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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1414 - (show annotations)
Sun Dec 22 16:27:35 2013 UTC (6 years, 1 month ago) by zherczeg
File size: 258189 byte(s)
Error occurred while calculating annotation data.
A new flag is set, when property checks are present in an XCLASS.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.35-RC1 xx-xxxx-201x
5 -----------------------------
6
7 1. A new flag is set, when property checks are present in an XCLASS.
8 When this flag is not set, PCRE can perform certain optimizations
9 such as studying these XCLASS-es.
10
11
12 Version 8.34 15-December-2013
13 -----------------------------
14
15 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
16 executable memory. Patch inspired by Carsten Klein.
17
18 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
19 this macro is never tested and has no effect, because the work to support
20 coverage involves only compiling and linking options and special targets in
21 the Makefile. The comment in config.h implied that defining the macro would
22 enable coverage support, which is totally false. There was also support for
23 setting this macro in the CMake files (my fault, I just copied it from
24 configure). SUPPORT_GCOV has now been removed.
25
26 3. Make a small performance improvement in strlen16() and strlen32() in
27 pcretest.
28
29 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
30 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
31 these cases). There was also one in pcretest.c.
32
33 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
34
35 6. In UTF mode, the code for checking whether a group could match an empty
36 string (which is used for indefinitely repeated groups to allow for
37 breaking an infinite loop) was broken when the group contained a repeated
38 negated single-character class with a character that occupied more than one
39 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
40 UTF-8 mode). The effect was undefined: the group might or might not be
41 deemed as matching an empty string, or the program might have crashed.
42
43 7. The code for checking whether a group could match an empty string was not
44 recognizing that \h, \H, \v, \V, and \R must match a character.
45
46 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
47 an empty string. If it can, pcretest shows this in its information output.
48
49 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
50 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
51 matched by pcre_exec() without using JIT:
52
53 (a) If the rest of the pattern did not match after a maximal run of
54 grapheme clusters, the code for backing up to try with fewer of them
55 did not always back up over a full grapheme when characters that do not
56 have the modifier quality were involved, e.g. Hangul syllables.
57
58 (b) If the match point in a subject started with modifier character, and
59 there was no match, the code could incorrectly back up beyond the match
60 point, and potentially beyond the first character in the subject,
61 leading to a segfault or an incorrect match result.
62
63 10. A conditional group with an assertion condition could lead to PCRE
64 recording an incorrect first data item for a match if no other first data
65 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
66 first data item, and therefore matched "ca" after "c" instead of at the
67 start.
68
69 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
70 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
71
72 12. The source of pcregrep now includes z/OS-specific code so that it can be
73 compiled for z/OS as part of the special z/OS distribution.
74
75 13. Added the -T and -TM options to pcretest.
76
77 14. The code in pcre_compile.c for creating the table of named capturing groups
78 has been refactored. Instead of creating the table dynamically during the
79 actual compiling pass, the information is remembered during the pre-compile
80 pass (on the stack unless there are more than 20 named groups, in which
81 case malloc() is used) and the whole table is created before the actual
82 compile happens. This has simplified the code (it is now nearly 150 lines
83 shorter) and prepared the way for better handling of references to groups
84 with duplicate names.
85
86 15. A back reference to a named subpattern when there is more than one of the
87 same name now checks them in the order in which they appear in the pattern.
88 The first one that is set is used for the reference. Previously only the
89 first one was inspected. This change makes PCRE more compatible with Perl.
90
91 16. Unicode character properties were updated from Unicode 6.3.0.
92
93 17. The compile-time code for auto-possessification has been refactored, based
94 on a patch by Zoltan Herczeg. It now happens after instead of during
95 compilation. The code is cleaner, and more cases are handled. The option
96 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
97 options in pcretest are provided to set it. It can also be set by
98 (*NO_AUTO_POSSESS) at the start of a pattern.
99
100 18. The character VT has been added to the default ("C" locale) set of
101 characters that match \s and are generally treated as white space,
102 following this same change in Perl 5.18. There is now no difference between
103 "Perl space" and "POSIX space". Whether VT is treated as white space in
104 other locales depends on the locale.
105
106 19. The code for checking named groups as conditions, either for being set or
107 for being recursed, has been refactored (this is related to 14 and 15
108 above). Processing unduplicated named groups should now be as fast at
109 numerical groups, and processing duplicated groups should be faster than
110 before.
111
112 20. Two patches to the CMake build system, by Alexander Barkov:
113
114 (1) Replace the "source" command by "." in CMakeLists.txt because
115 "source" is a bash-ism.
116
117 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
118 without these the CMake build does not work on Solaris.
119
120 21. Perl has changed its handling of \8 and \9. If there is no previously
121 encountered capturing group of those numbers, they are treated as the
122 literal characters 8 and 9 instead of a binary zero followed by the
123 literals. PCRE now does the same.
124
125 22. Following Perl, added \o{} to specify codepoints in octal, making it
126 possible to specify values greater than 0777 and also making them
127 unambiguous.
128
129 23. Perl now gives an error for missing closing braces after \x{... instead of
130 treating the string as literal. PCRE now does the same.
131
132 24. RunTest used to grumble if an inappropriate test was selected explicitly,
133 but just skip it when running all tests. This make it awkward to run ranges
134 of tests when one of them was inappropriate. Now it just skips any
135 inappropriate tests, as it always did when running all tests.
136
137 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
138 character types such as \d or \w, too many callouts were inserted, and the
139 data that they returned was rubbish.
140
141 26. In UCP mode, \s was not matching two of the characters that Perl matches,
142 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
143 were matched by \h. The code has now been refactored so that the lists of
144 the horizontal and vertical whitespace characters used for \h and \v (which
145 are defined only in one place) are now also used for \s.
146
147 27. Add JIT support for the 64 bit TileGX architecture.
148 Patch by Jiong Wang (Tilera Corporation).
149
150 28. Possessive quantifiers for classes (both explicit and automatically
151 generated) now use special opcodes instead of wrapping in ONCE brackets.
152
153 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
154 (because it's meaningless), this was not happening when PCRE_CASELESS was
155 set. Not wrong, but inefficient.
156
157 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
158 properties for \w, \d, etc) is present in a test regex. Otherwise if the
159 test contains no characters greater than 255, Perl doesn't realise it
160 should be using Unicode semantics.
161
162 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
163 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
164 does in Unicode mode.
165
166 32. Added the "forbid" facility to pcretest so that putting tests into the
167 wrong test files can sometimes be quickly detected.
168
169 33. There is now a limit (default 250) on the depth of nesting of parentheses.
170 This limit is imposed to control the amount of system stack used at compile
171 time. It can be changed at build time by --with-parens-nest-limit=xxx or
172 the equivalent in CMake.
173
174 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
175 errors. Perl warns for these when in warning mode, but PCRE has no facility
176 for giving warnings.
177
178 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
179 However, this was not working for (?!) because it is optimized to (*FAIL),
180 for which PCRE does not allow quantifiers. The optimization is now disabled
181 when a quantifier follows (?!). I can't see any use for this, but it makes
182 things uniform.
183
184 36. Perl no longer allows group names to start with digits, so I have made this
185 change also in PCRE. It simplifies the code a bit.
186
187 37. In extended mode, Perl ignores spaces before a + that indicates a
188 possessive quantifier. PCRE allowed a space before the quantifier, but not
189 before the possessive +. It now does.
190
191 38. The use of \K (reset reported match start) within a repeated possessive
192 group such as (a\Kb)*+ was not working.
193
194 40. Document that the same character tables must be used at compile time and
195 run time, and that the facility to pass tables to pcre_exec() and
196 pcre_dfa_exec() is for use only with saved/restored patterns.
197
198 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
199 features for Builds with MSVC:
200
201 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
202 builds). The .rc files can be used to set FileDescription and many other
203 attributes.
204
205 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
206 This allows higher-level build scripts which want .pdb files to avoid
207 hard-coding the exact files needed."
208
209 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
210 mean "start of word" and "end of word", respectively, as a transition aid.
211
212 43. A minimizing repeat of a class containing codepoints greater than 255 in
213 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
214 compiled to use the heap for recursion.
215
216 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
217 is configured.
218
219
220 Version 8.33 28-May-2013
221 ------------------------
222
223 1. Added 'U' to some constants that are compared to unsigned integers, to
224 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
225 variables that are added to signed variables, to ensure the result is
226 signed and can be negated.
227
228 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
229 CMake config files.
230
231 3. Revise the creation of config.h.generic so that all boolean macros are
232 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
233 overriding via -D on the command line possible.
234
235 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
236 to unsigned int is reported to make a quite noticeable speed difference in
237 a specific Windows environment. Testing on Linux did also appear to show
238 some benefit (and it is clearly not harmful). Also fixed the definition of
239 Xop which should be unsigned.
240
241 5. Related to (4), changing the definition of the intermediate variable cc
242 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
243 improvements.
244
245 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
246 unnecessary spaces.
247
248 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
249 and later.
250
251 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
252
253 9. Optimizing fast_forward_start_bits in JIT.
254
255 10. Adding support for callouts in JIT, and fixing some issues revealed
256 during this work. Namely:
257
258 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
259
260 (b) Minimum length was not checked before the matching is started.
261
262 11. The value of capture_last that is passed to callouts was incorrect in some
263 cases when there was a capture on one path that was subsequently abandoned
264 after a backtrack. Also, the capture_last value is now reset after a
265 recursion, since all captures are also reset in this case.
266
267 12. The interpreter no longer returns the "too many substrings" error in the
268 case when an overflowing capture is in a branch that is subsequently
269 abandoned after a backtrack.
270
271 13. In the pathological case when an offset vector of size 2 is used, pcretest
272 now prints out the matched string after a yield of 0 or 1.
273
274 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
275 Only supported by the JIT compiler at the moment.
276
277 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
278
279 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
280 offset of the starting point of the matching process, provided the offsets
281 vector is large enough.
282
283 17. The \A escape now records a lookbehind value of 1, though its execution
284 does not actually inspect the previous character. This is to ensure that,
285 in partial multi-segment matching, at least one character from the old
286 segment is retained when a new segment is processed. Otherwise, if there
287 are no lookbehinds in the pattern, \A might match incorrectly at the start
288 of a new segment.
289
290 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
291
292 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
293 gives some modest performance improvement in 8-bit mode.
294
295 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
296 be expressed in certain programming languages using Universal Character
297 Names.
298
299 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
300 which points out that "non characters" are not "characters that may not
301 appear in Unicode strings" but rather "characters that are reserved for
302 internal use and have only local meaning".
303
304 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
305 there was a conditional group that depended on an assertion, if the
306 assertion was false, the callout that immediately followed the alternation
307 in the condition was skipped when pcre_exec() was used for matching.
308
309 23. Allow an explicit callout to be inserted before an assertion that is the
310 condition for a conditional group, for compatibility with automatic
311 callouts, which always insert a callout at this point.
312
313 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
314 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
315
316 25. (*PRUNE) is now supported by the JIT compiler.
317
318 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
319
320 27. Fix the case where there are two or more SKIPs with arguments that may be
321 ignored.
322
323 28. (*SKIP) is now supported by the JIT compiler.
324
325 29. (*THEN) is now supported by the JIT compiler.
326
327 30. Update RunTest with additional test selector options.
328
329 31. The way PCRE handles backtracking verbs has been changed in two ways.
330
331 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
332 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
333 backtracking. In some cases this makes it more Perl-compatible, but Perl's
334 rather obscure rules do not always do the same thing.
335
336 (2) Previously, backtracking verbs were confined within assertions. This is
337 no longer the case for positive assertions, except for (*ACCEPT). Again,
338 this sometimes improves Perl compatibility, and sometimes does not.
339
340 32. A number of tests that were in test 2 because Perl did things differently
341 have been moved to test 1, because either Perl or PCRE has changed, and
342 these tests are now compatible.
343
344 32. Backtracking control verbs are now handled in the same way in JIT and
345 interpreter.
346
347 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
348 contained a forward subroutine reference caused a compile error.
349
350 34. Auto-detect and optimize limited repetitions in JIT.
351
352 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
353 blocking (*UTF) etc.
354
355 36. In the interpreter, maximizing pattern repetitions for characters and
356 character types now use tail recursion, which reduces stack usage.
357
358 37. The value of the max lookbehind was not correctly preserved if a compiled
359 and saved regex was reloaded on a host of different endianness.
360
361 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
362 of the compiled pattern block, expand the flags field from 16 to 32 bits
363 because it was almost full.
364
365 39. Try madvise first before posix_madvise.
366
367 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
368 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
369 original change locked out all matches of empty strings. This has been
370 changed so that one match of an empty string per line is recognized.
