the .txt form of the man
2720 pages, and it removes trailing spaces from listed files. It also creates
2721 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2722 case, it wraps all the #defines with #ifndefs. This script should be run
2723 before "make dist".
2724
2725 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2726 with Unicode property support.
2727
2728 (a) For a maximizing quantifier, if the two different cases of the
2729 character were of different lengths in their UTF-8 codings (there are
2730 some cases like this - I found 11), and the matching function had to
2731 back up over a mixture of the two cases, it incorrectly assumed they
2732 were both the same length.
2733
2734 (b) When PCRE was configured to use the heap rather than the stack for
2735 recursion during matching, it was not correctly preserving the data for
2736 the other case of a UTF-8 character when checking ahead for a match
2737 while processing a minimizing repeat. If the check also involved
2738 matching a wide character, but failed, corruption could cause an
2739 erroneous result when trying to check for a repeat of the original
2740 character.
2741
2742 13. Some tidying changes to the testing mechanism:
2743
2744 (a) The RunTest script now detects the internal link size and whether there
2745 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2746 values substituted by "configure". (The RunGrepTest script already did
2747 this for UTF-8.) The configure.ac script no longer substitutes the
2748 relevant variables.
2749
2750 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2751 with length and offset values. This means that the output is different
2752 for different internal link sizes. Test 2 is skipped for link sizes
2753 other than 2 because of this, bypassing the problem. Unfortunately,
2754 there was also a test in test 3 (the locale tests) that used /B and
2755 failed for link sizes other than 2. Rather than cut the whole test out,
2756 I have added a new /Z option to pcretest that replaces the length and
2757 offset values with spaces. This is now used to make test 3 independent
2758 of link size. (Test 2 will be tidied up later.)
2759
2760 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2761 segmentation fault instead of returning the appropriate error message.
2762
2763 15. In multiline mode when the newline sequence was set to "any", the pattern
2764 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2765 This doesn't seem right; it now treats the CRLF combination as the line
2766 ending, and so does not match in that case. It's only a pattern such as ^$
2767 that would hit this one: something like ^ABC$ would have failed after \r
2768 and then tried again after \r\n.
2769
2770 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2771 in an attempt to make files that differ only in their line terminators
2772 compare equal. This works on Linux.
2773
2774 17. Under certain error circumstances pcregrep might try to free random memory
2775 as it exited. This is now fixed, thanks to valgrind.
2776
2777 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2778 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2779 was because its rules for how to advance for /g after matching an empty
2780 string at the end of a line did not allow for this case. They now check for
2781 it specially.
2782
2783 20. pcretest is supposed to handle patterns and data of any length, by
2784 extending its buffers when necessary. It was getting this wrong when the
2785 buffer for a data line had to be extended.
2786
2787 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2788 CRLF as a newline sequence.
2789
2790 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2791 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2792 I have nevertheless tidied it up.
2793
2794 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2795
2796 24. Added a man page for pcre-config.
2797
2798
2799 Version 7.0 19-Dec-06
2800 ---------------------
2801
2802 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2803 moving to gcc 4.1.1.
2804
2805 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2806 sys/time.h, which is documented as needed for this function. It doesn't
2807 seem to matter on Linux, but it showed up on some releases of OS X.
2808
2809 3. It seems that there are systems where bytes whose values are greater than
2810 127 match isprint() in the "C" locale. The "C" locale should be the
2811 default when a C program starts up. In most systems, only ASCII printing
2812 characters match isprint(). This difference caused the output from pcretest
2813 to vary, making some of the tests fail. I have changed pcretest so that:
2814
2815 (a) When it is outputting text in the compiled version of a pattern, bytes
2816 other than 32-126 are always shown as hex escapes.
2817
2818 (b) When it is outputting text that is a matched part of a subject string,
2819 it does the same, unless a different locale has been set for the match
2820 (using the /L modifier). In this case, it uses isprint() to decide.
2821
2822 4. Fixed a major bug that caused incorrect computation of the amount of memory
2823 required for a compiled pattern when options that changed within the
2824 pattern affected the logic of the preliminary scan that determines the
2825 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2826 that the computed length was too small. The symptoms of this bug were
2827 either the PCRE error "internal error: code overflow" from pcre_compile(),
2828 or a glibc crash with a message such as "pcretest: free(): invalid next
2829 size (fast)". Examples of patterns that provoked this bug (shown in
2830 pcretest format) are:
2831
2832 /(?-x: )/x
2833 /(?x)(?-x: \s*#\s*)/
2834 /((?i)[\x{c0}])/8
2835 /(?i:[\x{c0}])/8
2836
2837 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2838 is now done differently.
2839
2840 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2841 wrapper classes; (b) implement a new function in the C++ scanner that is
2842 more efficient than the old way of doing things because it avoids levels of
2843 recursion in the regex matching; (c) add a paragraph to the documentation
2844 for the FullMatch() function.
2845
2846 6. The escape sequence \n was being treated as whatever was defined as
2847 "newline". Not only was this contrary to the documentation, which states
2848 that \n is character 10 (hex 0A), but it also went horribly wrong when
2849 "newline" was defined as CRLF. This has been fixed.
2850
2851 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2852 was being set to -1 for the "end of line" case (supposedly a value that no
2853 character can have). Though this value is never used (the check for end of
2854 line is "zero bytes in current character"), it caused compiler complaints.
2855 I've changed it to 0xffffffff.
2856
2857 8. In pcre_version.c, the version string was being built by a sequence of
2858 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2859 string (as it is for production releases) called a macro with an empty
2860 argument. The C standard says the result of this is undefined. The gcc
2861 compiler treats it as an empty string (which was what was wanted) but it is
2862 reported that Visual C gives an error. The source has been hacked around to
2863 avoid this problem.
2864
2865 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2866 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2867 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2868 of them did).
2869
2870 10. Originally, pcretest opened its input and output without "b"; then I was
2871 told that "b" was needed in some environments, so it was added for release
2872 5.0 to both the input and output. (It makes no difference on Unix-like
2873 systems.) Later I was told that it is wrong for the input on Windows. I've
2874 now abstracted the modes into two macros, to make it easier to fiddle with
2875 them, and removed "b" from the input mode under Windows.
2876
2877 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2878
2879 12. Added -help and --help to pcretest as an official way of being reminded
2880 of the options.
2881
2882 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2883 and pcrecpp.cc because they annoy compilers at high warning levels.
2884
2885 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2886
2887 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2888 scripts are not C programs :-) and which was not noticed because it works
2889 on Linux.
2890
2891 16. pcretest is supposed to handle any length of pattern and data line (as one
2892 line or as a continued sequence of lines) by extending its input buffer if
2893 necessary. This feature was broken for very long pattern lines, leading to
2894 a string of junk being passed to pcre_compile() if the pattern was longer
2895 than about 50K.
2896
2897 17. I have done a major re-factoring of the way pcre_compile() computes the
2898 amount of memory needed for a compiled pattern. Previously, there was code
2899 that made a preliminary scan of the pattern in order to do this. That was
2900 OK when PCRE was new, but as the facilities have expanded, it has become
2901 harder and harder to keep it in step with the real compile phase, and there
2902 have been a number of bugs (see for example, 4 above). I have now found a
2903 cunning way of running the real compile function in a "fake" mode that
2904 enables it to compute how much memory it would need, while actually only
2905 ever using a few hundred bytes of working memory and without too many
2906 tests of the mode. This should make future maintenance and development
2907 easier. A side effect of this work is that the limit of 200 on the nesting
2908 depth of parentheses has been removed (though this was never a serious
2909 limitation, I suspect). However, there is a downside: pcre_compile() now
2910 runs more slowly than before (30% or more, depending on the pattern). I
2911 hope this isn't a big issue. There is no effect on runtime performance.
2912
2913 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2914 newline (only possible for the last line of a file) and it was a
2915 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2916
2917 19. Added additional timing features to pcretest. (1) The -tm option now times
2918 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2919 separate command line item, by a number that specifies the number of
2920 repeats to use when timing. The default is 50000; this gives better
2921 precision, but takes uncomfortably long for very large patterns.
2922
2923 20. Extended pcre_study() to be more clever in cases where a branch of a
2924 subpattern has no definite first character. For example, (a*|b*)[cd] would
2925 previously give no result from pcre_study(). Now it recognizes that the
2926 first character must be a, b, c, or d.
2927
2928 21. There was an incorrect error "recursive call could loop indefinitely" if
2929 a subpattern (or the entire pattern) that was being tested for matching an
2930 empty string contained only one non-empty item after a nested subpattern.
2931 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2932 incorrectly, because the \d was being skipped in the check.
2933
2934 22. The pcretest program now has a new pattern option /B and a command line
2935 option -b, which is equivalent to adding /B to every pattern. This causes
2936 it to show the compiled bytecode, without the additional information that
2937 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2938 is the same as /B/I).
2939
2940 23. A new optimization is now able automatically to treat some sequences such
2941 as a*b as a*+b. More specifically, if something simple (such as a character
2942 or a simple class like \d) has an unlimited quantifier, and is followed by
2943 something that cannot possibly match the quantified thing, the quantifier
2944 is automatically "possessified".
2945
2946 24. A recursive reference to a subpattern whose number was greater than 39
2947 went wrong under certain circumstances in UTF-8 mode. This bug could also
2948 have affected the operation of pcre_study().
2949
2950 25. Realized that a little bit of performance could be had by replacing
2951 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2952
2953 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2954
2955 27. Possessive quantifiers such as a++ were previously implemented by turning
2956 them into atomic groups such as ($>a+). Now they have their own opcodes,
2957 which improves performance. This includes the automatically created ones
2958 from 23 above.
2959
2960 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2961 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2962 the first matched character to be a colon. This applied both to named and
2963 numbered groups.
2964
2965 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2966
2967 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2968 building PCRE on Windows easier, so I have included it in the distribution.
2969
2970 31. There is now a check in pcretest against a ridiculously large number being
2971 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2972 loop, the loop is abandoned.
2973
2974 32. Forward references to subpatterns in conditions such as (?(2)...) where
2975 subpattern 2 is defined later cause pcre_compile() to search forwards in
2976 the pattern for the relevant set of parentheses. This search went wrong
2977 when there were unescaped parentheses in a character class, parentheses
2978 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2979
2980 33. "Subroutine" calls and backreferences were previously restricted to
2981 referencing subpatterns earlier in the regex. This restriction has now
2982 been removed.
2983
2984 34. Added a number of extra features that are going to be in Perl 5.10. On the
2985 whole, these are just syntactic alternatives for features that PCRE had
2986 previously implemented using the Python syntax or my own invention. The
2987 other formats are all retained for compatibility.
2988
2989 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2990 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2991 also .NET compatible.
2992
2993 (b) A recursion or subroutine call to a named group can now be defined as
2994 (?&name) as well as (?P>name).
2995
2996 (c) A backreference to a named group can now be defined as \k<name> or
2997 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2998 5.10, are also .NET compatible.
2999
3000 (d) A conditional reference to a named group can now use the syntax
3001 (?(<name>) or (?('name') as well as (?(name).
3002
3003 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3004 groups (named and numbered) that are never evaluated inline, but can be
3005 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3006 is always false. There may be only one alternative in such a group.
3007
3008 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3009 as the simple (?(R). The condition is true only if the most recent
3010 recursion is that of the given number or name. It does not search out
3011 through the entire recursion stack.
3012
3013 (g) The escape \gN or \g{N} has been added, where N is a positive or
3014 negative number, specifying an absolute or relative reference.
3015
3016 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3017 some "unreachable code" warnings.
3018
3019 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3020 things, this adds five new scripts.
3021
3022 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3023 There were also incompatibilities regarding the handling of \Q..\E inside
3024 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3025 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3026
3027 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3028 matches an empty string, and forcibly breaks the loop. There were bugs in
3029 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3030 against aaaa the result was just "a" rather than "aaaa", for example. Two
3031 separate and independent bugs (that affected different cases) have been
3032 fixed.
3033
3034 39. Refactored the code to abolish the use of different opcodes for small
3035 capturing bracket numbers. This is a tidy that I avoided doing when I
3036 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3037 The new approach is not only tidier, it makes it possible to reduce the
3038 memory needed to fix the previous bug (38).
3039
3040 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3041 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3042 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3043 mode.
3044
3045 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3046 report.
3047
3048 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3049 copy construction and assignment in the C++ wrapper.
3050
3051 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3052 couple of bugs that could have given wrong results in the "--newline=crlf"
3053 case.
3054
3055 44. Added a number of casts and did some reorganization of signed/unsigned int
3056 variables following suggestions from Dair Grant. Also renamed the variable
3057 "this" as "item" because it is a C++ keyword.
3058
3059 45. Arranged for dftables to add
3060
3061 #include "pcre_internal.h"
3062
3063 to pcre_chartables.c because without it, gcc 4.x may remove the array
3064 definition from the final binary if PCRE is built into a static library and
3065 dead code stripping is activated.
3066
3067 46. For an unanchored pattern, if a match attempt fails at the start of a
3068 newline sequence, and the newline setting is CRLF or ANY, and the next two
3069 characters are CRLF, advance by two characters instead of one.
3070
3071
3072 Version 6.7 04-Jul-06
3073 ---------------------
3074
3075 1. In order to handle tests when input lines are enormously long, pcretest has
3076 been re-factored so that it automatically extends its buffers when
3077 necessary. The code is crude, but this _is_ just a test program. The
3078 default size has been increased from 32K to 50K.
3079
3080 2. The code in pcre_study() was using the value of the re argument before
3081 testing it for NULL. (Of course, in any sensible call of the function, it
3082 won't be NULL.)
3083
3084 3. The memmove() emulation function in pcre_internal.h, which is used on
3085 systems that lack both memmove() and bcopy() - that is, hardly ever -
3086 was missing a "static" storage class specifier.
3087
3088 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3089 containing an extended class (one that cannot be represented by a bitmap
3090 because it contains high-valued characters or Unicode property items, e.g.
3091 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3092 pattern, but PCRE should not loop if you do not (it no longer does).
3093 [Detail: two cases were found: (a) a repeated subpattern containing an
3094 extended class; (b) a recursive reference to a subpattern that followed a
3095 previous extended class. It wasn't skipping over the extended class
3096 correctly when UTF-8 mode was not set.]
3097
3098 5. A negated single-character class was not being recognized as fixed-length
3099 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3100 compile error "lookbehind assertion is not fixed length".
3101
3102 6. The RunPerlTest auxiliary script was showing an unexpected difference
3103 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3104 write a Perl script that can interpret lines of an input file either as
3105 byte characters or as UTF-8, which is what "perltest" was being required to
3106 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3107 can't do is switch easily at run time between having the "use utf8;" pragma
3108 or not. In the end, I fudged it by using the RunPerlTest script to insert
3109 "use utf8;" explicitly for the UTF-8 tests.
3110
3111 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3112 the end of the subject string, contrary to the documentation and to what
3113 Perl does. This was true of both matching functions. Now it matches only at
3114 the start of the subject and immediately after *internal* newlines.
3115
3116 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3117 a pointer to an int instead of a pointer to an unsigned long int. This
3118 caused problems on 64-bit systems.
3119
3120 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3121 instance of the 'standard' template library not being so standard".
3122
3123 10. There was no check on the number of named subpatterns nor the maximum
3124 length of a subpattern name. The product of these values is used to compute
3125 the size of the memory block for a compiled pattern. By supplying a very
3126 long subpattern name and a large number of named subpatterns, the size
3127 computation could be caused to overflow. This is now prevented by limiting
3128 the length of names to 32 characters, and the number of named subpatterns
3129 to 10,000.
3130
3131 11. Subpatterns that are repeated with specific counts have to be replicated in
3132 the compiled pattern. The size of memory for this was computed from the
3133 length of the subpattern and the repeat count. The latter is limited to
3134 65535, but there was no limit on the former, meaning that integer overflow
3135 could in principle occur. The compiled length of a repeated subpattern is
3136 now limited to 30,000 bytes in order to prevent this.
3137
3138 12. Added the optional facility to have named substrings with the same name.
3139
3140 13. Added the ability to use a named substring as a condition, using the
3141 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3142 are numbers (not recommended). Forward references are permitted.
3143
3144 14. Added forward references in named backreferences (if you see what I mean).
3145
3146 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3147 pattern could run off the end of the subject. For example, the pattern
3148 "(?s)(.{1,5})"8 did this with the subject "ab".
3149
3150 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3151 PCRE_CASELESS was set when matching characters that were quantified with ?
3152 or *.
3153
3154 17. A character class other than a single negated character that had a minimum
3155 but no maximum quantifier - for example [ab]{6,} - was not handled
3156 correctly by pce_dfa_exec(). It would match only one character.
3157
3158 18. A valid (though odd) pattern that looked like a POSIX character
3159 class but used an invalid character after [ (for example [[,abc,]]) caused
3160 pcre_compile() to give the error "Failed: internal error: code overflow" or
3161 in some cases to crash with a glibc free() error. This could even happen if
3162 the pattern terminated after [[ but there just happened to be a sequence of
3163 letters, a binary zero, and a closing ] in the memory that followed.
3164
3165 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3166 over the years. Originally (before any Unicode support), just the bottom 8
3167 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3168 output from "man perlunicode" includes this:
3169
3170 The regular expression compiler produces polymorphic opcodes. That
3171 is, the pattern adapts to the data and automatically switches to
3172 the Unicode character scheme when presented with Unicode data--or
3173 instead uses a traditional byte scheme when presented with byte
3174 data.
3175
3176 Sadly, a wide octal escape does not cause a switch, and in a string with
3177 no other multibyte characters, these octal escapes are treated as before.
3178 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3179 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3180 Unicode string.
3181
3182 I have not perpetrated such confusion in PCRE. Up till now, it took just
3183 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3184 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3185 translate to the appropriate multibyte character.
3186
3187 29. Applied some refactoring to reduce the number of warnings from Microsoft
3188 and Borland compilers. This has included removing the fudge introduced
3189 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3190 a warning about an unused variable.
3191
3192 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3193 characters since release 4.0, because Perl (from release 5.004) does not.
3194 [Or at least, is documented not to: some releases seem to be in conflict
3195 with the documentation.] However, when a pattern was studied with
3196 pcre_study() and all its branches started with \s, PCRE still included VT
3197 as a possible starting character. Of course, this did no harm; it just
3198 caused an unnecessary match attempt.
3199
3200 22. Removed a now-redundant internal flag bit that recorded the fact that case
3201 dependency changed within the pattern. This was once needed for "required
3202 byte" processing, but is no longer used. This recovers a now-scarce options
3203 bit. Also moved the least significant internal flag bit to the most-
3204 significant bit of the word, which was not previously used (hangover from
3205 the days when it was an int rather than a uint) to free up another bit for
3206 the future.
3207
3208 23. Added support for CRLF line endings as well as CR and LF. As well as the
3209 default being selectable at build time, it can now be changed at runtime
3210 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3211 specify that it is scanning data with non-default line endings.
3212
3213 24. Changed the definition of CXXLINK to make it agree with the definition of
3214 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3215
3216 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3217 recursions. This makes a big different to stack usage for some patterns.
3218
3219 26. If a subpattern containing a named recursion or subroutine reference such
3220 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3221 the space required for the compiled pattern went wrong and gave too small a
3222 value. Depending on the environment, this could lead to "Failed: internal
3223 error: code overflow at offset 49" or "glibc detected double free or
3224 corruption" errors.
3225
3226 27. Applied patches from Google (a) to support the new newline modes and (b) to
3227 advance over multibyte UTF-8 characters in GlobalReplace.
3228
3229 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3230 difference for some implementation of PCRE in some Windows version.
3231
3232 29. Added some extra testing facilities to pcretest:
3233
3234 \q<number> in a data line sets the "match limit" value
3235 \Q<number> in a data line sets the "match recursion limt" value
3236 -S <number> sets the stack size, where <number> is in megabytes
3237
3238 The -S option isn't available for Windows.
3239
3240
3241 Version 6.6 06-Feb-06
3242 ---------------------
3243
3244 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3245 in pcreposix.h. I have copied the definition from pcre.h.
3246
3247 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3248 because pcre.h is no longer a built file.
3249
3250 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3251 not normally included in the compiled code.
3252
3253
3254 Version 6.5 01-Feb-06
3255 ---------------------
3256
3257 1. When using the partial match feature with pcre_dfa_exec(), it was not
3258 anchoring the second and subsequent partial matches at the new starting
3259 point. This could lead to incorrect results. For example, with the pattern
3260 /1234/, partially matching against "123" and then "a4" gave a match.
3261
3262 2. Changes to pcregrep:
3263
3264 (a) All non-match returns from pcre_exec() were being treated as failures
3265 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3266 error message is output. Some extra information is given for the
3267 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3268 probably the only errors that are likely to be caused by users (by
3269 specifying a regex that has nested indefinite repeats, for instance).
3270 If there are more than 20 of these errors, pcregrep is abandoned.
3271
3272 (b) A binary zero was treated as data while matching, but terminated the
3273 output line if it was written out. This has been fixed: binary zeroes
3274 are now no different to any other data bytes.
3275
3276 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3277 used to set a locale for matching. The --locale=xxxx long option has
3278 been added (no short equivalent) to specify a locale explicitly on the
3279 pcregrep command, overriding the environment variables.
3280
3281 (d) When -B was used with -n, some line numbers in the output were one less
3282 than they should have been.
3283
3284 (e) Added the -o (--only-matching) option.
3285
3286 (f) If -A or -C was used with -c (count only), some lines of context were
3287 accidentally printed for the final match.
3288
3289 (g) Added the -H (--with-filename) option.
3290
3291 (h) The combination of options -rh failed to suppress file names for files
3292 that were found from directory arguments.
3293
3294 (i) Added the -D (--devices) and -d (--directories) options.
3295
3296 (j) Added the -F (--fixed-strings) option.
3297
3298 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3299
3300 (l) Added the --colo(u)r option.
3301
3302 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3303 is not present by default.
3304
3305 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3306 items such as (?R) or (?1), when the recursion could match a number of
3307 alternatives. If it matched one of the alternatives, but subsequently,
3308 outside the recursion, there was a failure, the code tried to back up into
3309 the recursion. However, because of the way PCRE is implemented, this is not
3310 possible, and the result was an incorrect result from the match.