371 Subsequent searches on the same line (for colouring or for --only-matching,
372 for example) do not recognize empty strings.
373
374 41. Applied a user patch to fix a number of spelling mistakes in comments.
375
376 42. Data lines longer than 65536 caused pcretest to crash.
377
378 43. Clarified the data type for length and startoffset arguments for pcre_exec
379 and pcre_dfa_exec in the function-specific man pages, where they were
380 explicitly stated to be in bytes, never having been updated. I also added
381 some clarification to the pcreapi man page.
382
383 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
384 a segmentation fault.
385
386
387 Version 8.32 30-November-2012
388 -----------------------------
389
390 1. Improved JIT compiler optimizations for first character search and single
391 character iterators.
392
393 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
394 Patch by Daniel Richard G.
395
396 3. Single character iterator optimizations in the JIT compiler.
397
398 4. Improved JIT compiler optimizations for character ranges.
399
400 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
401 Reported by Giuseppe D'Angelo.
402
403 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
404 of a line, was being set incorrectly in cases where .* appeared inside
405 atomic brackets at the start of a pattern, or where there was a subsequent
406 *PRUNE or *SKIP.
407
408 7. Improved instruction cache flush for POWER/PowerPC.
409 Patch by Daniel Richard G.
410
411 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
412 grep:
413
414 (a) There is now no limit to the number of patterns to be matched.
415
416 (b) An error is given if a pattern is too long.
417
418 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
419 are now supported.
420
421 (d) --exclude-from and --include-from (multiple use) have been added.
422
423 (e) Exclusions and inclusions now apply to all files and directories, not
424 just to those obtained from scanning a directory recursively.
425
426 (f) Multiple uses of -f and --file-list are now supported.
427
428 (g) In a Windows environment, the default for -d has been changed from
429 "read" (the GNU grep default) to "skip", because otherwise the presence
430 of a directory in the file list provokes an error.
431
432 (h) The documentation has been revised and clarified in places.
433
434 9. Improve the matching speed of capturing brackets.
435
436 10. Changed the meaning of \X so that it now matches a Unicode extended
437 grapheme cluster.
438
439 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
440 out POSIX threads when JIT support is configured.
441
442 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
443
444 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
445 structure could go wrong in environments where size_t is not the same size
446 as int.
447
448 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
449 set.
450
451 15. The EBCDIC support had decayed; later updates to the code had included
452 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
453 general tidy up of EBCDIC-related issues, and the documentation was also
454 not quite right. There is now a test that can be run on ASCII systems to
455 check some of the EBCDIC-related things (but is it not a full test).
456
457 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
458 in a small tidy to the code.
459
460 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
461
462 18. If the --only-matching (-o) option in pcregrep is specified multiple
463 times, each one causes appropriate output. For example, -o1 -o2 outputs the
464 substrings matched by the 1st and 2nd capturing parentheses. A separating
465 string can be specified by --om-separator (default empty).
466
467 19. Improving the first n character searches.
468
469 20. Turn case lists for horizontal and vertical white space into macros so that
470 they are defined only once.
471
472 21. This set of changes together give more compatible Unicode case-folding
473 behaviour for characters that have more than one other case when UCP
474 support is available.
475
476 (a) The Unicode property table now has offsets into a new table of sets of
477 three or more characters that are case-equivalent. The MultiStage2.py
478 script that generates these tables (the pcre_ucd.c file) now scans
479 CaseFolding.txt instead of UnicodeData.txt for character case
480 information.
481
482 (b) The code for adding characters or ranges of characters to a character
483 class has been abstracted into a generalized function that also handles
484 case-independence. In UTF-mode with UCP support, this uses the new data
485 to handle characters with more than one other case.
486
487 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
488 whose other case is greater than 256 are now correctly matched
489 caselessly. Previously, the high codepoint matched the low one, but not
490 vice versa.
491
492 (d) The processing of \h, \H, \v, and \ in character classes now makes use
493 of the new class addition function, using character lists defined as
494 macros alongside the case definitions of 20 above.
495
496 (e) Caseless back references now work with characters that have more than
497 one other case.
498
499 (f) General caseless matching of characters with more than one other case
500 is supported.
501
502 22. Unicode character properties were updated from Unicode 6.2.0
503
504 23. Improved CMake support under Windows. Patch by Daniel Richard G.
505
506 24. Add support for 32-bit character strings, and UTF-32
507
508 25. Major JIT compiler update (code refactoring and bugfixing).
509 Experimental Sparc 32 support is added.
510
511 26. Applied a modified version of Daniel Richard G's patch to create
512 pcre.h.generic and config.h.generic by "make" instead of in the
513 PrepareRelease script.
514
515 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
516 pcre_compile.c when checking for a zero character.
517
518 28. Introducing a native interface for JIT. Through this interface, the compiled
519 machine code can be directly executed. The purpose of this interface is to
520 provide fast pattern matching, so several sanity checks are not performed.
521 However, feature tests are still performed. The new interface provides
522 1.4x speedup compared to the old one.
523
524 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
525 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
526 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
527
528 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
529 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
530 the "old" RFC 2279). Instead, it ended up passing a negative length to
531 pcre_exec().
532
533 31. Add support for GCC's visibility feature to hide internal functions.
534
535 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
536 "unknown -C option" after outputting 0 or 1.
537
538 33. There is now support for generating a code coverage report for the test
539 suite in environments where gcc is the compiler and lcov is installed. This
540 is mainly for the benefit of the developers.
541
542 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
543 unaddressable using valgrind annotations, allowing valgrind to detect
544 invalid memory accesses. This is mainly for the benefit of the developers.
545
546 25. (*UTF) can now be used to start a pattern in any of the three libraries.
547
548 26. Give configure error if --enable-cpp but no C++ compiler found.
549
550
551 Version 8.31 06-July-2012
552 -------------------------
553
554 1. Fixing a wrong JIT test case and some compiler warnings.
555
556 2. Removed a bashism from the RunTest script.
557
558 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
559 to unsigned type, result still unsigned" that was given by an MS compiler
560 on encountering the code "-sizeof(xxx)".
561
562 4. Partial matching support is added to the JIT compiler.
563
564 5. Fixed several bugs concerned with partial matching of items that consist
565 of more than one character:
566
567 (a) /^(..)\1/ did not partially match "aba" because checking references was
568 done on an "all or nothing" basis. This also applied to repeated
569 references.
570
571 (b) \R did not give a hard partial match if \r was found at the end of the
572 subject.
573
574 (c) \X did not give a hard partial match after matching one or more
575 characters at the end of the subject.
576
577 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
578 a partial match for the string "\r".
579
580 (e) When newline was set to CRLF, the metacharacter "." did not recognize
581 a partial match for a CR character at the end of the subject string.
582
583 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
584 running pcretest, the text "(JIT)" added to the output whenever JIT is
585 actually used to run the match.
586
587 7. Individual JIT compile options can be set in pcretest by following -s+[+]
588 or /S+[+] with a digit between 1 and 7.
589
590 8. OP_NOT now supports any UTF character not just single-byte ones.
591
592 9. (*MARK) control verb is now supported by the JIT compiler.
593
594 10. The command "./RunTest list" lists the available tests without actually
595 running any of them. (Because I keep forgetting what they all are.)
596
597 11. Add PCRE_INFO_MAXLOOKBEHIND.
598
599 12. Applied a (slightly modified) user-supplied patch that improves performance
600 when the heap is used for recursion (compiled with --disable-stack-for-
601 recursion). Instead of malloc and free for each heap frame each time a
602 logical recursion happens, frames are retained on a chain and re-used where
603 possible. This sometimes gives as much as 30% improvement.
604
605 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
606 call.
607
608 14. As documented, (*COMMIT) is now confined to within a positive assertion.
609
610 15. It is now possible to link pcretest with libedit as an alternative to
611 libreadline.
612
613 16. (*COMMIT) control verb is now supported by the JIT compiler.
614
615 17. The Unicode data tables have been updated to Unicode 6.1.0.
616
617 18. Added --file-list option to pcregrep.
618
619 19. Added binary file support to pcregrep, including the -a, --binary-files,
620 -I, and --text options.
621
622 20. The madvise function is renamed for posix_madvise for QNX compatibility
623 reasons. Fixed by Giuseppe D'Angelo.
624
625 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
626
627 22. Changed the option for creating symbolic links for 16-bit man pages from
628 -s to -sf so that re-installing does not cause issues.
629
630 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
631
632 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
633 matches in certain environments (the workspace was not being correctly
634 retained). Also added to pcre_dfa_exec() a simple plausibility check on
635 some of the workspace data at the beginning of a restart.
636
637 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
638 was not doing so when it should - probably a typo introduced by SVN 528
639 (change 8.10/14).
640
641 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
642 \w+ when the character tables indicated that \x{c4} was a word character.
643 There were several related cases, all because the tests for doing a table
644 lookup were testing for characters less than 127 instead of 255.
645
646 27. If a pattern contains capturing parentheses that are not used in a match,
647 their slots in the ovector are set to -1. For those that are higher than
648 any matched groups, this happens at the end of processing. In the case when
649 there were back references that the ovector was too small to contain
650 (causing temporary malloc'd memory to be used during matching), and the
651 highest capturing number was not used, memory off the end of the ovector
652 was incorrectly being set to -1. (It was using the size of the temporary
653 memory instead of the true size.)
654
655 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
656 ovector size, it uses memory at the end of the block that it has got.
657
658 29. Check for an overlong MARK name and give an error at compile time. The
659 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
660
661 30. JIT compiler update.
662
663 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
664 Rill for the patch.
665
666 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
667
668 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
669
670 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
671 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
672 was enabled.
673
674 35. Improve JIT code generation for greedy plus quantifier.
675
676 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
677 1 to "aa" instead of to an empty string. The bug affected repeated groups
678 that could potentially match an empty string.
679
680 37. Optimizing single character iterators in JIT.
681
682 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
683 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
684 codepoints that are too big for the mode are faulted, and in a UTF mode,
685 disallowed codepoints are also faulted.
686
687 39. If PCRE was compiled with UTF support, in three places in the DFA
688 matcher there was code that should only have been obeyed in UTF mode, but
689 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
690 processing when bytes with values greater than 127 were present. In 16-bit
691 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
692 both cases the values are those that cannot be the first data item in a UTF
693 character. The three items that might have provoked this were recursions,
694 possessively repeated groups, and atomic groups.
695
696 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
697 and pcregrep, because some OS require shared objects to be explicitly
698 passed to ld, causing the link step to fail if they are not.
699
700 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
701 mode, patterns that started with \h* or \R* might be incorrectly matched.
702
703
704 Version 8.30 04-February-2012
705 -----------------------------
706
707 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
708 name is defined in ctype.h.
709
710 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
711 only in quite long subpatterns.
712
713 3. Removed the function pcre_info(), which has been obsolete and deprecated
714 since it was replaced by pcre_fullinfo() in February 2000.
715
716 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
717 match a (*MARK), and the match failed at the start of the subject, a
718 reference to memory before the start of the subject could occur. This bug
719 was introduced by fix 17 of release 8.21.
720
721 5. A reference to an unset group with zero minimum repetition was giving
722 totally wrong answers (in non-JavaScript-compatibility mode). For example,
723 /(another)?(\1?)test/ matched against "hello world test". This bug was
724 introduced in release 8.13.
725
726 6. Add support for 16-bit character strings (a large amount of work involving
727 many changes and refactorings).
728
729 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
730 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
731 from a file.
732
733 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
734 rounding is not applied in this particular case).
735
736 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
737 if they appear, or are escaped, in patterns.
738
739 10. Get rid of a number of -Wunused-but-set-variable warnings.
740
741 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
742 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
743 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
744 also returns the mark "x". This bug applied to capturing parentheses,
745 non-capturing parentheses, and atomic parentheses. It also applied to some
746 assertions.
747
748 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
749 information out of configure.ac instead of relying on pcre.h.generic, which
750 is not stored in the repository.
751
752 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
753 -lreadline.
754
755 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
756
757 15. Applied Graycode's patch to put the top-level frame on the stack rather
758 than the heap when not using the stack for recursion. This gives a
759 performance improvement in many cases when recursion is not deep.
760
761 16. Experimental code added to "pcretest -C" to output the stack frame size.
762
763
764 Version 8.21 12-Dec-2011
765 ------------------------
766
767 1. Updating the JIT compiler.
768
769 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
770 are added as well.
771
772 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
773 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
774 calling _pcre_jit_exec. Some extra comments are added.
775
776 4. (*MARK) settings inside atomic groups that do not contain any capturing
777 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
778 was introduced by change 18 for 8.20.