3311
3312 In order to prevent this happening, the specification of recursion has
3313 been changed so that all such subpatterns are automatically treated as
3314 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3315
3316 4. I had overlooked the fact that, in some locales, there are characters for
3317 which isalpha() is true but neither isupper() nor islower() are true. In
3318 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3319 and ordfeminine) are like this. This affected the treatment of \w and \W
3320 when they appeared in character classes, but not when they appeared outside
3321 a character class. The bit map for "word" characters is now created
3322 separately from the results of isalnum() instead of just taking it from the
3323 upper, lower, and digit maps. (Plus the underscore character, of course.)
3324
3325 5. The above bug also affected the handling of POSIX character classes such as
3326 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3327 permanent tables. Instead, the bit maps for such a class were previously
3328 created as the appropriate unions of the upper, lower, and digit bitmaps.
3329 Now they are created by subtraction from the [[:word:]] class, which has
3330 its own bitmap.
3331
3332 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3333 It is created by subtracting the vertical space characters (\x09, \x0a,
3334 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3335 subtraction was done in the overall bitmap for a character class, meaning
3336 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3337 be recognized. This bug has been fixed.
3338
3339 7. Patches from the folks at Google:
3340
3341 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3342 real life, but is still worth protecting against".
3343
3344 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3345 regular expressions".
3346
3347 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3348 have it.
3349
3350 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3351 "configure" and the latter not, in order to fix a problem somebody had
3352 with compiling the Arg class on HP-UX.
3353
3354 (e) Improve the error-handling of the C++ wrapper a little bit.
3355
3356 (f) New tests for checking recursion limiting.
3357
3358 8. The pcre_memmove() function, which is used only if the environment does not
3359 have a standard memmove() function (and is therefore rarely compiled),
3360 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3361 returning a result (though PCRE never actually uses the result).
3362
3363 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3364 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3365 returned instead of calling malloc() with an overflowing number that would
3366 most likely cause subsequent chaos.
3367
3368 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3369
3370 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3371 with this option is matched, the nmatch and pmatch options of regexec() are
3372 ignored.
3373
3374 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3375 provided in case anyone wants to the the POSIX interface with UTF-8
3376 strings.
3377
3378 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3379 C++ linking (needed for some HP-UX environments).
3380
3381 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3382 (unused parameter) and in the pcre_printint() function (omitted "default"
3383 switch label when the default is to do nothing).
3384
3385 15. Added some code to make it possible, when PCRE is compiled as a C++
3386 library, to replace subject pointers for pcre_exec() with a smart pointer
3387 class, thus making it possible to process discontinuous strings.
3388
3389 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3390 much the same function. They were added by different people who were trying
3391 to make PCRE easy to compile on non-Unix systems. It has been suggested
3392 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3393 for compiling on Windows systems. I have therefore replaced it with
3394 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3395 defaults to "extern" for C or "extern C" for C++, which works fine on
3396 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3397 SCOPE with something explicit in config.h. In addition:
3398
3399 (a) pcreposix.h still had just "extern" instead of either of these macros;
3400 I have replaced it with PCRE_DATA_SCOPE.
3401
3402 (b) Functions such as _pcre_xclass(), which are internal to the library,
3403 but external in the C sense, all had PCRE_EXPORT in their definitions.
3404 This is apparently wrong for the Windows case, so I have removed it.
3405 (It makes no difference on Unix-like systems.)
3406
3407 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3408 of recursive calls to match(). This is different to MATCH_LIMIT because
3409 that limits the total number of calls to match(), not all of which increase
3410 the depth of recursion. Limiting the recursion depth limits the amount of
3411 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3412 when PCRE is compiled, and changed at run time. A patch from Google adds
3413 this functionality to the C++ interface.
3414
3415 18. Changes to the handling of Unicode character properties:
3416
3417 (a) Updated the table to Unicode 4.1.0.
3418
3419 (b) Recognize characters that are not in the table as "Cn" (undefined).
3420
3421 (c) I revised the way the table is implemented to a much improved format
3422 which includes recognition of ranges. It now supports the ranges that
3423 are defined in UnicodeData.txt, and it also amalgamates other
3424 characters into ranges. This has reduced the number of entries in the
3425 table from around 16,000 to around 3,000, thus reducing its size
3426 considerably. I realized I did not need to use a tree structure after
3427 all - a binary chop search is just as efficient. Having reduced the
3428 number of entries, I extended their size from 6 bytes to 8 bytes to
3429 allow for more data.
3430
3431 (d) Added support for Unicode script names via properties such as \p{Han}.
3432
3433 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3434 matching that character.
3435
3436 20. When matching a repeated Unicode property with a minimum greater than zero,
3437 (for example \pL{2,}), PCRE could look past the end of the subject if it
3438 reached it while seeking the minimum number of characters. This could
3439 happen only if some of the characters were more than one byte long, because
3440 there is a check for at least the minimum number of bytes.
3441
3442 21. Refactored the implementation of \p and \P so as to be more general, to
3443 allow for more different types of property in future. This has changed the
3444 compiled form incompatibly. Anybody with saved compiled patterns that use
3445 \p or \P will have to recompile them.
3446
3447 22. Added "Any" and "L&" to the supported property types.
3448
3449 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3450 but give a compile time error if the value is greater than 0xff.
3451
3452 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3453 accidentally not being installed or uninstalled.
3454
3455 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3456 made were to insert the current release number. This seemed silly, because
3457 it made things harder for people building PCRE on systems that don't run
3458 "configure". I have turned pcre.h into a distributed file, no longer built
3459 by "configure", with the version identification directly included. There is
3460 no longer a pcre.h.in file.
3461
3462 However, this change necessitated a change to the pcre-config script as
3463 well. It is built from pcre-config.in, and one of the substitutions was the
3464 release number. I have updated configure.ac so that ./configure now finds
3465 the release number by grepping pcre.h.
3466
3467 26. Added the ability to run the tests under valgrind.
3468
3469
3470 Version 6.4 05-Sep-05
3471 ---------------------
3472
3473 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3474 "--" to be printed when multiple files were scanned, even when none of the
3475 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3476 consider it to be a bug, and have restored the previous behaviour.
3477
3478 2. A couple of code tidies to get rid of compiler warnings.
3479
3480 3. The pcretest program used to cheat by referring to symbols in the library
3481 whose names begin with _pcre_. These are internal symbols that are not
3482 really supposed to be visible externally, and in some environments it is
3483 possible to suppress them. The cheating is now confined to including
3484 certain files from the library's source, which is a bit cleaner.
3485
3486 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3487 file's purpose clearer.
3488
3489 5. Reorganized pcre_ucp_findchar().
3490
3491
3492 Version 6.3 15-Aug-05
3493 ---------------------
3494
3495 1. The file libpcre.pc.in did not have general read permission in the tarball.
3496
3497 2. There were some problems when building without C++ support:
3498
3499 (a) If C++ support was not built, "make install" and "make test" still
3500 tried to test it.
3501
3502 (b) There were problems when the value of CXX was explicitly set. Some
3503 changes have been made to try to fix these, and ...
3504
3505 (c) --disable-cpp can now be used to explicitly disable C++ support.
3506
3507 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3508 backslash in a target when C++ was disabled. This confuses some
3509 versions of "make", apparently. Using an intermediate variable solves
3510 this. (Same for CPP_LOBJ.)
3511
3512 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3513 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3514 necessary on certain architectures.
3515
3516 4. Added a setting of -export-symbols-regex to the link command to remove
3517 those symbols that are exported in the C sense, but actually are local
3518 within the library, and not documented. Their names all begin with
3519 "_pcre_". This is not a perfect job, because (a) we have to except some
3520 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3521 available (and never for static libraries). I have made a note to try to
3522 find a way round (a) in the future.
3523
3524
3525 Version 6.2 01-Aug-05
3526 ---------------------
3527
3528 1. There was no test for integer overflow of quantifier values. A construction
3529 such as {1111111111111111} would give undefined results. What is worse, if
3530 a minimum quantifier for a parenthesized subpattern overflowed and became
3531 negative, the calculation of the memory size went wrong. This could have
3532 led to memory overwriting.
3533
3534 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3535
3536 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3537 operating environments where this matters.
3538
3539 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3540 PCRE options from within the C++ wrapper.
3541
3542 5. Named capturing subpatterns were not being correctly counted when a pattern
3543 was compiled. This caused two problems: (a) If there were more than 100
3544 such subpatterns, the calculation of the memory needed for the whole
3545 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3546 back references of the form \12, where the number was greater than 9, were
3547 not recognized as back references, even though there were sufficient
3548 previous subpatterns.
3549
3550 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3551 versions of gcc, e.g. 2.95.4.
3552
3553
3554 Version 6.1 21-Jun-05
3555 ---------------------
3556
3557 1. There was one reference to the variable "posix" in pcretest.c that was not
3558 surrounded by "#if !defined NOPOSIX".
3559
3560 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3561 the cross-check on the old pcre_info() function, for the benefit of the
3562 cut-down version of PCRE that is currently imported into Exim.
3563
3564 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3565 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3566 patterns that start (?i) but I don't think that matters. The use of (?i) is
3567 just an example; this all applies to the other options as well.
3568
3569 4. Since libtool seems to echo the compile commands it is issuing, the output
3570 from "make" can be reduced a bit by putting "@" in front of each libtool
3571 compile command.
3572
3573 5. Patch from the folks at Google for configure.in to be a bit more thorough
3574 in checking for a suitable C++ installation before trying to compile the
3575 C++ stuff. This should fix a reported problem when a compiler was present,
3576 but no suitable headers.
3577
3578 6. The man pages all had just "PCRE" as their title. I have changed them to
3579 be the relevant file name. I have also arranged that these names are
3580 retained in the file doc/pcre.txt, which is a concatenation in text format
3581 of all the man pages except the little individual ones for each function.
3582
3583 7. The NON-UNIX-USE file had not been updated for the different set of source
3584 files that come with release 6. I also added a few comments about the C++
3585 wrapper.
3586
3587
3588 Version 6.0 07-Jun-05
3589 ---------------------
3590
3591 1. Some minor internal re-organization to help with my DFA experiments.
3592
3593 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3594 didn't matter for the library itself when fully configured, but did matter
3595 when compiling without UCP support, or within Exim, where the ucp files are
3596 not imported.
3597
3598 3. Refactoring of the library code to split up the various functions into
3599 different source modules. The addition of the new DFA matching code (see
3600 below) to a single monolithic source would have made it really too
3601 unwieldy, quite apart from causing all the code to be include in a
3602 statically linked application, when only some functions are used. This is
3603 relevant even without the DFA addition now that patterns can be compiled in
3604 one application and matched in another.
3605
3606 The downside of splitting up is that there have to be some external
3607 functions and data tables that are used internally in different modules of
3608 the library but which are not part of the API. These have all had their
3609 names changed to start with "_pcre_" so that they are unlikely to clash
3610 with other external names.
3611
3612 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3613 a different (DFA) algorithm. Although it is slower than the original
3614 function, it does have some advantages for certain types of matching
3615 problem.
3616
3617 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3618 including restarting after a partial match.
3619
3620 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3621 defined when compiling for Windows was sent to me. I have put it into the
3622 code, though I have no means of testing or verifying it.
3623
3624 7. Added the pcre_refcount() auxiliary function.
3625
3626 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3627 match before or at the first newline in the subject string. In pcretest,
3628 the /f option on a pattern can be used to set this.
3629
3630 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3631 would behave wrongly. This has been present in PCRE since release 4.0.
3632
3633 10. A number of changes to the pcregrep command:
3634
3635 (a) Refactored how -x works; insert ^(...)$ instead of setting
3636 PCRE_ANCHORED and checking the length, in preparation for adding
3637 something similar for -w.
3638
3639 (b) Added the -w (match as a word) option.
3640
3641 (c) Refactored the way lines are read and buffered so as to have more
3642 than one at a time available.
3643
3644 (d) Implemented a pcregrep test script.
3645
3646 (e) Added the -M (multiline match) option. This allows patterns to match
3647 over several lines of the subject. The buffering ensures that at least
3648 8K, or the rest of the document (whichever is the shorter) is available
3649 for matching (and similarly the previous 8K for lookbehind assertions).
3650
3651 (f) Changed the --help output so that it now says
3652
3653 -w, --word-regex(p)
3654
3655 instead of two lines, one with "regex" and the other with "regexp"
3656 because that confused at least one person since the short forms are the
3657 same. (This required a bit of code, as the output is generated
3658 automatically from a table. It wasn't just a text change.)
3659
3660 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3661 option but starts with a hyphen. Could be a pattern or a path name
3662 starting with a hyphen, for instance.
3663
3664 (h) "-" can be given as a file name to represent stdin.
3665
3666 (i) When file names are being printed, "(standard input)" is used for
3667 the standard input, for compatibility with GNU grep. Previously
3668 "<stdin>" was used.
3669
3670 (j) The option --label=xxx can be used to supply a name to be used for
3671 stdin when file names are being printed. There is no short form.
3672
3673 (k) Re-factored the options decoding logic because we are going to add
3674 two more options that take data. Such options can now be given in four
3675 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3676
3677 (l) Added the -A, -B, and -C options for requesting that lines of context
3678 around matches be printed.
3679
3680 (m) Added the -L option to print the names of files that do not contain
3681 any matching lines, that is, the complement of -l.
3682
3683 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3684 continue to scan other files.
3685
3686 (o) The -s option was incorrectly implemented. For compatibility with other
3687 greps, it now suppresses the error message for a non-existent or non-
3688 accessible file (but not the return code). There is a new option called
3689 -q that suppresses the output of matching lines, which was what -s was
3690 previously doing.
3691
3692 (p) Added --include and --exclude options to specify files for inclusion
3693 and exclusion when recursing.
3694
3695 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3696 Hopefully, it now does.
3697
3698 12. Missing cast in pcre_study().
3699
3700 13. Added an "uninstall" target to the makefile.
3701
3702 14. Replaced "extern" in the function prototypes in Makefile.in with
3703 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3704 world, but is set differently for Windows.
3705
3706 15. Added a second compiling function called pcre_compile2(). The only
3707 difference is that it has an extra argument, which is a pointer to an
3708 integer error code. When there is a compile-time failure, this is set
3709 non-zero, in addition to the error test pointer being set to point to an
3710 error message. The new argument may be NULL if no error number is required
3711 (but then you may as well call pcre_compile(), which is now just a
3712 wrapper). This facility is provided because some applications need a
3713 numeric error indication, but it has also enabled me to tidy up the way
3714 compile-time errors are handled in the POSIX wrapper.
3715
3716 16. Added VPATH=.libs to the makefile; this should help when building with one
3717 prefix path and installing with another. (Or so I'm told by someone who
3718 knows more about this stuff than I do.)
3719
3720 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3721 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3722 match everything, including newlines. This is not POSIX-compatible, but
3723 somebody wanted the feature. From pcretest it can be activated by using
3724 both the P and the s flags.
3725
3726 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3727
3728 19. libpcre.pc was being incorrectly installed as executable.
3729
3730 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3731 it now also looks for '\r' so that it will work unmodified on Windows.
3732
3733 21. Added Google's contributed C++ wrapper to the distribution.
3734
3735 22. Added some untidy missing memory free() calls in pcretest, to keep
3736 Electric Fence happy when testing.
3737
3738
3739
3740 Version 5.0 13-Sep-04
3741 ---------------------
3742
3743 1. Internal change: literal characters are no longer packed up into items
3744 containing multiple characters in a single byte-string. Each character
3745 is now matched using a separate opcode. However, there may be more than one
3746 byte in the character in UTF-8 mode.
3747
3748 2. The pcre_callout_block structure has two new fields: pattern_position and
3749 next_item_length. These contain the offset in the pattern to the next match
3750 item, and its length, respectively.
3751
3752 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3753 insertion of callouts before each pattern item. Added the /C option to
3754 pcretest to make use of this.
3755
3756 4. On the advice of a Windows user, the lines
3757
3758 #if defined(_WIN32) || defined(WIN32)
3759 _setmode( _fileno( stdout ), 0x8000 );
3760 #endif /* defined(_WIN32) || defined(WIN32) */
3761
3762 have been added to the source of pcretest. This apparently does useful
3763 magic in relation to line terminators.
3764
3765 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3766 for the benefit of those environments where the "b" makes a difference.
3767
3768 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3769 to know about it. I have put a hack into configure.in that adds in code
3770 to set GCC=yes if CC=icc. This seems to end up at a point in the
3771 generated configure script that is early enough to affect the setting of
3772 compiler options, which is what is needed, but I have no means of testing
3773 whether it really works. (The user who reported this had patched the
3774 generated configure script, which of course I cannot do.)
3775
3776 LATER: After change 22 below (new libtool files), the configure script
3777 seems to know about icc (and also ecc). Therefore, I have commented out
3778 this hack in configure.in.
3779
3780 7. Added support for pkg-config (2 patches were sent in).
3781
3782 8. Negated POSIX character classes that used a combination of internal tables
3783 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3784 [[:^ascii]]. Typically, they would match almost any characters. The other
3785 POSIX classes were not broken in this way.
3786
3787 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3788 to find the match, as PCRE was deluded into thinking that the match had to
3789 start at the start point or following a newline. The same bug applied to
3790 patterns with negative forward assertions or any backward assertions
3791 preceding ".*" at the start, unless the pattern required a fixed first
3792 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3793
3794 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3795 starting at the last subject character, bytes beyond the end of the subject
3796 string were read.
3797
3798 11. Renamed the variable "class" as "classbits" to make life easier for C++
3799 users. (Previously there was a macro definition, but it apparently wasn't
3800 enough.)
3801
3802 12. Added the new field "tables" to the extra data so that tables can be passed
3803 in at exec time, or the internal tables can be re-selected. This allows
3804 a compiled regex to be saved and re-used at a later time by a different
3805 program that might have everything at different addresses.
3806
3807 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3808 -R library as well as a -L library.
3809
3810 14. The debugging options of pcretest (-d on the command line or D on a
3811 pattern) showed incorrect output for anything following an extended class
3812 that contained multibyte characters and which was followed by a quantifier.
3813
3814 15. Added optional support for general category Unicode character properties
3815 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3816 support. It adds about 90K to the size of the library. The meanings of the
3817 inbuilt class escapes such as \d and \s have NOT been changed.
3818
3819 16. Updated pcredemo.c to include calls to free() to release the memory for the
3820 compiled pattern.
3821
3822 17. The generated file chartables.c was being created in the source directory
3823 instead of in the building directory. This caused the build to fail if the
3824 source directory was different from the building directory, and was
3825 read-only.
3826
3827 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3828 file. No doubt somebody will tell me if they don't make sense... Also added
3829 Dan Mooney's comments about building on OpenVMS.
3830
3831 19. Added support for partial matching via the PCRE_PARTIAL option for
3832 pcre_exec() and the \P data escape in pcretest.
3833
3834 20. Extended pcretest with 3 new pattern features:
3835
3836 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3837 write the compiled pattern to the file whose name is "rest-of-line".
3838 This is a straight binary dump of the data, with the saved pointer to
3839 the character tables forced to be NULL. The study data, if any, is
3840 written too. After writing, pcretest reads a new pattern.
3841
3842 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3843 compiled pattern from the given file. There must not be any
3844 occurrences of "<" in the file name (pretty unlikely); if there are,
3845 pcretest will instead treat the initial "<" as a pattern delimiter.
3846 After reading in the pattern, pcretest goes on to read data lines as
3847 usual.
3848
3849 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3850 and 16-bit fields in a compiled pattern, to simulate a pattern that
3851 was compiled on a host of opposite endianness.
3852
3853 21. The pcre-exec() function can now cope with patterns that were compiled on
3854 hosts of opposite endianness, with this restriction:
3855
3856 As for any compiled expression that is saved and used later, the tables
3857 pointer field cannot be preserved; the extra_data field in the arguments
3858 to pcre_exec() should be used to pass in a tables address if a value
3859 other than the default internal tables were used at compile time.
3860
3861 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3862 now diagnosed as an error. Previously, most of the time, a negative number
3863 would have been treated as zero, but if in addition "ovector" was passed as
3864 NULL, a crash could occur.
3865
3866 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3867 new versions from the libtool 1.5 distribution (the last one is a copy of
3868 a file called libtool.m4). This seems to have fixed the need to patch
3869 "configure" to support Darwin 1.3 (which I used to do). However, I still
3870 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3871 workstation).
3872
3873 24. Changed the PCRE licence to be the more standard "BSD" licence.
3874
3875
3876 Version 4.5 01-Dec-03
3877 ---------------------
3878
3879 1. There has been some re-arrangement of the code for the match() function so
3880 that it can be compiled in a version that does not call itself recursively.
3881 Instead, it keeps those local variables that need separate instances for
3882 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3883 needs to "recurse". Keeping track of where control must go is done by means
3884 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3885 hide most of the details from the main code, and operates only if
3886 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3887 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3888 operating.
3889
3890 To make it easier for callers to provide specially tailored get/free
3891 functions for this usage, two new functions, pcre_stack_malloc, and
3892 pcre_stack_free, are used. They are always called in strict stacking order,
3893 and the size of block requested is always the same.
3894
3895 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3896 PCRE has been compiled to use the stack or the heap for recursion. The
3897 -C option of pcretest uses this to show which version is compiled.
3898
3899 A new data escape \S, is added to pcretest; it causes the amounts of store
3900 obtained and freed by both kinds of malloc/free at match time to be added
3901 to the output.
3902
3903 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3904 what's available on my current Linux desktop machine.
3905
3906 3. When matching a UTF-8 string, the test for a valid string at the start has
3907 been extended. If start_offset is not zero, PCRE now checks that it points
3908 to a byte that is the start of a UTF-8 character. If not, it returns
3909 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3910 this is necessary because there may be backward assertions in the pattern.
3911 When matching the same subject several times, it may save resources to use
3912 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3913
3914 4. The code for checking the validity of UTF-8 strings has been tightened so
3915 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3916 containing "overlong sequences".
3917
3918 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3919 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3920 should have been "&", but it just so happened that all the cases this let
3921 through by mistake were picked up later in the function.
3922
3923 6. I had used a variable called "isblank" - this is a C99 function, causing
3924 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3925
3926 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3927 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3928 so the version is shown; (c) stop "make test" echoing "./RunTest".
3929
3930 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3931
3932 9. The prototype for memmove() for systems that don't have it was using
3933 size_t, but the inclusion of the header that defines size_t was later. I've
3934 moved the #includes for the C headers earlier to avoid this.
3935
3936 10. Added some adjustments to the code to make it easier to compiler on certain
3937 special systems:
3938
3939 (a) Some "const" qualifiers were missing.
3940 (b) Added the macro EXPORT before all exported functions; by default this
3941 is defined to be empty.