779
780 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
781 ECMA-262 standard.
782
783 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
784 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
785 This bug was probably introduced by change 9 of 8.13.
786
787 7. While fixing 6 above, I noticed that a number of other items were being
788 incorrectly rejected as "not fixed length". This arose partly because newer
789 opcodes had not been added to the fixed-length checking code. I have (a)
790 corrected the bug and added tests for these items, and (b) arranged for an
791 error to occur if an unknown opcode is encountered while checking for fixed
792 length instead of just assuming "not fixed length". The items that were
793 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
794 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
795 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
796
797 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
798 being incorrectly compiled and would have given unpredicatble results.
799
800 9. A possessively repeated subpattern with minimum repeat count greater than
801 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
802 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
803 the first (A) could occur when it should not.
804
805 10. Add a cast and remove a redundant test from the code.
806
807 11. JIT should use pcre_malloc/pcre_free for allocation.
808
809 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
810 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
811 is anything other than /usr, -L is still shown).
812
813 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
814
815 14. Perl does not support \N without a following name in a [] class; PCRE now
816 also gives an error.
817
818 15. If a forward reference was repeated with an upper limit of around 2000,
819 it caused the error "internal error: overran compiling workspace". The
820 maximum number of forward references (including repeats) was limited by the
821 internal workspace, and dependent on the LINK_SIZE. The code has been
822 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
823 the default depends on LINK_SIZE. There is a new upper limit (for safety)
824 of around 200,000 forward references. While doing this, I also speeded up
825 the filling in of repeated forward references.
826
827 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
828 incorrectly expecting the subject to contain another "a" after the start.
829
830 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
831 in the match, the SKIP should be ignored. This was not happening; instead
832 the SKIP was being treated as NOMATCH. For patterns such as
833 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
834 tested.
835
836 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
837 now much more compatible with Perl, in particular in cases where the result
838 is a non-match for a non-anchored pattern. For example, if
839 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
840 "m", where previously it did not return a name. A side effect of this
841 change is that for partial matches, the last encountered mark name is
842 returned, as for non matches. A number of tests that were previously not
843 Perl-compatible have been moved into the Perl-compatible test files. The
844 refactoring has had the pleasing side effect of removing one argument from
845 the match() function, thus reducing its stack requirements.
846
847 19. If the /S+ option was used in pcretest to study a pattern using JIT,
848 subsequent uses of /S (without +) incorrectly behaved like /S+.
849
850 21. Retrieve executable code size support for the JIT compiler and fixing
851 some warnings.
852
853 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
854 not work when the shorter character appeared right at the end of the
855 subject string.
856
857 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
858 systems.
859
860 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
861 output it when the /M option is used in pcretest.
862
863 25. The CheckMan script was not being included in the distribution. Also, added
864 an explicit "perl" to run Perl scripts from the PrepareRelease script
865 because this is reportedly needed in Windows.
866
867 26. If study data was being save in a file and studying had not found a set of
868 "starts with" bytes for the pattern, the data written to the file (though
869 never used) was taken from uninitialized memory and so caused valgrind to
870 complain.
871
872 27. Updated RunTest.bat as provided by Sheri Pierce.
873
874 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
875
876 29. Computation of memory usage for the table of capturing group names was
877 giving an unnecessarily large value.
878
879
880 Version 8.20 21-Oct-2011
881 ------------------------
882
883 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
884 a POSIX class. After further experiments with Perl, which convinced me that
885 Perl has bugs and confusions, a closing square bracket is no longer allowed
886 in a POSIX name. This bug also affected patterns with classes that started
887 with full stops.
888
889 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
890 captured substring, but while checking the failing first alternative,
891 substring 1 is temporarily captured. If the output vector supplied to
892 pcre_exec() was not big enough for this capture, the yield of the function
893 was still zero ("insufficient space for captured substrings"). This cannot
894 be totally fixed without adding another stack variable, which seems a lot
895 of expense for a edge case. However, I have improved the situation in cases
896 such as /(a)(b)x|abc/ matched against "abc", where the return code
897 indicates that fewer than the maximum number of slots in the ovector have
898 been set.
899
900 3. Related to (2) above: when there are more back references in a pattern than
901 slots in the output vector, pcre_exec() uses temporary memory during
902 matching, and copies in the captures as far as possible afterwards. It was
903 using the entire output vector, but this conflicts with the specification
904 that only 2/3 is used for passing back captured substrings. Now it uses
905 only the first 2/3, for compatibility. This is, of course, another edge
906 case.
907
908 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
909 main code base, and can be used by building with --enable-jit. When this is
910 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
911 runtime --no-jit option is given.
912
913 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
914 ovector, the return from the function was zero, implying that there were
915 other matches that did not fit. The correct "exactly full" value is now
916 returned.
917
918 6. If a subpattern that was called recursively or as a subroutine contained
919 (*PRUNE) or any other control that caused it to give a non-standard return,
920 invalid errors such as "Error -26 (nested recursion at the same subject
921 position)" or even infinite loops could occur.
922
923 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
924 computing the minimum length on reaching *ACCEPT, and so ended up with the
925 wrong value of 1 rather than 0. Further investigation indicates that
926 computing a minimum subject length in the presence of *ACCEPT is difficult
927 (think back references, subroutine calls), and so I have changed the code
928 so that no minimum is registered for a pattern that contains *ACCEPT.
929
930 8. If (*THEN) was present in the first (true) branch of a conditional group,
931 it was not handled as intended. [But see 16 below.]
932
933 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
934 Sheri Pierce.
935
936 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
937 the first byte in a match must be "a".
938
939 11. Change 17 for 8.13 increased the recursion depth for patterns like
940 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
941 pattern contains any instances of (*THEN). If it does not, the old
942 optimizations are restored. It would be nice to do this on a per-group
943 basis, but at the moment that is not feasible.
944
945 12. In some environments, the output of pcretest -C is CRLF terminated. This
946 broke RunTest's code that checks for the link size. A single white space
947 character after the value is now allowed for.
948
949 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
950 For "fr", it uses the Windows-specific input and output files.
951
952 14. If (*THEN) appeared in a group that was called recursively or as a
953 subroutine, it did not work as intended. [But see next item.]
954
955 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
956 pattern fragments (but not containing any | characters). If A and B are
957 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
958 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
959 D. In other words, Perl considers parentheses that do not contain any |
960 characters to be part of a surrounding alternative, whereas PCRE was
961 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
962 differently. PCRE now behaves in the same way as Perl, except in the case
963 of subroutine/recursion calls such as (?1) which have in any case always
964 been different (but PCRE had them first :-).
965
966 16. Related to 15 above: Perl does not treat the | in a conditional group as
967 creating alternatives. Such a group is treated in the same way as an
968 ordinary group without any | characters when processing (*THEN). PCRE has
969 been changed to match Perl's behaviour.
970
971 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
972 RunGrepTest script failed.
973
974 18. Change 22 for version 13 caused atomic groups to use more stack. This is
975 inevitable for groups that contain captures, but it can lead to a lot of
976 stack use in large patterns. The old behaviour has been restored for atomic
977 groups that do not contain any capturing parentheses.
978
979 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
980 suppress the check for a minimum subject length at run time. (If it was
981 given to pcre_exec() or pcre_dfa_exec() it did work.)
982
983 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
984 fail to work when decoding hex characters in data strings in EBCDIC
985 environments.
986
987 21. It appears that in at least one Mac OS environment, the isxdigit() function
988 is implemented as a macro that evaluates to its argument more than once,
989 contravening the C 90 Standard (I haven't checked a later standard). There
990 was an instance in pcretest which caused it to go wrong when processing
991 \x{...} escapes in subject strings. The has been rewritten to avoid using
992 things like p++ in the argument of isxdigit().
993
994
995 Version 8.13 16-Aug-2011
996 ------------------------
997
998 1. The Unicode data tables have been updated to Unicode 6.0.0.
999
1000 2. Two minor typos in pcre_internal.h have been fixed.
1001
1002 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1003 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1004 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1005
1006 4. There were a number of related bugs in the code for matching backrefences
1007 caselessly in UTF-8 mode when codes for the characters concerned were
1008 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1009 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1010 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1011 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1012 2-byte code at the end of the subject (it thought there wasn't enough data
1013 left).
1014
1015 5. Comprehensive information about what went wrong is now returned by
1016 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1017 as the output vector has at least 2 elements. The offset of the start of
1018 the failing character and a reason code are placed in the vector.
1019
1020 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1021 now returned is for the first byte of the failing character, instead of the
1022 last byte inspected. This is an incompatible change, but I hope it is small
1023 enough not to be a problem. It makes the returned offset consistent with
1024 pcre_exec() and pcre_dfa_exec().
1025
1026 7. pcretest now gives a text phrase as well as the error number when
1027 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1028 failure, the offset and reason code are output.
1029
1030 8. When \R was used with a maximizing quantifier it failed to skip backwards
1031 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1032 back over a single character (\n). This seems wrong (because it treated the
1033 two characters as a single entity when going forwards), conflicts with the
1034 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1035 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1036 has been changed.
1037
1038 9. Some internal refactoring has changed the processing so that the handling
1039 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1040 time (the PCRE_DOTALL option was changed this way some time ago: version
1041 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1042 which was always a bit of a fudge. It also means that there is one less
1043 argument for the match() function, which reduces its stack requirements
1044 slightly. This change also fixes an incompatibility with Perl: the pattern
1045 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1046
1047 10. More internal refactoring has drastically reduced the number of recursive
1048 calls to match() for possessively repeated groups such as (abc)++ when
1049 using pcre_exec().
1050
1051 11. While implementing 10, a number of bugs in the handling of groups were
1052 discovered and fixed:
1053
1054 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1055 (a|)*(?1) gave a compile-time internal error.
1056 ((a|)+)+ did not notice that the outer group could match an empty string.
1057 (^a|^)+ was not marked as anchored.
1058 (.*a|.*)+ was not marked as matching at start or after a newline.
1059
1060 12. Yet more internal refactoring has removed another argument from the match()
1061 function. Special calls to this function are now indicated by setting a
1062 value in a variable in the "match data" data block.
1063
1064 13. Be more explicit in pcre_study() instead of relying on "default" for
1065 opcodes that mean there is no starting character; this means that when new
1066 ones are added and accidentally left out of pcre_study(), testing should
1067 pick them up.
1068
1069 14. The -s option of pcretest has been documented for ages as being an old
1070 synonym of -m (show memory usage). I have changed it to mean "force study
1071 for every regex", that is, assume /S for every regex. This is similar to -i
1072 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1073 using it. It makes it easier to run collections of tests with and without
1074 study enabled, and thereby test pcre_study() more easily. All the standard
1075 tests are now run with and without -s (but some patterns can be marked as
1076 "never study" - see 20 below).
1077
1078 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1079 restoration of the capturing data to the outer values was not happening
1080 correctly.
1081
1082 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1083 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1084 pattern had matched an empty string, and so incorrectly returned a no
1085 match.
1086
1087 17. There was optimizing code for the last branch of non-capturing parentheses,
1088 and also for the obeyed branch of a conditional subexpression, which used
1089 tail recursion to cut down on stack usage. Unfortunately, now that there is
1090 the possibility of (*THEN) occurring in these branches, tail recursion is
1091 no longer possible because the return has to be checked for (*THEN). These
1092 two optimizations have therefore been removed. [But see 8.20/11 above.]
1093
1094 18. If a pattern containing \R was studied, it was assumed that \R always
1095 matched two bytes, thus causing the minimum subject length to be
1096 incorrectly computed because \R can also match just one byte.
1097
1098 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1099 was incorrectly computed.
1100
1101 20. If /S is present twice on a test pattern in pcretest input, it now
1102 *disables* studying, thereby overriding the use of -s on the command line
1103 (see 14 above). This is necessary for one or two tests to keep the output
1104 identical in both cases.
1105
1106 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1107 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1108
1109 22. When an atomic group that contained a capturing parenthesis was
1110 successfully matched, but the branch in which it appeared failed, the
1111 capturing was not being forgotten if a higher numbered group was later
1112 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1113 group 1 to "a", when in fact it should be unset. This applied to multi-
1114 branched capturing and non-capturing groups, repeated or not, and also to
1115 positive assertions (capturing in negative assertions does not happen
1116 in PCRE) and also to nested atomic groups.
1117
1118 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1119 subject after a captured substring, to make it easier to tell which of a
1120 number of identical substrings has been captured.
1121
1122 24. The way atomic groups are processed by pcre_exec() has been changed so that
1123 if they are repeated, backtracking one repetition now resets captured
1124 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1125 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1126 "aaa". Previously, it would have been "a". As part of this code
1127 refactoring, the way recursive calls are handled has also been changed.