3942 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3943 that it reads its output file name as an argument instead of writing
3944 to the standard output and assuming this can be redirected.
3945
3946 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3947 class containing characters with values greater than 255, PCRE compilation
3948 went into a loop.
3949
3950 12. A recursive reference to a subpattern that was within another subpattern
3951 that had a minimum quantifier of zero caused PCRE to crash. For example,
3952 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3953 recursion. If the recursively-called subpattern itself had a zero repeat,
3954 that was OK.
3955
3956 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3957 buffer into which it was copied (for escape processing) was still set at
3958 1024, so long lines caused crashes.
3959
3960 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3961 "internal error: code overflow...". This applied to any character class
3962 that was followed by a possessive quantifier.
3963
3964 15. Modified the Makefile to add libpcre.la as a prerequisite for
3965 libpcreposix.la because I was told this is needed for a parallel build to
3966 work.
3967
3968 16. If a pattern that contained .* following optional items at the start was
3969 studied, the wrong optimizing data was generated, leading to matching
3970 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3971 matching string must start with a or b or c. The correct conclusion for
3972 this pattern is that a match can start with any character.
3973
3974
3975 Version 4.4 13-Aug-03
3976 ---------------------
3977
3978 1. In UTF-8 mode, a character class containing characters with values between
3979 127 and 255 was not handled correctly if the compiled pattern was studied.
3980 In fixing this, I have also improved the studying algorithm for such
3981 classes (slightly).
3982
3983 2. Three internal functions had redundant arguments passed to them. Removal
3984 might give a very teeny performance improvement.
3985
3986 3. Documentation bug: the value of the capture_top field in a callout is *one
3987 more than* the number of the hightest numbered captured substring.
3988
3989 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3990 in incorrectly linking with a previously installed version. They now link
3991 explicitly with libpcre.la.
3992
3993 5. configure.in no longer needs to recognize Cygwin specially.
3994
3995 6. A problem in pcre.in for Windows platforms is fixed.
3996
3997 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3998 pcretest, it used to include the size of the study block as part of its
3999 output. Unfortunately, the structure contains a field that has a different
4000 size on different hardware architectures. This meant that the tests that
4001 showed this size failed. As the block is currently always of a fixed size,
4002 this information isn't actually particularly useful in pcretest output, so
4003 I have just removed it.
4004
4005 8. Three pre-processor statements accidentally did not start in column 1.
4006 Sadly, there are *still* compilers around that complain, even though
4007 standard C has not required this for well over a decade. Sigh.
4008
4009 9. In pcretest, the code for checking callouts passed small integers in the
4010 callout_data field, which is a void * field. However, some picky compilers
4011 complained about the casts involved for this on 64-bit systems. Now
4012 pcretest passes the address of the small integer instead, which should get
4013 rid of the warnings.
4014
4015 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4016 both compile and run time, and gives an error if an invalid UTF-8 sequence
4017 is found. There is a option for disabling this check in cases where the
4018 string is known to be correct and/or the maximum performance is wanted.
4019
4020 11. In response to a bug report, I changed one line in Makefile.in from
4021
4022 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4023 to
4024 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4025
4026 to look similar to other lines, but I have no way of telling whether this
4027 is the right thing to do, as I do not use Windows. No doubt I'll get told
4028 if it's wrong...
4029
4030
4031 Version 4.3 21-May-03
4032 ---------------------
4033
4034 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4035 Makefile.
4036
4037 2. Some refactoring to improve the quality of the code:
4038
4039 (i) The utf8_table... variables are now declared "const".
4040
4041 (ii) The code for \cx, which used the "case flipping" table to upper case
4042 lower case letters, now just substracts 32. This is ASCII-specific,
4043 but the whole concept of \cx is ASCII-specific, so it seems
4044 reasonable.
4045
4046 (iii) PCRE was using its character types table to recognize decimal and
4047 hexadecimal digits in the pattern. This is silly, because it handles
4048 only 0-9, a-f, and A-F, but the character types table is locale-
4049 specific, which means strange things might happen. A private
4050 table is now used for this - though it costs 256 bytes, a table is
4051 much faster than multiple explicit tests. Of course, the standard
4052 character types table is still used for matching digits in subject
4053 strings against \d.
4054
4055 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4056 ending in _t are). So I've renamed it as ESC_tee.
4057
4058 3. The first argument for regexec() in the POSIX wrapper should have been
4059 defined as "const".
4060
4061 4. Changed pcretest to use malloc() for its buffers so that they can be
4062 Electric Fenced for debugging.
4063
4064 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4065 to read one or more bytes before the start of the subject string. Often this
4066 had no effect on PCRE's behaviour, but in some circumstances it could
4067 provoke a segmentation fault.
4068
4069 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4070 to try to read one or more bytes before the start of the subject string.
4071
4072 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4073 UTF-8 support could misbehave in various ways if the subject string
4074 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4075 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4076 back over UTF-8 characters.)
4077
4078
4079 Version 4.2 14-Apr-03
4080 ---------------------
4081
4082 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4083
4084 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4085 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4086 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4087 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4088 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4089 and BUILD_EXEEXT
4090 Note: automatic setting of the BUILD variables is not yet working
4091 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4092 compile-time but not at link-time
4093 [LINK]: use for linking executables only
4094 make different versions for Windows and non-Windows
4095 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4096 libraries
4097 [LINK_FOR_BUILD]: new variable
4098 [OBJEXT]: use throughout
4099 [EXEEXT]: use throughout
4100 <winshared>: new target
4101 <wininstall>: new target
4102 <dftables.o>: use native compiler
4103 <dftables>: use native linker
4104 <install>: handle Windows platform correctly
4105 <clean>: ditto
4106 <check>: ditto
4107 copy DLL to top builddir before testing
4108
4109 As part of these changes, -no-undefined was removed again. This was reported
4110 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4111 in any case.
4112
4113 3. Some tidies to get rid of compiler warnings:
4114
4115 . In the match_data structure, match_limit was an unsigned long int, whereas
4116 match_call_count was an int. I've made them both unsigned long ints.
4117
4118 . In pcretest the fact that a const uschar * doesn't automatically cast to
4119 a void * provoked a warning.
4120
4121 . Turning on some more compiler warnings threw up some "shadow" variables
4122 and a few more missing casts.
4123
4124 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4125 option, a class that contained a single character with a value between 128
4126 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4127
4128 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4129 option, a class that contained several characters, but with at least one
4130 whose value was between 128 and 255 caused PCRE to crash.
4131
4132
4133 Version 4.1 12-Mar-03
4134 ---------------------
4135
4136 1. Compiling with gcc -pedantic found a couple of places where casts were
4137 needed, and a string in dftables.c that was longer than standard compilers are
4138 required to support.
4139
4140 2. Compiling with Sun's compiler found a few more places where the code could
4141 be tidied up in order to avoid warnings.
4142
4143 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4144 first of these names is deprecated in the latest Autoconf in favour of the name
4145 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4146 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4147 analogy I have changed it to CFLAGS_FOR_BUILD.
4148
4149 4. Added -no-undefined to the linking command in the Makefile, because this is
4150 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4151 linking step for the pcreposix library.
4152
4153 5. PCRE was failing to diagnose the case of two named groups with the same
4154 name.
4155
4156 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4157 literal character that is needed in the subject for a match, and scans along to
4158 ensure that it is present before embarking on the full matching process. This
4159 saves time in cases of nested unlimited repeats that are never going to match.
4160 Problem: the scan can take a lot of time if the subject is very long (e.g.
4161 megabytes), thus penalizing straightforward matches. It is now done only if the
4162 amount of subject to be scanned is less than 1000 bytes.
4163
4164 7. A lesser problem with the same optimization is that it was recording the
4165 first character of an anchored pattern as "needed", thus provoking a search
4166 right along the subject, even when the first match of the pattern was going to
4167 fail. The "needed" character is now not set for anchored patterns, unless it
4168 follows something in the pattern that is of non-fixed length. Thus, it still
4169 fulfils its original purpose of finding quick non-matches in cases of nested
4170 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4171
4172
4173 Version 4.0 17-Feb-03
4174 ---------------------
4175
4176 1. If a comment in an extended regex that started immediately after a meta-item
4177 extended to the end of string, PCRE compiled incorrect data. This could lead to
4178 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4179
4180 2. Moved to autoconf 2.53 and libtool 1.4.2.
4181
4182 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4183 the special perltest8 script is no longer needed - all the tests can be run
4184 from a single perltest script.
4185
4186 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4187 by \s. It has now been removed in PCRE. This means it isn't recognized as
4188 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4189 class [:space:] *does* include VT, thereby creating a mess.
4190
4191 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4192 space and tab.
4193
4194 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4195 its new features into the main test script, reducing the number of scripts.
4196
4197 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4198 were backward compatible, and made the (?i) apply to the whole pattern, as if
4199 /i were given. Now it behaves more logically, and applies the option setting
4200 only to what follows. PCRE has been changed to follow suit. However, if it
4201 finds options settings right at the start of the pattern, it extracts them into
4202 the global options, as before. Thus, they show up in the info data.
4203
4204 8. Added support for the \Q...\E escape sequence. Characters in between are
4205 treated as literals. This is slightly different from Perl in that $ and @ are
4206 also handled as literals inside the quotes. In Perl, they will cause variable
4207 interpolation. Note the following examples:
4208
4209 Pattern PCRE matches Perl matches
4210
4211 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4212 \Qabc\$xyz\E abc\$xyz abc\$xyz
4213 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4214
4215 For compatibility with Perl, \Q...\E sequences are recognized inside character
4216 classes as well as outside them.
4217
4218 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4219 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4220 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4221 signed/unsigned warnings.
4222
4223 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4224 option for pcretest, so I've replaced it by a simple function that does just
4225 that job.
4226
4227 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4228 "pcregrep -".
4229
4230 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4231 Java package. This provides some syntactic sugar for simple cases of what my
4232 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4233 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4234 item, you can use this simplified notation. Note that only makes sense with
4235 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4236 greediness, whatever the setting of the PCRE_UNGREEDY option.
4237
4238 13. A change of greediness default within a pattern was not taking effect at
4239 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4240 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4241 was abstracted outside.
4242
4243 14. PCRE now supports the \G assertion. It is true when the current matching
4244 position is at the start point of the match. This differs from \A when the
4245 starting offset is non-zero. Used with the /g option of pcretest (or similar
4246 code), it works in the same way as it does for Perl's /g option. If all
4247 alternatives of a regex begin with \G, the expression is anchored to the start
4248 match position, and the "anchored" flag is set in the compiled expression.
4249
4250 15. Some bugs concerning the handling of certain option changes within patterns
4251 have been fixed. These applied to options other than (?ims). For example,
4252 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4253 been the other way round. Some of this was related to change 7 above.
4254
4255 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4256 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4257 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4258 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4259
4260 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4261 mode. Unlike ".", it always matches newline, whatever the setting of
4262 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4263 assertions. Perl allows it, but it doesn't (in general) work because it can't
4264 calculate the length of the lookbehind. At least, that's the case for Perl
4265 5.8.0 - I've been told they are going to document that it doesn't work in
4266 future.
4267
4268 18. Added an error diagnosis for escapes that PCRE does not support: these are
4269 \L, \l, \N, \P, \p, \U, \u, and \X.
4270
4271 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4272 reading past the end of the pattern in cases such as /[abcd/.
4273
4274 20. PCRE was getting more memory than necessary for patterns with classes that
4275 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4276
4277 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4278 compiling PCRE for use with Virtual Pascal.
4279
4280 22. Small fix to the Makefile to make it work properly if the build is done
4281 outside the source tree.
4282
4283 23. Added a new extension: a condition to go with recursion. If a conditional
4284 subpattern starts with (?(R) the "true" branch is used if recursion has
4285 happened, whereas the "false" branch is used only at the top level.
4286
4287 24. When there was a very long string of literal characters (over 255 bytes
4288 without UTF support, over 250 bytes with UTF support), the computation of how
4289 much memory was required could be incorrect, leading to segfaults or other
4290 strange effects.
4291
4292 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4293 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4294 there was a subsequent back reference to those brackets. This meant that, for
4295 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4296 possible to check for precisely this case. All we can do is abandon the
4297 optimization if .* occurs inside capturing brackets when there are any back
4298 references whatsoever. (See below for a better fix that came later.)
4299
4300 26. The handling of the optimization for finding the first character of a
4301 non-anchored pattern, and for finding a character that is required later in the
4302 match were failing in some cases. This didn't break the matching; it just
4303 failed to optimize when it could. The way this is done has been re-implemented.
4304
4305 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4306
4307 28. Added a new feature that provides some of the functionality that Perl
4308 provides with (?{...}). The facility is termed a "callout". The way it is done
4309 in PCRE is for the caller to provide an optional function, by setting
4310 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4311 global variable. By default it is unset, which disables all calling out. To get
4312 the function called, the regex must include (?C) at appropriate points. This
4313 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4314 This provides a means of identifying different callout points. When PCRE
4315 reaches such a point in the regex, if pcre_callout has been set, the external
4316 function is called. It is provided with data in a structure called
4317 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4318 matching continues; if it returns a non-zero value, the match at the current
4319 point fails. However, backtracking will occur if possible. [This was changed
4320 later and other features added - see item 49 below.]
4321
4322 29. pcretest is upgraded to test the callout functionality. It provides a
4323 callout function that displays information. By default, it shows the start of
4324 the match and the current position in the text. There are some new data escapes
4325 to vary what happens:
4326
4327 \C+ in addition, show current contents of captured substrings
4328 \C- do not supply a callout function
4329 \C!n return 1 when callout number n is reached
4330 \C!n!m return 1 when callout number n is reached for the mth time
4331
4332 30. If pcregrep was called with the -l option and just a single file name, it
4333 output "<stdin>" if a match was found, instead of the file name.
4334
4335 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4336 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4337 pcre_exec(). This saves a malloc/free per call. The default value of
4338 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4339 when configuring.
4340
4341 32. The default maximum size of a compiled pattern is 64K. There have been a
4342 few cases of people hitting this limit. The code now uses macros to handle the
4343 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4344 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4345 configuring. Tests 2 and 5 work only with 2-byte links because they output
4346 debugging information about compiled patterns.
4347
4348 33. Internal code re-arrangements:
4349
4350 (a) Moved the debugging function for printing out a compiled regex into
4351 its own source file (printint.c) and used #include to pull it into
4352 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4353 separate copies.
4354
4355 (b) Defined the list of op-code names for debugging as a macro in
4356 internal.h so that it is next to the definition of the opcodes.
4357
4358 (c) Defined a table of op-code lengths for simpler skipping along compiled
4359 code. This is again a macro in internal.h so that it is next to the
4360 definition of the opcodes.
4361
4362 34. Added support for recursive calls to individual subpatterns, along the
4363 lines of Robin Houston's patch (but implemented somewhat differently).
4364
4365 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4366 allow it to read and process whole directories in Win32. This code was
4367 contributed by Lionel Fourquaux; it has not been tested by me.
4368
4369 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4370 used to name a group. Names consist of alphanumerics and underscores, and must
4371 be unique. Back references use the syntax (?P=name) and recursive calls use
4372 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4373 numbers. The function pcre_fullinfo() can be used after compilation to extract
4374 a name/number map. There are three relevant calls:
4375
4376 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4377 PCRE_INFO_NAMECOUNT yields the number of entries
4378 PCRE_INFO_NAMETABLE yields a pointer to the map.
4379
4380 The map is a vector of fixed-size entries. The size of each entry depends on
4381 the length of the longest name used. The first two bytes of each entry are the
4382 group number, most significant byte first. There follows the corresponding
4383 name, zero terminated. The names are in alphabetical order.
4384
4385 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4386 case instead of 255. Making it the same both with and without UTF-8 support
4387 means that the same test output works with both.
4388
4389 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4390 calling malloc() with a zero argument.
4391
4392 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4393 optimization. I've improved things by keeping a bitmap of backreferences with
4394 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4395 fact referenced, the optimization can be applied. It is unlikely that a
4396 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4397 the match to follow \n) will appear inside brackets with a number greater than
4398 31, but if it does, any back reference > 31 suppresses the optimization.
4399
4400 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4401 of disabling numbered capturing parentheses. Any opening parenthesis that is
4402 not followed by ? behaves as if it were followed by ?: but named parentheses
4403 can still be used for capturing (and they will acquire numbers in the usual
4404 way).
4405
4406 41. Redesigned the return codes from the match() function into yes/no/error so
4407 that errors can be passed back from deep inside the nested calls. A malloc
4408 failure while inside a recursive subpattern call now causes the
4409 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4410
4411 42. It is now possible to set a limit on the number of times the match()
4412 function is called in a call to pcre_exec(). This facility makes it possible to
4413 limit the amount of recursion and backtracking, though not in a directly
4414 obvious way, because the match() function is used in a number of different
4415 circumstances. The count starts from zero for each position in the subject
4416 string (for non-anchored patterns). The default limit is, for compatibility, a
4417 large number, namely 10 000 000. You can change this in two ways:
4418
4419 (a) When configuring PCRE before making, you can use --with-match-limit=n
4420 to set a default value for the compiled library.
4421
4422 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4423 a different value is set. See 45 below.
4424
4425 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4426
4427 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4428 of things that can be changed at compile time. The first argument specifies
4429 what is wanted and the second points to where the information is to be placed.
4430 The current list of available information is:
4431
4432 PCRE_CONFIG_UTF8
4433
4434 The output is an integer that is set to one if UTF-8 support is available;
4435 otherwise it is set to zero.
4436
4437 PCRE_CONFIG_NEWLINE
4438
4439 The output is an integer that it set to the value of the code that is used for
4440 newline. It is either LF (10) or CR (13).
4441
4442 PCRE_CONFIG_LINK_SIZE
4443
4444 The output is an integer that contains the number of bytes used for internal
4445 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4446
4447 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4448
4449 The output is an integer that contains the threshold above which the POSIX
4450 interface uses malloc() for output vectors. See item 31 above.
4451
4452 PCRE_CONFIG_MATCH_LIMIT
4453
4454 The output is an unsigned integer that contains the default limit of the number
4455 of match() calls in a pcre_exec() execution. See 42 above.
4456
4457 44. pcretest has been upgraded by the addition of the -C option. This causes it
4458 to extract all the available output from the new pcre_config() function, and to
4459 output it. The program then exits immediately.
4460
4461 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4462 order to support additional features. One way would have been to define
4463 pcre_exec2() (for example) with extra arguments, but this would not have been
4464 extensible, and would also have required all calls to the original function to
4465 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4466 is used for passing in "extra" data from pcre_study().
4467
4468 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4469 contains the following fields:
4470
4471 flags a bitmap indicating which of the following fields are set
4472 study_data opaque data from pcre_study()
4473 match_limit a way of specifying a limit on match() calls for a specific
4474 call to pcre_exec()
4475 callout_data data for callouts (see 49 below)
4476
4477 The flag bits are also defined in pcre.h, and are
4478
4479 PCRE_EXTRA_STUDY_DATA
4480 PCRE_EXTRA_MATCH_LIMIT
4481 PCRE_EXTRA_CALLOUT_DATA
4482
4483 The pcre_study() function now returns one of these new pcre_extra blocks, with
4484 the actual study data pointed to by the study_data field, and the
4485 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4486 before. That is, this change is entirely upwards-compatible and requires no
4487 change to existing code.
4488
4489 If you want to pass in additional data to pcre_exec(), you can either place it
4490 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4491 block.
4492
4493 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4494 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4495 times with different match limits, until it finds the minimum value needed for
4496 pcre_exec() to complete. The value is then output. This can be instructive; for
4497 most simple matches the number is quite small, but for pathological cases it
4498 gets very large very quickly.
4499
4500 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4501 returns the size of the data block pointed to by the study_data field in a
4502 pcre_extra block, that is, the value that was passed as the argument to
4503 pcre_malloc() when PCRE was getting memory in which to place the information
4504 created by pcre_study(). The fourth argument should point to a size_t variable.
4505 pcretest has been extended so that this information is shown after a successful
4506 pcre_study() call when information about the compiled regex is being displayed.
4507
4508 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4509 because what follows is always an absolute path. (Later: it turns out that this
4510 is more than cosmetic for MinGW, because it doesn't like empty path
4511 components.)
4512
4513 49. Some changes have been made to the callout feature (see 28 above):
4514
4515 (i) A callout function now has three choices for what it returns:
4516
4517 0 => success, carry on matching
4518 > 0 => failure at this point, but backtrack if possible
4519 < 0 => serious error, return this value from pcre_exec()
4520
4521 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4522 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4523 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4524 use by callout functions. It will never be used by PCRE itself.
4525
4526 (ii) The pcre_extra structure (see 45 above) has a void * field called
4527 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4528 pcre_callout_block structure has a field of the same name. The contents of
4529 the field passed in the pcre_extra structure are passed to the callout
4530 function in the corresponding field in the callout block. This makes it
4531 easier to use the same callout-containing regex from multiple threads. For
4532 testing, the pcretest program has a new data escape
4533
4534 \C*n pass the number n (may be negative) as callout_data
4535
4536 If the callout function in pcretest receives a non-zero value as
4537 callout_data, it returns that value.
4538
4539 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4540 there were some redundant $(CFLAGS) in commands that are now specified as
4541 $(LINK), which already includes $(CFLAGS).
4542
4543 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4544 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4545 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4546 one-byte characters throughout. Note that case-insensitive matching applies
4547 only to characters whose values are less than 256. PCRE doesn't support the
4548 notion of cases for higher-valued characters.
4549
4550 (i) A character class whose characters are all within 0-255 is handled as
4551 a bit map, and the map is inverted for negative classes. Previously, a
4552 character > 255 always failed to match such a class; however it should
4553 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4554
4555 (ii) A negated character class with a single character < 255 is coded as
4556 "not this character" (OP_NOT). This wasn't working properly when the test
4557 character was multibyte, either singly or repeated.
4558
4559 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4560 mode, for example: \x{100}{2,3}.
4561
4562 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4563 singly or repeated) now correctly test multibyte characters. However,
4564 PCRE doesn't recognize any characters with values greater than 255 as
4565 digits, spaces, or word characters. Such characters always match \D, \S,
4566 and \W, and never match \d, \s, or \w.
4567
4568 (v) Classes may now contain characters and character ranges with values
4569 greater than 255. For example: [ab\x{100}-\x{400}].
4570
4571 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4572 PCRE in UTF-8 mode.
4573
4574 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4575 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4576 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4577 value.)