1128
1129 25. If an assertion condition captured any substrings, they were not passed
1130 back unless some other capturing happened later. For example, if
1131 (?(?=(a))a) was matched against "a", no capturing was returned.
1132
1133 26. When studying a pattern that contained subroutine calls or assertions,
1134 the code for finding the minimum length of a possible match was handling
1135 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1136 group 1 called group 2 while simultaneously a separate group 2 called group
1137 1). A stack overflow occurred in this case. I have fixed this by limiting
1138 the recursion depth to 10.
1139
1140 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1141 Fortmann. This supports explicit test numbers on the command line, and has
1142 argument validation and error reporting.
1143
1144 28. An instance of \X with an unlimited repeat could fail if at any point the
1145 first character it looked at was a mark character.
1146
1147 29. Some minor code refactoring concerning Unicode properties and scripts
1148 should reduce the stack requirement of match() slightly.
1149
1150 30. Added the '=' option to pcretest to check the setting of unused capturing
1151 slots at the end of the pattern, which are documented as being -1, but are
1152 not included in the return count.
1153
1154 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1155 compiled something random. Now it gives a compile-time error (as does
1156 Perl).
1157
1158 32. A *MARK encountered during the processing of a positive assertion is now
1159 recorded and passed back (compatible with Perl).
1160
1161 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1162 had alternative anchored branches, the search for a second match in a line
1163 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1164 matched the line "0102" twice. The same bug affected patterns that started
1165 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1166 twice.
1167
1168 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1169 does, and because of capturing effects, quantifying parenthesized
1170 assertions may at times be useful. Quantifiers are now allowed for
1171 parenthesized assertions.
1172
1173 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1174
1175 36. \g was being checked for fancy things in a character class, when it should
1176 just be a literal "g".
1177
1178 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1179 appearance of a nested POSIX class supersedes an apparent external class.
1180 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1181 unescaped square brackets may also appear as part of class names. For
1182 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1183 more like Perl. (But see 8.20/1 above.)
1184
1185 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1186 was because it thought it was \N{name}, which is not supported).
1187
1188 39. Add minix to OS list not supporting the -S option in pcretest.
1189
1190 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1191 cannot analyze patterns in sufficient detail to catch mutual recursions
1192 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1193 subgroup is called recursively as a subpattern for a second time at the
1194 same position in the subject string. In previous releases this might have
1195 been caught by the recursion limit, or it might have run out of stack.
1196
1197 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1198 happen only once. PCRE was, however incorrectly giving a compile time error
1199 "recursive call could loop indefinitely" because it cannot analyze the
1200 pattern in sufficient detail. The compile time test no longer happens when
1201 PCRE is compiling a conditional subpattern, but actual runaway loops are
1202 now caught at runtime (see 40 above).
1203
1204 42. It seems that Perl allows any characters other than a closing parenthesis
1205 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1206 has been changed to be the same.
1207
1208 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1209 as not to get warnings when autogen.sh is called. Also changed
1210 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1211
1212 44. To help people who use pcregrep to scan files containing exceedingly long
1213 lines, the following changes have been made:
1214
1215 (a) The default value of the buffer size parameter has been increased from
1216 8K to 20K. (The actual buffer used is three times this size.)
1217
1218 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1219 PCRE is built.
1220
1221 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1222 to be set at run time.
1223
1224 (d) Numerical values in pcregrep options can be followed by K or M, for
1225 example --buffer-size=50K.
1226
1227 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1228 given and the return code is set to 2.
1229
1230 45. Add a pointer to the latest mark to the callout data block.
1231
1232 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1233 partial match of an empty string instead of no match. This was specific to
1234 the use of ".".
1235
1236 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1237 complete match instead of a partial match. This bug was dependent on both
1238 the PCRE_UTF8 and PCRE_DOTALL options being set.
1239
1240 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1241 starting byte set, because \b was not being ignored.
1242
1243
1244 Version 8.12 15-Jan-2011
1245 ------------------------
1246
1247 1. Fixed some typos in the markup of the man pages, and wrote a script that
1248 checks for such things as part of the documentation building process.
1249
1250 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1251 --match-limit and --recursion-limit options (added for 8.11). In
1252 particular, this made one of the standard tests fail. (The integer value
1253 went into the wrong half of a long int.)
1254
1255 3. If the --colour option was given to pcregrep with -v (invert match), it
1256 did strange things, either producing crazy output, or crashing. It should,
1257 of course, ignore a request for colour when reporting lines that do not
1258 match.
1259
1260 4. Another pcregrep bug caused similar problems if --colour was specified with
1261 -M (multiline) and the pattern match finished with a line ending.
1262
1263 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1264 matched in multiline mode, the following line was shown as part of the
1265 match. This seems wrong, so I have changed it.
1266
1267 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1268 the check for further matches in the same line (so they could be coloured)
1269 to overrun the end of the current line. If another match was found, it was
1270 incorrectly shown (and then shown again when found in the next line).
1271
1272 7. If pcregrep was compiled under Windows, there was a reference to the
1273 function pcregrep_exit() before it was defined. I am assuming this was
1274 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1275 reported by a user. I've moved the definition above the reference.
1276
1277
1278 Version 8.11 10-Dec-2010
1279 ------------------------
1280
1281 1. (*THEN) was not working properly if there were untried alternatives prior
1282 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1283 backtracked to try for "b" instead of moving to the next alternative branch
1284 at the same level (in this case, to look for "c"). The Perl documentation
1285 is clear that when (*THEN) is backtracked onto, it goes to the "next
1286 alternative in the innermost enclosing group".
1287
1288 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1289 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1290 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1291 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1292 (*THEN).
1293
1294 3. If \s appeared in a character class, it removed the VT character from
1295 the class, even if it had been included by some previous item, for example
1296 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1297 of \s, but is part of the POSIX "space" class.)
1298
1299 4. A partial match never returns an empty string (because you can always
1300 match an empty string at the end of the subject); however the checking for
1301 an empty string was starting at the "start of match" point. This has been
1302 changed to the "earliest inspected character" point, because the returned
1303 data for a partial match starts at this character. This means that, for
1304 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1305 (previously it gave "no match").
1306
1307 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1308 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1309 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1310 has an implication that the given string is incomplete (because a partial
1311 match is preferred over a full match). For this reason, these items now
1312 give a partial match in this situation. [Aside: previously, the one case
1313 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1314 match rather than a full match, which was wrong by the old rules, but is
1315 now correct.]
1316
1317 6. There was a bug in the handling of #-introduced comments, recognized when
1318 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1319 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1320 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1321 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1322 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1323 places in pcre_compile().
1324
1325 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1326 comments when looking ahead for named forward references to subpatterns,
1327 the only newline sequence it recognized was NL. It now handles newlines
1328 according to the set newline convention.
1329
1330 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1331 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1332 cater for a lack of strerror(). These oversights have been fixed.
1333
1334 9. Added --match-limit and --recursion-limit to pcregrep.
1335
1336 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1337
1338 11. When the -o option was used, pcregrep was setting a return code of 1, even
1339 when matches were found, and --line-buffered was not being honoured.
1340
1341 12. Added an optional parentheses number to the -o and --only-matching options
1342 of pcregrep.
1343
1344 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1345 can match an empty string. The code to do it in pcretest and pcredemo
1346 needed fixing:
1347
1348 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1349 only one byte after an empty string match just before CRLF (this case
1350 just got forgotten; "any" and "anycrlf" were OK).
1351
1352 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1353 mode when an empty string match preceded an ASCII character followed by
1354 a non-ASCII character. (The code for advancing by one character rather
1355 than one byte was nonsense.)
1356
1357 (c) The pcredemo.c sample program did not have any code at all to handle
1358 the cases when CRLF is a valid newline sequence.
1359
1360 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1361 as a starting offset was within the subject string. There is now a new
1362 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1363 negative or greater than the length of the string. In order to test this,
1364 pcretest is extended to allow the setting of negative starting offsets.
1365
1366 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1367 starting offset points to the beginning of a UTF-8 character was
1368 unnecessarily clumsy. I tidied it up.
1369
1370 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1371 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1372
1373 17. Nobody had reported that the --include_dir option, which was added in
1374 release 7.7 should have been called --include-dir (hyphen, not underscore)
1375 for compatibility with GNU grep. I have changed it to --include-dir, but
1376 left --include_dir as an undocumented synonym, and the same for
1377 --exclude-dir, though that is not available in GNU grep, at least as of
1378 release 2.5.4.
1379
1380 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1381 characters from a string of bytes) have been redefined so as not to use
1382 loops, in order to improve performance in some environments. At the same
1383 time, I abstracted some of the common code into auxiliary macros to save
1384 repetition (this should not affect the compiled code).
1385
1386 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1387 compile-time error is now given if \c is not followed by an ASCII
1388 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1389 different, and any byte value is allowed.)
1390
1391 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1392 START_OPTIMIZE option, which is now allowed at compile time - but just
1393 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1394 to pcregrep and other applications that have no direct access to PCRE
1395 options. The new /Y option in pcretest sets this option when calling
1396 pcre_compile().
1397
1398 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1399 back references. Groups containing recursive back references were forced to
1400 be atomic by that change, but in the case of named groups, the amount of
1401 memory required was incorrectly computed, leading to "Failed: internal
1402 error: code overflow". This has been fixed.
1403
1404 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1405 pcretest.c, to avoid build problems in some Borland environments.
1406
1407
1408 Version 8.10 25-Jun-2010
1409 ------------------------
1410
1411 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1412 THEN.
1413
1414 2. (*ACCEPT) was not working when inside an atomic group.
1415
1416 3. Inside a character class, \B is treated as a literal by default, but
1417 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1418 causes the error). The code is unchanged, but I tidied the documentation.
1419
1420 4. Inside a character class, PCRE always treated \R and \X as literals,
1421 whereas Perl faults them if its -w option is set. I have changed PCRE so
1422 that it faults them when PCRE_EXTRA is set.
1423
1424 5. Added support for \N, which always matches any character other than
1425 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1426
1427 6. When compiling pcregrep with newer versions of gcc which may have
1428 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1429 declared with attribute warn_unused_result" were given. Just casting the
1430 result to (void) does not stop the warnings; a more elaborate fudge is
1431 needed. I've used a macro to implement this.
1432
1433 7. Minor change to pcretest.c to avoid a compiler warning.
1434
1435 8. Added four artifical Unicode properties to help with an option to make
1436 \s etc use properties (see next item). The new properties are: Xan
1437 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1438
1439 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1440 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1441 this option. Modified pcretest to add /W to test this facility. Added
1442 REG_UCP to make it available via the POSIX interface.
1443
1444 10. Added --line-buffered to pcregrep.
1445
1446 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1447 studied, and the match started with a letter with a code point greater than
1448 127 whose first byte was different to the first byte of the other case of
1449 the letter, the other case of this starting letter was not recognized
1450 (#976).
1451
1452 12. If a pattern that was studied started with a repeated Unicode property
1453 test, for example, \p{Nd}+, there was the theoretical possibility of
1454 setting up an incorrect bitmap of starting bytes, but fortunately it could
1455 not have actually happened in practice until change 8 above was made (it
1456 added property types that matched character-matching opcodes).
1457
1458 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1459 possible starting bytes for non-anchored patterns.
1460
1461 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1462 \R, and also a number of cases that involve Unicode properties, both
1463 explicit and implicit when PCRE_UCP is set.
1464
1465 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1466 input, it could crash or give wrong results if characters with values
1467 greater than 0xc0 were present in the subject string. (Detail: it assumed
1468 UTF-8 input when processing these items.)
1469
1470 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1471 size_t is 64-bit (#991).
1472
1473 17. Added a check for running out of memory when PCRE is compiled with
1474 --disable-stack-for-recursion (#990).
1475
1476 18. If the last data line in a file for pcretest does not have a newline on
1477 the end, a newline was missing in the output.
1478
1479 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1480 less than 128) in its various bitmaps. However, there is a facility for
1481 generating tables according to the current locale when PCRE is compiled. It
1482 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1483 characters, are recognized by isspace() and therefore were getting set in
1484 these tables, and indeed these tables seem to approximate to ISO 8859. This
1485 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1486 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1487 which of course cannot start UTF-8 characters. I have changed the code so
1488 that only real ASCII characters (less than 128) and the correct starting
1489 bytes for UTF-8 encodings are set for characters greater than 127 when in
1490 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1491 altogether.)
1492
1493 20. Added the /T option to pcretest so as to be able to run tests with non-
1494 standard character tables, thus making it possible to include the tests
1495 used for 19 above in the standard set of tests.
1496
1497 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1498 reference to a subpattern the other side of a comment that contains an
1499 opening parenthesis caused either an internal compiling error, or a
1500 reference to the wrong subpattern.