4578
4579 53. The single man page has become too large. I have therefore split it up into
4580 a number of separate man pages. These also give rise to individual HTML pages;
4581 these are now put in a separate directory, and there is an index.html page that
4582 lists them all. Some hyperlinking between the pages has been installed.
4583
4584 54. Added convenience functions for handling named capturing parentheses.
4585
4586 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4587 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4588 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4589 are faulted.
4590
4591 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4592 calling configure. These values are used when compiling the dftables.c program
4593 which is run to generate the source of the default character tables. They
4594 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4595 you will need to set these values.
4596
4597 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4598
4599
4600 Version 3.9 02-Jan-02
4601 ---------------------
4602
4603 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4604
4605 2. If --disable-static was given, the building process failed when trying to
4606 build pcretest and pcregrep. (For some reason it was using libtool to compile
4607 them, which is not right, as they aren't part of the library.)
4608
4609
4610 Version 3.8 18-Dec-01
4611 ---------------------
4612
4613 1. The experimental UTF-8 code was completely screwed up. It was packing the
4614 bytes in the wrong order. How dumb can you get?
4615
4616
4617 Version 3.7 29-Oct-01
4618 ---------------------
4619
4620 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4621 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4622 this didn't happen under Solaris 8, where I normally test things.
4623
4624 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4625 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4626 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4627 this fix an hour or so after the initial 3.7 release.)
4628
4629
4630 Version 3.6 23-Oct-01
4631 ---------------------
4632
4633 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4634 offsets passed as NULL with zero offset count.
4635
4636 2. The config.guess and config.sub files had not been updated when I moved to
4637 the latest autoconf.
4638
4639
4640 Version 3.5 15-Aug-01
4641 ---------------------
4642
4643 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4644 had been forgotten.
4645
4646 2. By using declared but undefined structures, we can avoid using "void"
4647 definitions in pcre.h while keeping the internal definitions of the structures
4648 private.
4649
4650 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4651 user point of view, this means that both static and shared libraries are built
4652 by default, but this can be individually controlled. More of the work of
4653 handling this static/shared cases is now inside libtool instead of PCRE's make
4654 file.
4655
4656 4. The pcretest utility is now installed along with pcregrep because it is
4657 useful for users (to test regexs) and by doing this, it automatically gets
4658 relinked by libtool. The documentation has been turned into a man page, so
4659 there are now .1, .txt, and .html versions in /doc.
4660
4661 5. Upgrades to pcregrep:
4662 (i) Added long-form option names like gnu grep.
4663 (ii) Added --help to list all options with an explanatory phrase.
4664 (iii) Added -r, --recursive to recurse into sub-directories.
4665 (iv) Added -f, --file to read patterns from a file.
4666
4667 6. pcre_exec() was referring to its "code" argument before testing that
4668 argument for NULL (and giving an error if it was NULL).
4669
4670 7. Upgraded Makefile.in to allow for compiling in a different directory from
4671 the source directory.
4672
4673 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4674 options bits, the pointer it was passed was to an int instead of to an unsigned
4675 long int. This mattered only on 64-bit systems.
4676
4677 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4678 generated) instead of pcre.in, which it its source. Also made the same change
4679 in several of the .c files.
4680
4681 10. A new release of gcc defines printf() as a macro, which broke pcretest
4682 because it had an ifdef in the middle of a string argument for printf(). Fixed
4683 by using separate calls to printf().
4684
4685 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4686 script, to force use of CR or LF instead of \n in the source. On non-Unix
4687 systems, the value can be set in config.h.
4688
4689 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4690 absolute limit. Changed the text of the error message to make this clear, and
4691 likewise updated the man page.
4692
4693 13. The limit of 99 on the number of capturing subpatterns has been removed.
4694 The new limit is 65535, which I hope will not be a "real" limit.
4695
4696
4697 Version 3.4 22-Aug-00
4698 ---------------------
4699
4700 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4701
4702 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4703
4704
4705 Version 3.3 01-Aug-00
4706 ---------------------
4707
4708 1. If an octal character was given, but the value was greater than \377, it
4709 was not getting masked to the least significant bits, as documented. This could
4710 lead to crashes in some systems.
4711
4712 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4713 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4714
4715 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4716 These just pass their arguments on to (pcre_free)(), but they are provided
4717 because some uses of PCRE bind it to non-C systems that can call its functions,
4718 but cannot call free() or pcre_free() directly.
4719
4720 4. Add "make test" as a synonym for "make check". Corrected some comments in
4721 the Makefile.
4722
4723 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4724 Makefile.
4725
4726 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4727 command called pgrep for grepping around the active processes.
4728
4729 7. Added the beginnings of support for UTF-8 character strings.
4730
4731 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4732 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4733 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4734 out for the ar command.)
4735
4736
4737 Version 3.2 12-May-00
4738 ---------------------
4739
4740 This is purely a bug fixing release.
4741
4742 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4743 of ZA. This was just one example of several cases that could provoke this bug,
4744 which was introduced by change 9 of version 2.00. The code for breaking
4745 infinite loops after an iteration that matches an empty string was't working
4746 correctly.
4747
4748 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4749 when matched against abbab (for example). After matching an empty string, it
4750 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4751 caused it to match further down the string than it should.
4752
4753 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4754 was there because it doesn't seem to be needed, and it causes trouble on some
4755 systems, as it is not a Standard C header. It has been removed.
4756
4757 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4758 were reported on the Macintosh. The changes were from
4759
4760 while ((c = *(++ptr)) != 0 && c != '\n');
4761 to
4762 while ((c = *(++ptr)) != 0 && c != '\n') ;
4763
4764 Totally extraordinary, but if that's what it takes...
4765
4766 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4767 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4768 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4769 assumes the way PCRE uses memmove() (always moving upwards).
4770
4771 6. PCRE is being used in one environment where strchr() is not available. There
4772 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4773 faster code anyway.
4774
4775
4776 Version 3.1 09-Feb-00
4777 ---------------------
4778
4779 The only change in this release is the fixing of some bugs in Makefile.in for
4780 the "install" target:
4781
4782 (1) It was failing to install pcreposix.h.
4783
4784 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4785
4786
4787 Version 3.0 01-Feb-00
4788 ---------------------
4789
4790 1. Add support for the /+ modifier to perltest (to output $` like it does in
4791 pcretest).
4792
4793 2. Add support for the /g modifier to perltest.
4794
4795 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4796 matches null strings.
4797
4798 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4799 pattern. Perl treats empty patterns specially - it reuses the most recent
4800 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4801 effect.
4802
4803 5. The POSIX interface was broken in that it was just handing over the POSIX
4804 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4805 required a bigger vector, with some working space on the end. This means that
4806 the POSIX wrapper now has to get and free some memory, and copy the results.
4807
4808 6. Added some simple autoconf support, placing the test data and the
4809 documentation in separate directories, re-organizing some of the
4810 information files, and making it build pcre-config (a GNU standard). Also added
4811 libtool support for building PCRE as a shared library, which is now the
4812 default.
4813
4814 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4815 09 are not valid octal constants. Single digits will be used for minor values
4816 less than 10.
4817
4818 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4819 existing programs that set these in the POSIX interface can use PCRE without
4820 modification.
4821
4822 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4823 return all that pcre_info() returns, plus additional data. The pcre_info()
4824 function is retained for compatibility, but is considered to be obsolete.
4825
4826 10. Added experimental recursion feature (?R) to handle one common case that
4827 Perl 5.6 will be able to do with (?p{...}).
4828
4829 11. Added support for POSIX character classes like [:alpha:], which Perl is
4830 adopting.
4831
4832
4833 Version 2.08 31-Aug-99
4834 ----------------------
4835
4836 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4837 trying to match at the startoffset position, but instead was moving forward to
4838 the next newline as if a previous match had failed.
4839
4840 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4841 and could get into a loop if a null string was matched other than at the start
4842 of the subject.
4843
4844 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4845 be distinguished at compile time, and for completeness also added PCRE_DATE.
4846
4847 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4848 in GnuWin32 environments.
4849
4850
4851 Version 2.07 29-Jul-99
4852 ----------------------
4853
4854 1. The documentation is now supplied in plain text form and HTML as well as in
4855 the form of man page sources.
4856
4857 2. C++ compilers don't like assigning (void *) values to other pointer types.
4858 In particular this affects malloc(). Although there is no problem in Standard
4859 C, I've put in casts to keep C++ compilers happy.
4860
4861 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4862 should be (const char *).
4863
4864 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4865 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4866 However, I haven't made this a standard facility. The documentation doesn't
4867 mention it, and the Makefile doesn't support it.
4868
4869 5. The Makefile now contains an "install" target, with editable destinations at
4870 the top of the file. The pcretest program is not installed.
4871
4872 6. pgrep -V now gives the PCRE version number and date.
4873
4874 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4875 causing the entire string to be ignored, instead of just the last character.
4876
4877 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4878 non-matching string, it can take a very, very long time, even for strings of
4879 quite modest length, because of the nested recursion. PCRE now does better in
4880 some of these cases. It does this by remembering the last required literal
4881 character in the pattern, and pre-searching the subject to ensure it is present
4882 before running the real match. In other words, it applies a heuristic to detect
4883 some types of certain failure quickly, and in the above example, if presented
4884 with a string that has no trailing " it gives "no match" very quickly.
4885
4886 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4887 other alternatives are tried instead.
4888
4889
4890 Version 2.06 09-Jun-99
4891 ----------------------
4892
4893 1. Change pcretest's output for amount of store used to show just the code
4894 space, because the remainder (the data block) varies in size between 32-bit and
4895 64-bit systems.
4896
4897 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4898 start matching at. This allows lookbehinds to work when searching for multiple
4899 occurrences in a string.
4900
4901 3. Added additional options to pcretest for testing multiple occurrences:
4902
4903 /+ outputs the rest of the string that follows a match
4904 /g loops for multiple occurrences, using the new startoffset argument
4905 /G loops for multiple occurrences by passing an incremented pointer
4906
4907 4. PCRE wasn't doing the "first character" optimization for patterns starting
4908 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4909 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4910 the letter 'x'. On long subject strings, this gives a significant speed-up.
4911
4912
4913 Version 2.05 21-Apr-99
4914 ----------------------
4915
4916 1. Changed the type of magic_number from int to long int so that it works
4917 properly on 16-bit systems.
4918
4919 2. Fixed a bug which caused patterns starting with .* not to work correctly
4920 when the subject string contained newline characters. PCRE was assuming
4921 anchoring for such patterns in all cases, which is not correct because .* will
4922 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4923 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4924 must be retried after every newline in the subject.
4925
4926
4927 Version 2.04 18-Feb-99
4928 ----------------------
4929
4930 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4931 computation of the store needed to hold the pattern was incorrect (too large).
4932 If such patterns were nested a few deep, this could multiply and become a real
4933 problem.
4934
4935 2. Added /M option to pcretest to show the memory requirement of a specific
4936 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4937
4938 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4939 compiled in such a way that the backtracking after subsequent failure was
4940 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4941 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4942
4943
4944 Version 2.03 02-Feb-99
4945 ----------------------
4946
4947 1. Fixed typo and small mistake in man page.
4948
4949 2. Added 4th condition (GPL supersedes if conflict) and created separate
4950 LICENCE file containing the conditions.
4951
4952 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4953 Perl, that is the internal \ allows the delimiter to be included in the
4954 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4955 the final delimiter, add \ to the end of the pattern (to test the error).