1501
1502
1503 Version 8.02 19-Mar-2010
1504 ------------------------
1505
1506 1. The Unicode data tables have been updated to Unicode 5.2.0.
1507
1508 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1509 configured.
1510
1511 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1512 original author of that file, following a query about its status.
1513
1514 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1515 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1516
1517 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1518 quantifier applied to a forward-referencing subroutine call, could compile
1519 incorrect code or give the error "internal error: previously-checked
1520 referenced subpattern not found".
1521
1522 6. Both MS Visual Studio and Symbian OS have problems with initializing
1523 variables to point to external functions. For these systems, therefore,
1524 pcre_malloc etc. are now initialized to local functions that call the
1525 relevant global functions.
1526
1527 7. There were two entries missing in the vectors called coptable and poptable
1528 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1529 I've fixed the data, and added a kludgy way of testing at compile time that
1530 the lengths are correct (equal to the number of opcodes).
1531
1532 8. Following on from 7, I added a similar kludge to check the length of the
1533 eint vector in pcreposix.c.
1534
1535 9. Error texts for pcre_compile() are held as one long string to avoid too
1536 much relocation at load time. To find a text, the string is searched,
1537 counting zeros. There was no check for running off the end of the string,
1538 which could happen if a new error number was added without updating the
1539 string.
1540
1541 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1542
1543 11. \K was not working if it appeared in an atomic group or in a group that
1544 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1545 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1546 the assertion is positive, but not if it is negative.
1547
1548 12. Change 11 fortuitously reduced the size of the stack frame used in the
1549 "match()" function of pcre_exec.c by one pointer. Forthcoming
1550 implementation of support for (*MARK) will need an extra pointer on the
1551 stack; I have reserved it now, so that the stack frame size does not
1552 decrease.
1553
1554 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1555 item in branch that calls a recursion is a subroutine call - as in the
1556 second branch in the above example - was incorrectly given the compile-
1557 time error "recursive call could loop indefinitely" because pcre_compile()
1558 was not correctly checking the subroutine for matching a non-empty string.
1559
1560 14. The checks for overrunning compiling workspace could trigger after an
1561 overrun had occurred. This is a "should never occur" error, but it can be
1562 triggered by pathological patterns such as hundreds of nested parentheses.
1563 The checks now trigger 100 bytes before the end of the workspace.
1564
1565 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1566
1567
1568 Version 8.01 19-Jan-2010
1569 ------------------------
1570
1571 1. If a pattern contained a conditional subpattern with only one branch (in
1572 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1573 computed the wrong minimum data length (which is of course zero for such
1574 subpatterns). This could cause incorrect "no match" results.
1575
1576 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1577 the pattern is reset in the first branch, pcre_compile() failed with
1578 "internal error: code overflow at offset...". This happened only when
1579 the reset was to the original external option setting. (An optimization
1580 abstracts leading options settings into an external setting, which was the
1581 cause of this.)
1582
1583 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1584 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1585 assertion pattern did not match (meaning that the assertion was true), it
1586 was incorrectly treated as false if the SKIP had been reached during the
1587 matching. This also applied to assertions used as conditions.
1588
1589 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1590 assertion subpattern, including such a pattern used as a condition,
1591 unpredictable results occurred, instead of the error return
1592 PCRE_ERROR_DFA_UITEM.
1593
1594 5. The C++ GlobalReplace function was not working like Perl for the special
1595 situation when an empty string is matched. It now does the fancy magic
1596 stuff that is necessary.
1597
1598 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1599 removed. (These were left over from very, very early versions of PCRE.)
1600
1601 7. Some cosmetic changes to the code to make life easier when compiling it
1602 as part of something else:
1603
1604 (a) Change DEBUG to PCRE_DEBUG.
1605
1606 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1607 called "current" as "current_branch", to prevent a collision with the
1608 Linux macro when compiled as a kernel module.
1609
1610 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1611 prevent a collision with the Linux macro when compiled as a kernel
1612 module.
1613
1614 8. In pcre_compile() there are some checks for integer overflows that used to
1615 cast potentially large values to (double). This has been changed to that
1616 when building, a check for int64_t is made, and if it is found, it is used
1617 instead, thus avoiding the use of floating point arithmetic. (There is no
1618 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1619 double.
1620
1621 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1622 2005 (difference between two addresses compared to an unsigned value).
1623
1624 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1625 custom one, because of the following reported problem in Windows:
1626
1627 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1628 under Win32.
1629 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1630 therefore missing the function definition.
1631 - The compiler thus generates a "C" signature for the test function.
1632 - The linker fails to find the "C" function.
1633 - PCRE fails to configure if asked to do so against libbz2.
1634
1635 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1636 messages were output:
1637
1638 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1639 rerunning libtoolize, to keep the correct libtool macros in-tree.
1640 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1641
1642 I have done both of these things.
1643
1644 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1645 most of the time, it *can* run out if it is given a pattern that contains a
1646 runaway infinite recursion. I updated the discussion in the pcrestack man
1647 page.
1648
1649 13. Now that we have gone to the x.xx style of version numbers, the minor
1650 version may start with zero. Using 08 or 09 is a bad idea because users
1651 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1652 interpreted as invalid octal numbers. I've updated the previous comment in
1653 configure.ac, and also added a check that gives an error if 08 or 09 are
1654 used.
1655
1656 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1657 causing partial matching to fail when the end of the subject matched \W
1658 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1659
1660 15. There were some discrepancies between the declarations in pcre_internal.h
1661 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1662 their definitions. The declarations used "const uschar *" and the
1663 definitions used USPTR. Even though USPTR is normally defined as "const
1664 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1665 reported that: "This difference in casting confuses some C++ compilers, for
1666 example, SunCC recognizes above declarations as different functions and
1667 generates broken code for hbpcre." I have changed the declarations to use
1668 USPTR.
1669
1670 16. GNU libtool is named differently on some systems. The autogen.sh script now
1671 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1672 (FreeBSD).
1673
1674 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1675 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1676 comment: "Figure out how to create a longlong from a string: strtoll and
1677 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1678 instance, but it only takes 2 args instead of 3!"
1679
1680 18. A subtle bug concerned with back references has been fixed by a change of
1681 specification, with a corresponding code fix. A pattern such as
1682 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1683 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1684 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1685 same bug. Such groups have to be quantified to be useful, or contained
1686 inside another quantified group. (If there's no repetition, the reference
1687 can never match.) The problem arises because, having left the group and
1688 moved on to the rest of the pattern, a later failure that backtracks into
1689 the group uses the captured value from the final iteration of the group
1690 rather than the correct earlier one. I have fixed this in PCRE by forcing
1691 any group that contains a reference to itself to be an atomic group; that
1692 is, there cannot be any backtracking into it once it has completed. This is
1693 similar to recursive and subroutine calls.
1694
1695
1696 Version 8.00 19-Oct-09
1697 ----------------------
1698
1699 1. The table for translating pcre_compile() error codes into POSIX error codes
1700 was out-of-date, and there was no check on the pcre_compile() error code
1701 being within the table. This could lead to an OK return being given in
1702 error.
1703
1704 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1705 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1706 in a Windows environment.
1707
1708 3. The pcregrep --count option prints the count for each file even when it is
1709 zero, as does GNU grep. However, pcregrep was also printing all files when
1710 --files-with-matches was added. Now, when both options are given, it prints
1711 counts only for those files that have at least one match. (GNU grep just
1712 prints the file name in this circumstance, but including the count seems
1713 more useful - otherwise, why use --count?) Also ensured that the
1714 combination -clh just lists non-zero counts, with no names.
1715
1716 4. The long form of the pcregrep -F option was incorrectly implemented as
1717 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1718 but it seems right to fix it, and I didn't think it was worth preserving
1719 the old behaviour.
1720
1721 5. The command line items --regex=pattern and --regexp=pattern were not
1722 recognized by pcregrep, which required --regex pattern or --regexp pattern
1723 (with a space rather than an '='). The man page documented the '=' forms,
1724 which are compatible with GNU grep; these now work.
1725
1726 6. No libpcreposix.pc file was created for pkg-config; there was just
1727 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1728
1729 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1730 when UCP support is not needed, by modifying the Python script that
1731 generates it from Unicode data files. This should not matter if the module
1732 is correctly used as a library, but I received one complaint about 50K of
1733 unwanted data. My guess is that the person linked everything into his
1734 program rather than using a library. Anyway, it does no harm.
1735
1736 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1737 was a minimum greater than 1 for a wide character in a possessive
1738 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1739 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1740 character. Chaos in the form of incorrect output or a compiling loop could
1741 result.
1742
1743 9. The restrictions on what a pattern can contain when partial matching is
1744 requested for pcre_exec() have been removed. All patterns can now be
1745 partially matched by this function. In addition, if there are at least two
1746 slots in the offset vector, the offset of the earliest inspected character
1747 for the match and the offset of the end of the subject are set in them when
1748 PCRE_ERROR_PARTIAL is returned.
1749
1750 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1751 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1752 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1753 and may be more useful for multi-segment matching.
1754
1755 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1756 used to be given if ever the end of the subject was reached; now it is
1757 given only if matching could not proceed because another character was
1758 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1759 string "Z", which now yields "no match" instead of "partial match". In the
1760 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1761 final character ended with (*FAIL).
1762
1763 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1764 if the pattern had a "must contain" character that was already found in the
1765 earlier partial match, unless partial matching was again requested. For
1766 example, with the pattern /dog.(body)?/, the "must contain" character is
1767 "g". If the first part-match was for the string "dog", restarting with
1768 "sbody" failed. This bug has been fixed.
1769
1770 13. The string returned by pcre_dfa_exec() after a partial match has been
1771 changed so that it starts at the first inspected character rather than the
1772 first character of the match. This makes a difference only if the pattern
1773 starts with a lookbehind assertion or \b or \B (\K is not supported by
1774 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1775 matching functions compatible, and I think it's the right thing to do.
1776
1777 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1778 so that the demonstration program is easily available in environments where
1779 PCRE has not been installed from source.
1780
1781 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1782 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1783 library.
1784
1785 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1786 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1787 is not the first non-POSIX option to be added. Clearly some people find
1788 these options useful.
1789
1790 17. If a caller to the POSIX matching function regexec() passes a non-zero
1791 value for nmatch with a NULL value for pmatch, the value of
1792 nmatch is forced to zero.
1793
1794 18. RunGrepTest did not have a test for the availability of the -u option of
1795 the diff command, as RunTest does. It now checks in the same way as
1796 RunTest, and also checks for the -b option.
1797
1798 19. If an odd number of negated classes containing just a single character
1799 interposed, within parentheses, between a forward reference to a named
1800 subpattern and the definition of the subpattern, compilation crashed with
1801 an internal error, complaining that it could not find the referenced
1802 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1803 [The bug was that it was starting one character too far in when skipping
1804 over the character class, thus treating the ] as data rather than
1805 terminating the class. This meant it could skip too much.]
1806
1807 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1808 /g option in pcretest when the pattern contains \K, which makes it possible
1809 to have an empty string match not at the start, even when the pattern is
1810 anchored. Updated pcretest and pcredemo to use this option.
1811
1812 21. If the maximum number of capturing subpatterns in a recursion was greater
1813 than the maximum at the outer level, the higher number was returned, but
1814 with unset values at the outer level. The correct (outer level) value is
1815 now given.
1816
1817 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1818 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1819 make it do so. The string so far is captured, making this feature
1820 compatible with Perl.
1821
1822 23. The tests have been re-organized, adding tests 11 and 12, to make it
1823 possible to check the Perl 5.10 features against Perl 5.10.
1824
1825 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1826 pattern matches a fixed length string. PCRE did not allow this; now it
1827 does. Neither allows recursion.
1828
1829 25. I finally figured out how to implement a request to provide the minimum
1830 length of subject string that was needed in order to match a given pattern.
1831 (It was back references and recursion that I had previously got hung up
1832 on.) This code has now been added to pcre_study(); it finds a lower bound
1833 to the length of subject needed. It is not necessarily the greatest lower
1834 bound, but using it to avoid searching strings that are too short does give
1835 some useful speed-ups. The value is available to calling programs via
1836 pcre_fullinfo().
1837
1838 26. While implementing 25, I discovered to my embarrassment that pcretest had
1839 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1840 study optimizations had never been tested with that matching function.
1841 Oops. What is worse, even when it was passed study data, there was a bug in
1842 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1843 were also very few tests of studied patterns with pcre_dfa_exec().
1844
1845 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1846 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1847 on the other side of the coin, they are no longer allowed to have different
1848 names, because these cannot be distinguished in PCRE, and this has caused
1849 confusion. (This is a difference from Perl.)
1850
1851 28. When duplicate subpattern names are present (necessarily with different
1852 numbers, as required by 27 above), and a test is made by name in a
1853 conditional pattern, either for a subpattern having been matched, or for
1854 recursion in such a pattern, all the associated numbered subpatterns are
1855 tested, and the overall condition is true if the condition is true for any
1856 one of them. This is the way Perl works, and is also more like the way
1857 testing by number works.
1858
1859
1860 Version 7.9 11-Apr-09
1861 ---------------------
1862
1863 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1864 (pcretest), all targets were linked against these libraries. This included
1865 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1866 libraries. This caused unwanted dependencies to be created. This problem
1867 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1868 pcretest is linked with readline.
1869
1870 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1871 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1872 moved outside it again, because FALSE and TRUE are already defined in AIX,
1873 but BOOL is not.
1874
1875 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1876 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1877
1878 4. The pcregrep documentation said spaces were inserted as well as colons (or
1879 hyphens) following file names and line numbers when outputting matching
1880 lines. This is not true; no spaces are inserted. I have also clarified the
1881 wording for the --colour (or --color) option.
1882
1883 5. In pcregrep, when --colour was used with -o, the list of matching strings
1884 was not coloured; this is different to GNU grep, so I have changed it to be
1885 the same.
1886
1887 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1888 each matching line was coloured. Now it goes on to look for further matches
1889 of any of the test patterns, which is the same behaviour as GNU grep.
1890
1891 7. A pattern that could match an empty string could cause pcregrep to loop; it
1892 doesn't make sense to accept an empty string match in pcregrep, so I have
1893 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1894 seems to be how GNU grep behaves. [But see later change 40 for release
1895 8.33.]
1896
1897 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1898 start or after a newline", because the conditional assertion was not being
1899 correctly handled. The rule now is that both the assertion and what follows
1900 in the first alternative must satisfy the test.
1901
1902 9. If auto-callout was enabled in a pattern with a conditional group whose
1903 condition was an assertion, PCRE could crash during matching, both with
1904 pcre_exec() and pcre_dfa_exec().
1905
1906 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1907 used for matching.
1908
1909 11. Unicode property support in character classes was not working for
1910 characters (bytes) greater than 127 when not in UTF-8 mode.
1911
1912 12. Added the -M command line option to pcretest.
1913
1914 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1915
1916 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1917
1918 16. Added comments and documentation about mis-use of no_arg in the C++
1919 wrapper.
1920
1921 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1922 from Martin Jerabek that uses macro names for all relevant character and
1923 string constants.
1924
1925 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1926 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1927 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1928 these, but not everybody uses configure.
1929
1930 19. A conditional group that had only one branch was not being correctly
1931 recognized as an item that could match an empty string. This meant that an
1932 enclosing group might also not be so recognized, causing infinite looping
1933 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1934 with the subject "ab", where knowledge that the repeated group can match
1935 nothing is needed in order to break the loop.
1936
1937 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1938 exec(), but without supplying a callout function, matching went wrong.
1939
1940 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1941 leak if the size of the offset vector was greater than 30. When the vector
1942 is smaller, the saved offsets during recursion go onto a local stack
1943 vector, but for larger vectors malloc() is used. It was failing to free
1944 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1945 error, in fact).
1946
1947 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1948 heapframe that is used only when UTF-8 support is enabled. This caused no
1949 problem, but was untidy.
1950
1951 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1952 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1953 included within another project.
1954
1955 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1956 slightly modified by me:
1957
1958 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1959 not building pcregrep.
1960
1961 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1962 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1963
1964 25. Forward references, both numeric and by name, in patterns that made use of
1965 duplicate group numbers, could behave incorrectly or give incorrect errors,
1966 because when scanning forward to find the reference group, PCRE was not
1967 taking into account the duplicate group numbers. A pattern such as
1968 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1969
1970 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1971 the feature of a custom pointer more persuasive (as requested by a user).
1972
1973 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1974 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1975 pre-defined.
1976
1977 28. Added support for (*UTF8) at the start of a pattern.
1978
1979 29. Arrange for flags added by the "release type" setting in CMake to be shown
1980 in the configuration summary.
1981
1982
1983 Version 7.8 05-Sep-08
1984 ---------------------
1985
1986 1. Replaced UCP searching code with optimized version as implemented for Ad
1987 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1988 stage table and inline lookup instead of a function, giving speed ups of 2
1989 to 5 times on some simple patterns that I tested. Permission was given to
1990 distribute the MultiStage2.py script that generates the tables (it's not in
1991 the tarball, but is in the Subversion repository).
1992
1993 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1994 scripts.
1995
1996 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1997 a group with a zero qualifier. The result of the study could be incorrect,
1998 or the function might crash, depending on the pattern.
1999
2000 4. Caseless matching was not working for non-ASCII characters in back
2001 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2002 It now works when Unicode Property Support is available.
2003
2004 5. In pcretest, an escape such as \x{de} in the data was always generating
2005 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2006 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2007 truncation.
2008
2009 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2010
2011 7. Added two (int) casts to pcregrep when printing the difference of two
2012 pointers, in case they are 64-bit values.
2013
2014 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2015 test 2 if it fails.
2016
2017 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2018 and a #define of that name to empty if it is not externally set. This is to
2019 allow users of MSVC to set it if necessary.
2020
2021 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2022 the convenience functions in the pcre_get.c source file.
2023
2024 11. An option change at the start of a pattern that had top-level alternatives
2025 could cause overwriting and/or a crash. This command provoked a crash in
2026 some environments:
2027
2028 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2029
2030 This potential security problem was recorded as CVE-2008-2371.
2031
2032 12. For a pattern where the match had to start at the beginning or immediately
2033 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2034 pcre_dfa_exec() could read past the end of the passed subject if there was
2035 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2036 pcretest so that it places the subject at the end of its malloc-ed buffer.
2037
2038 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2039 exec() might read past the end of the data buffer in UTF-8 mode.
2040
2041 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2042 the data contained the byte 0x85 as part of a UTF-8 character within its
2043 first line. This applied both to normal and DFA matching.
2044
2045 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2046 /^[^d]*?$/8 failed to match "abc".
2047
2048 16. Added a missing copyright notice to pcrecpp_internal.h.
2049
2050 17. Make it more clear in the documentation that values returned from
2051 pcre_exec() in ovector are byte offsets, not character counts.
2052
2053 18. Tidied a few places to stop certain compilers from issuing warnings.
2054
2055 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2056 supplied by Stefan Weber. I made a further small update for 7.8 because
2057 there is a change of source arrangements: the pcre_searchfuncs.c module is
2058 replaced by pcre_ucd.c.
2059
2060
2061 Version 7.7 07-May-08
2062 ---------------------
2063
2064 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2065 a string to a long long, pretend we don't even have a long long." This is
2066 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2067
2068 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2069 pre-7.6 versions, which defined a global no_arg variable instead of putting
2070 it in the RE class. (See also #8 below.)
2071
2072 3. Remove a line of dead code, identified by coverity and reported by Nuno
2073 Lopes.
2074
2075 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2076
2077 (1) The include/exclude patterns were being applied to the whole pathnames
2078 of files, instead of just to the final components.
2079
2080 (2) If there was more than one level of directory, the subdirectories were
2081 skipped unless they satisfied the include/exclude conditions. This is
2082 inconsistent with GNU grep (and could even be seen as contrary to the
2083 pcregrep specification - which I improved to make it absolutely clear).
2084 The action now is always to scan all levels of directory, and just
2085 apply the include/exclude patterns to regular files.
2086
2087 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2088 --exclude_dir in the tests to avoid scanning .svn directories.
2089
2090 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2091 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2092 doesn't support NULs in patterns.
2093
2094 7. Added some missing "const"s to declarations of static tables in
2095 pcre_compile.c and pcre_dfa_exec.c.
2096
2097 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2098 caused by fix #2 above. (Subsequently also a second patch to fix the
2099 first patch. And a third patch - this was a messy problem.)
2100
2101 9. Applied Craig's patch to remove the use of push_back().
2102
2103 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2104 matching function regexec().
2105
2106 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2107 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2108 references. PCRE supports relative numbers with this syntax (I don't think
2109 Oniguruma does).
2110
2111 12. Previously, a group with a zero repeat such as (...){0} was completely
2112 omitted from the compiled regex. However, this means that if the group
2113 was called as a subroutine from elsewhere in the pattern, things went wrong
2114 (an internal error was given). Such groups are now left in the compiled
2115 pattern, with a new opcode that causes them to be skipped at execution
2116 time.
2117
2118 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2119 to the way PCRE behaves:
2120
2121 (a) A lone ] character is dis-allowed (Perl treats it as data).
2122
2123 (b) A back reference to an unmatched subpattern matches an empty string
2124 (Perl fails the current match path).
2125
2126 (c) A data ] in a character class must be notated as \] because if the
2127 first data character in a class is ], it defines an empty class. (In
2128 Perl it is not possible to have an empty class.) The empty class []
2129 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2130 The negative empty class [^] matches any one character, independently
2131 of the DOTALL setting.
2132
2133 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2134 non-existent subpattern following a character class starting with ']' and
2135 containing () gave an internal compiling error instead of "reference to
2136 non-existent subpattern". Fortunately, when the pattern did exist, the
2137 compiled code was correct. (When scanning forwards to check for the
2138 existence of the subpattern, it was treating the data ']' as terminating
2139 the class, so got the count wrong. When actually compiling, the reference
2140 was subsequently set up correctly.)
2141
2142 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2143 it was being rejected as not supported by pcre_dfa_exec(), even though
2144 other assertions are supported. I have made pcre_dfa_exec() support
2145 (*FAIL).
2146
2147 16. The implementation of 13c above involved the invention of a new opcode,
2148 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2149 cannot be changed at match time, I realized I could make a small
2150 improvement to matching performance by compiling OP_ALLANY instead of
2151 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2152 on the OP_ANY path.
2153
2154 17. Compiling pcretest on Windows with readline support failed without the
2155 following two fixes: (1) Make the unistd.h include conditional on
2156 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2157
2158 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2159 ncurses library to be included for pcretest when ReadLine support is
2160 requested, but also to allow for it to be overridden. This patch came from
2161 Daniel Bergström.
2162
2163 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2164 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2165 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2166 spotting this.
2167
2168
2169 Version 7.6 28-Jan-08
2170 ---------------------
2171
2172 1. A character class containing a very large number of characters with
2173 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2174 overflow.
2175
2176 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2177 HAVE_LONG_LONG is not defined.
2178
2179 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2180 bring them up to date and include new features. This patch includes:
2181
2182 - Fixed PH's badly added libz and libbz2 support.
2183 - Fixed a problem with static linking.
2184 - Added pcredemo. [But later removed - see 7 below.]
2185 - Fixed dftables problem and added an option.
2186 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2187 HAVE_LONG_LONG.
2188 - Added readline support for pcretest.
2189 - Added an listing of the option settings after cmake has run.
2190
2191 4. A user submitted a patch to Makefile that makes it easy to create
2192 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2193 Makefile.am that cause it to include this special target, without
2194 affecting anything else. Note that the same mingw target plus all
2195 the other distribution libraries and programs are now supported
2196 when configuring with CMake (see 6 below) instead of with
2197 Configure/Make.
2198
2199 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2200 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2201 exported in the Windows port". It has not yet been confirmed that the patch
2202 solves the problem, but it does no harm.
2203
2204 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2205 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2206 with CMake, and also correct the comment about stack recursion.
2207
2208 7. Remove the automatic building of pcredemo from the ./configure system and
2209 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2210 of a program that users should build themselves after PCRE is installed, so
2211 building it automatically is not really right. What is more, it gave
2212 trouble in some build environments.
2213
2214 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2215
2216
2217 Version 7.5 10-Jan-08
2218 ---------------------
2219
2220 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2221 values in parens when parsing an RE using the C++ wrapper."
2222
2223 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2224 Characters greater than 255 were excluded from the class instead of being
2225 included.
2226
2227 3. The same bug as (2) above applied to negated POSIX classes such as
2228 [:^space:].
2229
2230 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2231 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2232 I have changed it.
2233
2234 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2235 first named subpattern) and a construct such as (?&a) would reference the
2236 first named subpattern whose name started with "a" (in other words, the
2237 length check was missing). Both these problems are fixed. "Subpattern name
2238 expected" is now given for (?&) (a zero-length name), and this patch also
2239 makes it give the same error for \k'' (previously it complained that that
2240 was a reference to a non-existent subpattern).
2241
2242 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2243 this is right because (?- can be followed by option settings as well as by
2244 digits. I have, however, made the messages clearer.