4956
4957 4. Added the convenience functions for extracting substrings after a successful
4958 match. Updated pcretest to make it able to test these functions.
4959
4960
4961 Version 2.02 14-Jan-99
4962 ----------------------
4963
4964 1. Initialized the working variables associated with each extraction so that
4965 their saving and restoring doesn't refer to uninitialized store.
4966
4967 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4968 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4969 fix the problem.
4970
4971 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4972 calls, and wasn't printing the correct value for compiling calls. Increased the
4973 default value of LOOPREPEAT, and the number of significant figures in the
4974 times.
4975
4976 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4977
4978 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4979 a building problem on Windows NT with a FAT file system.
4980
4981
4982 Version 2.01 21-Oct-98
4983 ----------------------
4984
4985 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4986 to character tables built by pcre_maketables() in the current locale. If NULL
4987 is passed, the default tables are used.
4988
4989
4990 Version 2.00 24-Sep-98
4991 ----------------------
4992
4993 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4994 it any more.
4995
4996 2. Allow quantification of (?>) groups, and make it work correctly.
4997
4998 3. The first character computation wasn't working for (?>) groups.
4999
5000 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5001 end of the subject) and add 5.005's \z, which really does match only at the
5002 very end of the subject.
5003
5004 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5005
5006 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5007 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5008 localized options. All options to pcre_study() were also removed.
5009
5010 7. Add other new features from 5.005:
5011
5012 $(?<= positive lookbehind
5013 $(?<! negative lookbehind
5014 (?imsx-imsx) added the unsetting capability
5015 such a setting is global if at outer level; local otherwise
5016 (?imsx-imsx:) non-capturing groups with option setting
5017 (?(cond)re|re) conditional pattern matching
5018
5019 A backreference to itself in a repeated group matches the previous
5020 captured string.
5021
5022 8. General tidying up of studying (both automatic and via "study")
5023 consequential on the addition of new assertions.
5024
5025 9. As in 5.005, unlimited repeated groups that could match an empty substring
5026 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5027 runtime if any iteration does actually match an empty substring.
5028
5029 10. Include the RunTest script in the distribution.
5030
5031 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5032 discrepancies, some of which were old and were also with respect to 5.004. They
5033 have now been fixed.
5034
5035
5036 Version 1.09 28-Apr-98
5037 ----------------------
5038
5039 1. A negated single character class followed by a quantifier with a minimum
5040 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5041 program crashes, or just wrong answers. This did not apply to negated classes
5042 containing more than one character, or to minima other than one.
5043
5044
5045 Version 1.08 27-Mar-98
5046 ----------------------
5047
5048 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5049
5050 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5051 latter must appear before anything that relies on it in the pattern.
5052
5053
5054 Version 1.07 16-Feb-98
5055 ----------------------
5056
5057 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5058 repeat of a potentially empty string).
5059
5060
5061 Version 1.06 23-Jan-98
5062 ----------------------
5063
5064 1. Added Markus Oberhumer's little patches for C++.
5065
5066 2. Literal strings longer than 255 characters were broken.
5067
5068
5069 Version 1.05 23-Dec-97
5070 ----------------------
5071
5072 1. Negated character classes containing more than one character were failing if
5073 PCRE_CASELESS was set at run time.
5074
5075
5076 Version 1.04 19-Dec-97
5077 ----------------------
5078
5079 1. Corrected the man page, where some "const" qualifiers had been omitted.
5080
5081 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5082 input syntax.
5083
5084 3. Fixed memory leak which occurred when a regex with back references was
5085 matched with an offsets vector that wasn't big enough. The temporary memory
5086 that is used in this case wasn't being freed if the match failed.
5087
5088 4. Tidied pcretest to ensure it frees memory that it gets.
5089
5090 5. Temporary memory was being obtained in the case where the passed offsets
5091 vector was exactly big enough.
5092
5093 6. Corrected definition of offsetof() from change 5 below.
5094
5095 7. I had screwed up change 6 below and broken the rules for the use of
5096 setjmp(). Now fixed.
5097
5098
5099 Version 1.03 18-Dec-97
5100 ----------------------
5101
5102 1. A erroneous regex with a missing opening parenthesis was correctly
5103 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5104 on some systems.
5105
5106 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5107 it was reported that one broken compiler failed on the former because "code" is
5108 also an independent variable.
5109
5110 3. The erroneous regex a[]b caused an array overrun reference.
5111
5112 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5113 fail on data ending with that character. (It was going on too far, and checking
5114 the next character, typically a binary zero.) This was specific to the
5115 optimized code for single-character negative classes.
5116
5117 5. Added a contributed patch from the TIN world which does the following:
5118
5119 + Add an undef for memmove, in case the the system defines a macro for it.
5120
5121 + Add a definition of offsetof(), in case there isn't one. (I don't know
5122 the reason behind this - offsetof() is part of the ANSI standard - but
5123 it does no harm).
5124
5125 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5126 most of the places where whitespace preceded '#'. I have given up and
5127 allowed the remaining 2 cases to be at the margin.
5128
5129 + Rename some variables in pcre to eliminate shadowing. This seems very
5130 pedantic, but does no harm, of course.
5131
5132 6. Moved the call to setjmp() into its own function, to get rid of warnings
5133 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5134
5135 7. Constructs such as \d{8,} were compiling into the equivalent of
5136 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5137 outcome, but in this particular case used more store than had been allocated,
5138 which caused the bug to be discovered because it threw up an internal error.
5139
5140 8. The debugging code in both pcre and pcretest for outputting the compiled
5141 form of a regex was going wrong in the case of back references followed by
5142 curly-bracketed repeats.
5143
5144
5145 Version 1.02 12-Dec-97
5146 ----------------------
5147
5148 1. Typos in pcre.3 and comments in the source fixed.
5149
5150 2. Applied a contributed patch to get rid of places where it used to remove
5151 'const' from variables, and fixed some signed/unsigned and uninitialized
5152 variable warnings.
5153
5154 3. Added the "runtest" target to Makefile.
5155
5156 4. Set default compiler flag to -O2 rather than just -O.
5157
5158
5159 Version 1.01 19-Nov-97
5160 ----------------------
5161
5162 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5163 like /([ab]*)*/, that is, for classes with more than one character in them.
5164
5165 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5166 as /((?>a*))*/ (a PCRE_EXTRA facility).
5167
5168
5169 Version 1.00 18-Nov-97
5170 ----------------------
5171
5172 1. Added compile-time macros to support systems such as SunOS4 which don't have
5173 memmove() or strerror() but have other things that can be used instead.
5174
5175 2. Arranged that "make clean" removes the executables.
5176
5177
5178 Version 0.99 27-Oct-97
5179 ----------------------
5180
5181 1. Fixed bug in code for optimizing classes with only one character. It was
5182 initializing a 32-byte map regardless, which could cause it to run off the end
5183 of the memory it had got.
5184
5185 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5186
5187
5188 Version 0.98 22-Oct-97
5189 ----------------------
5190
5191 1. Fixed bug in code for handling temporary memory usage when there are more
5192 back references than supplied space in the ovector. This could cause segfaults.
5193
5194
5195 Version 0.97 21-Oct-97
5196 ----------------------
5197
5198 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5199
5200 2. Optimized negated single characters not to use a bit map.
5201
5202 3. Brought error texts together as macro definitions; clarified some of them;
5203 fixed one that was wrong - it said "range out of order" when it meant "invalid
5204 escape sequence".
5205
5206 4. Changed some char * arguments to const char *.
5207
5208 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5209
5210 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5211 pcretest.
5212
5213
5214 Version 0.96 16-Oct-97
5215 ----------------------
5216
5217 1. Added a simple "pgrep" utility to the distribution.
5218
5219 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5220 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5221 where "ddd" means "one or more decimal digits".
5222
5223 3. Fixed serious bug. If a pattern had a back reference, but the call to
5224 pcre_exec() didn't supply a large enough ovector to record the related
5225 identifying subpattern, the match always failed. PCRE now remembers the number
5226 of the largest back reference, and gets some temporary memory in which to save
5227 the offsets during matching if necessary, in order to ensure that
5228 backreferences always work.
5229
5230 4. Increased the compatibility with Perl in a number of ways:
5231
5232 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5233 to request this handling. The option can be set at compile or exec time.
5234
5235 (b) $ matches before a terminating newline by default; an option
5236 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5237 mode). The option can be set at compile or exec time.
5238
5239 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5240 the same as Perl's. If the decimal number it represents is less than 10
5241 or there aren't that many previous left capturing parentheses, an octal
5242 escape is read. Inside a character class, it's always an octal escape,
5243 even if it is a single digit.
5244
5245 (d) An escaped but undefined alphabetic character is taken as a literal,
5246 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5247 escapes.
5248
5249 (e) {0} is now permitted. (The previous item is removed from the compiled
5250 pattern).
5251
5252 5. Changed all the names of code files so that the basic parts are no longer
5253 than 10 characters, and abolished the teeny "globals.c" file.
5254
5255 6. Changed the handling of character classes; they are now done with a 32-byte
5256 bit map always.
5257
5258 7. Added the -d and /D options to pcretest to make it possible to look at the
5259 internals of compilation without having to recompile pcre.
5260
5261
5262 Version 0.95 23-Sep-97
5263 ----------------------
5264
5265 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5266 \x20 at the start of a run of normal characters. These were being treated as
5267 real characters, instead of the source characters being re-checked.
5268
5269
5270 Version 0.94 18-Sep-97
5271 ----------------------
5272
5273 1. The functions are now thread-safe, with the caveat that the global variables
5274 containing pointers to malloc() and free() or alternative functions are the
5275 same for all threads.
5276
5277 2. Get pcre_study() to generate a bitmap of initial characters for non-
5278 anchored patterns when this is possible, and use it if passed to pcre_exec().
5279
5280
5281 Version 0.93 15-Sep-97
5282 ----------------------
5283
5284 1. /(b)|(:+)/ was computing an incorrect first character.
5285
5286 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5287 but not actually doing anything yet.
5288
5289 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5290 as Perl does (e.g. [-az] or [az-]).
5291
5292 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5293 all possible positions.
5294
5295 5. Split up into different modules to avoid including unneeded functions in a
5296 compiled binary. However, compile and exec are still in one module. The "study"
5297 function is split off.
5298
5299 6. The character tables are now in a separate module whose source is generated
5300 by an auxiliary program - but can then be edited by hand if required. There are
5301 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5302 toupper() in the code.
5303
5304 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5305 make them global. Abolish the function for setting them, as the caller can now
5306 set them directly.
5307
5308
5309 Version 0.92 11-Sep-97
5310 ----------------------
5311
5312 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5313 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5314
5315 2. Caseless matching was not working in character classes if the characters in
5316 the pattern were in upper case.
5317
5318 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5319
5320 4. Make PCRE_ANCHORED public and accept as a compile option.
5321
5322 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5323 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5324 pass them.
5325
5326 6. Give an error if bad option bits passed at compile or run time.
5327
5328 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5329 pcretest to cause it to pass that flag.
5330
5331 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5332 options, and the first character, if set.
5333
5334 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5335
5336
5337 Version 0.91 10-Sep-97
5338 ----------------------
5339
5340 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5341 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5342
5343 2. PCRE was looping on encountering an indefinitely repeated back reference to
5344 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5345 Perl does - treats the match as successful.
5346
5347 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5