2245
2246 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2247 than the number used in the conditional) now cause a compile-time error.
2248 This is actually not compatible with Perl, which accepts such patterns, but
2249 treats the conditional as always being FALSE (as PCRE used to), but it
2250 seems to me that giving a diagnostic is better.
2251
2252 8. Change "alphameric" to the more common word "alphanumeric" in comments
2253 and messages.
2254
2255 9. Fix two occurrences of "backslash" in comments that should have been
2256 "backspace".
2257
2258 10. Remove two redundant lines of code that can never be obeyed (their function
2259 was moved elsewhere).
2260
2261 11. The program that makes PCRE's Unicode character property table had a bug
2262 which caused it to generate incorrect table entries for sequences of
2263 characters that have the same character type, but are in different scripts.
2264 It amalgamated them into a single range, with the script of the first of
2265 them. In other words, some characters were in the wrong script. There were
2266 thirteen such cases, affecting characters in the following ranges:
2267
2268 U+002b0 - U+002c1
2269 U+0060c - U+0060d
2270 U+0061e - U+00612
2271 U+0064b - U+0065e
2272 U+0074d - U+0076d
2273 U+01800 - U+01805
2274 U+01d00 - U+01d77
2275 U+01d9b - U+01dbf
2276 U+0200b - U+0200f
2277 U+030fc - U+030fe
2278 U+03260 - U+0327f
2279 U+0fb46 - U+0fbb1
2280 U+10450 - U+1049d
2281
2282 12. The -o option (show only the matching part of a line) for pcregrep was not
2283 compatible with GNU grep in that, if there was more than one match in a
2284 line, it showed only the first of them. It now behaves in the same way as
2285 GNU grep.
2286
2287 13. If the -o and -v options were combined for pcregrep, it printed a blank
2288 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2289 does the same. The return code can be used to tell if there were any
2290 non-matching lines.
2291
2292 14. Added --file-offsets and --line-offsets to pcregrep.
2293
2294 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2295 infinitely looping recursion. The bug was that positive lookaheads were not
2296 being skipped when checking for a possible empty match (negative lookaheads
2297 and both kinds of lookbehind were skipped).
2298
2299 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2300 inclusion of <windows.h> to before rather than after the definition of
2301 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2302
2303 17. Specifying a possessive quantifier with a specific limit for a Unicode
2304 character property caused pcre_compile() to compile bad code, which led at
2305 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2306 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2307 caused the error; without that there was no problem.
2308
2309 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2310
2311 19. Added --enable-pcretest-libreadline.
2312
2313 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2314 RE::GlobalReplace(). As a result, the number of replacements returned was
2315 double what it should be. I removed one of the increments, but Craig sent a
2316 later patch that removed the other one (the right fix) and added unit tests
2317 that check the return values (which was not done before).
2318
2319 21. Several CMake things:
2320
2321 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2322 the names libpcre and libpcreposix, not just pcre and pcreposix.
2323
2324 (2) The above change means that pcretest and pcregrep are now correctly
2325 linked with the newly-built libraries, not previously installed ones.
2326
2327 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2328
2329 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2330 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2331 UTF-8 newline character). The key issue is that the pattern starts .*;
2332 this means that the match must be either at the beginning, or after a
2333 newline. The bug was in the code for advancing after a failed match and
2334 checking that the new position followed a newline. It was not taking
2335 account of UTF-8 characters correctly.
2336
2337 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2338 character classes. PCRE was not treating the sequence [:...:] as a
2339 character class unless the ... were all letters. Perl, however, seems to
2340 allow any characters between [: and :], though of course it rejects as
2341 unknown any "names" that contain non-letters, because all the known class
2342 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2343 for example, whereas PCRE did not - it did not recognize a POSIX character
2344 class. This seemed a bit dangerous, so the code has been changed to be
2345 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2346 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2347 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2348 Perl does, and where it didn't before.
2349
2350 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2351 Windows environments %n is disabled by default.
2352
2353
2354 Version 7.4 21-Sep-07
2355 ---------------------
2356
2357 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2358 means that a class such as [\s] counted as "explicit reference to CR or
2359 LF". That isn't really right - the whole point of the change was to try to
2360 help when there was an actual mention of one of the two characters. So now
2361 the change happens only if \r or \n (or a literal CR or LF) character is
2362 encountered.
2363
2364 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2365 of both had grown to the point where there were only 3 bits left.
2366 Fortunately, there was spare space in the data structure, and so I have
2367 moved the internal flags into a new 16-bit field to free up more option
2368 bits.
2369
2370 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2371 but did not set the internal JCHANGED flag - either of these is enough to
2372 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2373 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2374 start sets both bits.
2375
2376 4. Added options (at build time, compile time, exec time) to change \R from
2377 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2378
2379 5. doc/pcresyntax.html was missing from the distribution.
2380
2381 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2382 compatibility, even though it is no longer used.
2383
2384 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2385 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2386 windows.h file is present (where different names are used). [This was
2387 reversed later after testing - see 16 below.]
2388
2389 8. Changed all #include <config.h> to #include "config.h". There were also
2390 some further <pcre.h> cases that I changed to "pcre.h".
2391
2392 9. When pcregrep was used with the --colour option, it missed the line ending
2393 sequence off the lines that it output.
2394
2395 10. It was pointed out to me that arrays of string pointers cause lots of
2396 relocations when a shared library is dynamically loaded. A technique of
2397 using a single long string with a table of offsets can drastically reduce
2398 these. I have refactored PCRE in four places to do this. The result is
2399 dramatic:
2400
2401 Originally: 290
2402 After changing UCP table: 187
2403 After changing error message table: 43
2404 After changing table of "verbs" 36
2405 After changing table of Posix names 22
2406
2407 Thanks to the folks working on Gregex for glib for this insight.
2408
2409 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2410 unicode-properties was also set.
2411
2412 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2413
2414 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2415 checked only for CRLF.
2416
2417 14. Added casts to pcretest.c to avoid compiler warnings.
2418
2419 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2420
2421 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2422 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2423 entirely. This removes changes made in 7 above.
2424
2425 17. The CMake files have been updated, and there is now more information about
2426 building with CMake in the NON-UNIX-USE document.
2427
2428
2429 Version 7.3 28-Aug-07
2430 ---------------------
2431
2432 1. In the rejigging of the build system that eventually resulted in 7.1, the
2433 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2434 brackets there is not right, since it causes compilers to look for an
2435 installed pcre.h, not the version that is in the source that is being
2436 compiled (which of course may be different). I have changed it back to:
2437
2438 #include "pcre.h"
2439
2440 I have a vague recollection that the change was concerned with compiling in
2441 different directories, but in the new build system, that is taken care of
2442 by the VPATH setting the Makefile.
2443
2444 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2445 when the subject happened to end in the byte 0x85 (e.g. if the last
2446 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2447 characters but of course it shouldn't be taken as a newline when it is part
2448 of another character. The bug was that, for an unlimited repeat of . in
2449 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2450 characters when looking for a newline.
2451
2452 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2453
2454 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2455 in debug output.
2456
2457 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2458 long printing in the pcrecpp unittest when running under MinGW.
2459
2460 6. ESC_K was left out of the EBCDIC table.
2461
2462 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2463 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2464 limit also applies to "virtual nesting" when a pattern is recursive, and in
2465 this case 1000 isn't so big. I have been able to remove this limit at the
2466 expense of backing off one optimization in certain circumstances. Normally,
2467 when pcre_exec() would call its internal match() function recursively and
2468 immediately return the result unconditionally, it uses a "tail recursion"
2469 feature to save stack. However, when a subpattern that can match an empty
2470 string has an unlimited repetition quantifier, it no longer makes this
2471 optimization. That gives it a stack frame in which to save the data for
2472 checking that an empty string has been matched. Previously this was taken
2473 from the 1000-entry workspace that had been reserved. So now there is no
2474 explicit limit, but more stack is used.
2475
2476 8. Applied Daniel's patches to solve problems with the import/export magic
2477 syntax that is required for Windows, and which was going wrong for the
2478 pcreposix and pcrecpp parts of the library. These were overlooked when this
2479 problem was solved for the main library.
2480
2481 9. There were some crude static tests to avoid integer overflow when computing
2482 the size of patterns that contain repeated groups with explicit upper
2483 limits. As the maximum quantifier is 65535, the maximum group length was
2484 set at 30,000 so that the product of these two numbers did not overflow a
2485 32-bit integer. However, it turns out that people want to use groups that
2486 are longer than 30,000 bytes (though not repeat them that many times).
2487 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2488 made it possible to implement the integer overflow checks in a much more
2489 dynamic way, which I have now done. The artificial limitation on group
2490 length has been removed - we now have only the limit on the total length of
2491 the compiled pattern, which depends on the LINK_SIZE setting.
2492
2493 10. Fixed a bug in the documentation for get/copy named substring when
2494 duplicate names are permitted. If none of the named substrings are set, the
2495 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2496 empty string.
2497
2498 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2499 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2500 because the ] is interpreted as the first data character and the
2501 terminating ] is not found. PCRE has been made compatible with Perl in this
2502 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2503 cause memory overwriting.
2504
2505 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2506 string has been matched (to stop an infinite loop). It was not recognizing
2507 a conditional subpattern that could match an empty string if that
2508 subpattern was within another subpattern. For example, it looped when
2509 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2510 condition was not nested. This bug has been fixed.
2511
2512 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2513 past the start of the subject in the presence of bytes with the top bit
2514 set, for example "\x8aBCD".
2515
2516 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2517 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2518
2519 14. Optimized (?!) to (*FAIL).
2520
2521 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2522 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2523 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2524 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2525 does: it's just the validity check that is more restrictive.
2526
2527 16. Inserted checks for integer overflows during escape sequence (backslash)
2528 processing, and also fixed erroneous offset values for syntax errors during
2529 backslash processing.
2530
2531 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2532 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2533
2534 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2535 caused an overrun.
2536
2537 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2538 something other than just ASCII characters) inside a group that had an
2539 unlimited repeat caused a loop at compile time (while checking to see
2540 whether the group could match an empty string).
2541
2542 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2543 [\P{Any}] did so. (Error in the code for printing property names.)
2544
2545 21. An orphan \E inside a character class could cause a crash.
2546
2547 22. A repeated capturing bracket such as (A)? could cause a wild memory
2548 reference during compilation.
2549
2550 23. There are several functions in pcre_compile() that scan along a compiled
2551 expression for various reasons (e.g. to see if it's fixed length for look
2552 behind). There were bugs in these functions when a repeated \p or \P was
2553 present in the pattern. These operators have additional parameters compared
2554 with \d, etc, and these were not being taken into account when moving along
2555 the compiled data. Specifically:
2556
2557 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2558 length.
2559
2560 (b) An item such as \pL+ within a repeated group could cause crashes or
2561 loops.
2562
2563 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2564 "reference to non-existent subpattern" error.
2565
2566 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2567
2568 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2569 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2570
2571 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2572
2573 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2574 character were causing crashes (broken optimization).
2575
2576 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2577 \p or \P) caused a compile-time loop.
2578
2579 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2580 break. For example, the unstudied pattern [\r\n]A does not match the string
2581 "\r\nA" because change 7.0/46 below moves the current point on by two
2582 characters after failing to match at the start. However, the pattern \nA
2583 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2584 the same is true. There doesn't seem any very clean way out of this, but
2585 what I have chosen to do makes the common cases work: PCRE now takes note
2586 of whether there can be an explicit match for \r or \n anywhere in the
2587 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2588 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2589 pattern has explicit CR or LF references.
2590
2591 29. Added (*CR) etc for changing newline setting at start of pattern.
2592
2593
2594 Version 7.2 19-Jun-07
2595 ---------------------
2596
2597 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2598 which is apparently normally available under Windows.
2599
2600 2. Re-jig the pcregrep tests with different newline settings in an attempt
2601 to make them independent of the local environment's newline setting.
2602
2603 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2604
2605 4. Some of the "internals" tests were previously cut out when the link size
2606 was not 2, because the output contained actual offsets. The recent new
2607 "Z" feature of pcretest means that these can be cut out, making the tests
2608 usable with all link sizes.
2609
2610 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2611 stack recursion. This gives a massive performance boost under BSD, but just
2612 a small improvement under Linux. However, it saves one field in the frame
2613 in all cases.
2614
2615 6. Added more features from the forthcoming Perl 5.10:
2616
2617 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2618 recursion call. It refers to the nth most recently opened parentheses.
2619
2620 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2621 to be opened parentheses.
2622
2623 (c) Conditions that refer to capturing parentheses can be specified
2624 relatively, for example, (?(-2)... or (?(+3)...
2625
2626 (d) \K resets the start of the current match so that everything before
2627 is not part of it.
2628
2629 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2630
2631 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2632 reference syntax.
2633
2634 (g) (?| introduces a group in which the numbering of parentheses in each
2635 alternative starts with the same number.
2636
2637 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2638
2639 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2640 PCRE_INFO_JCHANGED.
2641
2642 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2643 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2644 for detecting groups that can match an empty string.
2645
2646 9. A pattern with a very large number of alternatives (more than several
2647 hundred) was running out of internal workspace during the pre-compile
2648 phase, where pcre_compile() figures out how much memory will be needed. A
2649 bit of new cunning has reduced the workspace needed for groups with
2650 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2651 workspace instead of running out of the 4096 that are available.
2652
2653 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2654
2655 11. Applied patch from Google to remove an optimization that didn't quite work.
2656 The report of the bug said:
2657
2658 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2659 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2660 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2661
2662 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2663 it matched the wrong number of bytes.
2664
2665
2666 Version 7.1 24-Apr-07
2667 ---------------------
2668
2669 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2670 that is more "standard", making use of automake and other Autotools. There
2671 is some re-arrangement of the files and adjustment of comments consequent
2672 on this.
2673
2674 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2675 for recursive directory scanning broke on some systems because the files
2676 are not scanned in any specific order and on different systems the order
2677 was different. A call to "sort" has been inserted into RunGrepTest for the
2678 approprate test as a short-term fix. In the longer term there may be an
2679 alternative.
2680
2681 3. I had an email from Eric Raymond about problems translating some of PCRE's
2682 man pages to HTML (despite the fact that I distribute HTML pages, some
2683 people do their own conversions for various reasons). The problems
2684 concerned the use of low-level troff macros .br and .in. I have therefore
2685 removed all such uses from the man pages (some were redundant, some could
2686 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2687 HTML has been updated to handle .nf/.fi and to complain if it encounters
2688 .br or .in.
2689
2690 4. Updated comments in configure.ac that get placed in config.h.in and also
2691 arranged for config.h to be included in the distribution, with the name
2692 config.h.generic, for the benefit of those who have to compile without
2693 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2694
2695 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2696 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2697 makevp.bat for the current PCRE, using the additional files
2698 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2699
2700 6. A Windows user reported a minor discrepancy with test 2, which turned out
2701 to be caused by a trailing space on an input line that had got lost in his
2702 copy. The trailing space was an accident, so I've just removed it.
2703
2704 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2705 that is needed.
2706
2707 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2708 as "const" (a) because they are and (b) because it helps the PHP
2709 maintainers who have recently made a script to detect big data structures
2710 in the php code that should be moved to the .rodata section. I remembered
2711 to update Builducptable as well, so it won't revert if ucptable.h is ever
2712 re-created.
2713
2714 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2715 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2716 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2717 support is not required. This saves 1.5-2K of code, which is important in
2718 some applications.
2719
2720 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2721 so as not to refer to the tables, even though these functions will never be
2722 called when UTF-8 support is disabled. Otherwise there are problems with a
2723 shared library.
2724
2725 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2726
2727 (a) It was defining its arguments as char * instead of void *.
2728
2729 (b) It was assuming that all moves were upwards in memory; this was true
2730 a long time ago when I wrote it, but is no longer the case.
2731
2732 The emulated memove() is provided for those environments that have neither
2733 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2734 is clearly not the case, as these two bugs were recently reported.
2735
2736 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2737 and Detrail to create the HTML documentation, the .txt form of the man
2738 pages, and it removes trailing spaces from listed files. It also creates
2739 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2740 case, it wraps all the #defines with #ifndefs. This script should be run
2741 before "make dist".
2742
2743 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2744 with Unicode property support.
2745
2746 (a) For a maximizing quantifier, if the two different cases of the
2747 character were of different lengths in their UTF-8 codings (there are
2748 some cases like this - I found 11), and the matching function had to
2749 back up over a mixture of the two cases, it incorrectly assumed they
2750 were both the same length.
2751
2752 (b) When PCRE was configured to use the heap rather than the stack for
2753 recursion during matching, it was not correctly preserving the data for
2754 the other case of a UTF-8 character when checking ahead for a match
2755 while processing a minimizing repeat. If the check also involved
2756 matching a wide character, but failed, corruption could cause an
2757 erroneous result when trying to check for a repeat of the original
2758 character.
2759
2760 13. Some tidying changes to the testing mechanism:
2761
2762 (a) The RunTest script now detects the internal link size and whether there
2763 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2764 values substituted by "configure". (The RunGrepTest script already did
2765 this for UTF-8.) The configure.ac script no longer substitutes the
2766 relevant variables.
2767
2768 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2769 with length and offset values. This means that the output is different
2770 for different internal link sizes. Test 2 is skipped for link sizes
2771 other than 2 because of this, bypassing the problem. Unfortunately,
2772 there was also a test in test 3 (the locale tests) that used /B and
2773 failed for link sizes other than 2. Rather than cut the whole test out,
2774 I have added a new /Z option to pcretest that replaces the length and
2775 offset values with spaces. This is now used to make test 3 independent
2776 of link size. (Test 2 will be tidied up later.)
2777
2778 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2779 segmentation fault instead of returning the appropriate error message.
2780
2781 15. In multiline mode when the newline sequence was set to "any", the pattern
2782 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2783 This doesn't seem right; it now treats the CRLF combination as the line
2784 ending, and so does not match in that case. It's only a pattern such as ^$
2785 that would hit this one: something like ^ABC$ would have failed after \r
2786 and then tried again after \r\n.
2787
2788 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2789 in an attempt to make files that differ only in their line terminators
2790 compare equal. This works on Linux.
2791
2792 17. Under certain error circumstances pcregrep might try to free random memory
2793 as it exited. This is now fixed, thanks to valgrind.
2794
2795 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2796 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2797 was because its rules for how to advance for /g after matching an empty
2798 string at the end of a line did not allow for this case. They now check for
2799 it specially.
2800
2801 20. pcretest is supposed to handle patterns and data of any length, by
2802 extending its buffers when necessary. It was getting this wrong when the
2803 buffer for a data line had to be extended.
2804
2805 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2806 CRLF as a newline sequence.
2807
2808 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2809 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2810 I have nevertheless tidied it up.
2811
2812 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2813
2814 24. Added a man page for pcre-config.
2815
2816
2817 Version 7.0 19-Dec-06
2818 ---------------------
2819
2820 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2821 moving to gcc 4.1.1.
2822
2823 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2824 sys/time.h, which is documented as needed for this function. It doesn't
2825 seem to matter on Linux, but it showed up on some releases of OS X.
2826
2827 3. It seems that there are systems where bytes whose values are greater than
2828 127 match isprint() in the "C" locale. The "C" locale should be the
2829 default when a C program starts up. In most systems, only ASCII printing
2830 characters match isprint(). This difference caused the output from pcretest
2831 to vary, making some of the tests fail. I have changed pcretest so that:
2832
2833 (a) When it is outputting text in the compiled version of a pattern, bytes
2834 other than 32-126 are always shown as hex escapes.
2835
2836 (b) When it is outputting text that is a matched part of a subject string,
2837 it does the same, unless a different locale has been set for the match
2838 (using the /L modifier). In this case, it uses isprint() to decide.
2839
2840 4. Fixed a major bug that caused incorrect computation of the amount of memory
2841 required for a compiled pattern when options that changed within the
2842 pattern affected the logic of the preliminary scan that determines the
2843 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2844 that the computed length was too small. The symptoms of this bug were
2845 either the PCRE error "internal error: code overflow" from pcre_compile(),
2846 or a glibc crash with a message such as "pcretest: free(): invalid next
2847 size (fast)". Examples of patterns that provoked this bug (shown in
2848 pcretest format) are:
2849
2850 /(?-x: )/x
2851 /(?x)(?-x: \s*#\s*)/
2852 /((?i)[\x{c0}])/8
2853 /(?i:[\x{c0}])/8
2854
2855 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2856 is now done differently.
2857
2858 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2859 wrapper classes; (b) implement a new function in the C++ scanner that is
2860 more efficient than the old way of doing things because it avoids levels of
2861 recursion in the regex matching; (c) add a paragraph to the documentation
2862 for the FullMatch() function.
2863
2864 6. The escape sequence \n was being treated as whatever was defined as
2865 "newline". Not only was this contrary to the documentation, which states
2866 that \n is character 10 (hex 0A), but it also went horribly wrong when
2867 "newline" was defined as CRLF. This has been fixed.
2868
2869 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2870 was being set to -1 for the "end of line" case (supposedly a value that no
2871 character can have). Though this value is never used (the check for end of
2872 line is "zero bytes in current character"), it caused compiler complaints.
2873 I've changed it to 0xffffffff.
2874
2875 8. In pcre_version.c, the version string was being built by a sequence of
2876 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2877 string (as it is for production releases) called a macro with an empty
2878 argument. The C standard says the result of this is undefined. The gcc
2879 compiler treats it as an empty string (which was what was wanted) but it is
2880 reported that Visual C gives an error. The source has been hacked around to
2881 avoid this problem.
2882
2883 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2884 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2885 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2886 of them did).
2887
2888 10. Originally, pcretest opened its input and output without "b"; then I was
2889 told that "b" was needed in some environments, so it was added for release
2890 5.0 to both the input and output. (It makes no difference on Unix-like
2891 systems.) Later I was told that it is wrong for the input on Windows. I've
2892 now abstracted the modes into two macros, to make it easier to fiddle with
2893 them, and removed "b" from the input mode under Windows.
2894
2895 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2896
2897 12. Added -help and --help to pcretest as an official way of being reminded
2898 of the options.
2899
2900 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2901 and pcrecpp.cc because they annoy compilers at high warning levels.
2902
2903 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2904
2905 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2906 scripts are not C programs :-) and which was not noticed because it works
2907 on Linux.
2908
2909 16. pcretest is supposed to handle any length of pattern and data line (as one
2910 line or as a continued sequence of lines) by extending its input buffer if
2911 necessary. This feature was broken for very long pattern lines, leading to
2912 a string of junk being passed to pcre_compile() if the pattern was longer
2913 than about 50K.
2914
2915 17. I have done a major re-factoring of the way pcre_compile() computes the
2916 amount of memory needed for a compiled pattern. Previously, there was code
2917 that made a preliminary scan of the pattern in order to do this. That was
2918 OK when PCRE was new, but as the facilities have expanded, it has become
2919 harder and harder to keep it in step with the real compile phase, and there
2920 have been a number of bugs (see for example, 4 above). I have now found a
2921 cunning way of running the real compile function in a "fake" mode that
2922 enables it to compute how much memory it would need, while actually only
2923 ever using a few hundred bytes of working memory and without too many
2924 tests of the mode. This should make future maintenance and development
2925 easier. A side effect of this work is that the limit of 200 on the nesting
2926 depth of parentheses has been removed (though this was never a serious
2927 limitation, I suspect). However, there is a downside: pcre_compile() now
2928 runs more slowly than before (30% or more, depending on the pattern). I
2929 hope this isn't a big issue. There is no effect on runtime performance.
2930
2931 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2932 newline (only possible for the last line of a file) and it was a
2933 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2934
2935 19. Added additional timing features to pcretest. (1) The -tm option now times
2936 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2937 separate command line item, by a number that specifies the number of
2938 repeats to use when timing. The default is 50000; this gives better
2939 precision, but takes uncomfortably long for very large patterns.
2940
2941 20. Extended pcre_study() to be more clever in cases where a branch of a
2942 subpattern has no definite first character. For example, (a*|b*)[cd] would
2943 previously give no result from pcre_study(). Now it recognizes that the
2944 first character must be a, b, c, or d.
2945
2946 21. There was an incorrect error "recursive call could loop indefinitely" if
2947 a subpattern (or the entire pattern) that was being tested for matching an
2948 empty string contained only one non-empty item after a nested subpattern.
2949 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2950 incorrectly, because the \d was being skipped in the check.
2951
2952 22. The pcretest program now has a new pattern option /B and a command line
2953 option -b, which is equivalent to adding /B to every pattern. This causes
2954 it to show the compiled bytecode, without the additional information that
2955 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2956 is the same as /B/I).
2957
2958 23. A new optimization is now able automatically to treat some sequences such
2959 as a*b as a*+b. More specifically, if something simple (such as a character
2960 or a simple class like \d) has an unlimited quantifier, and is followed by
2961 something that cannot possibly match the quantified thing, the quantifier
2962 is automatically "possessified".
2963
2964 24. A recursive reference to a subpattern whose number was greater than 39
2965 went wrong under certain circumstances in UTF-8 mode. This bug could also
2966 have affected the operation of pcre_study().
2967
2968 25. Realized that a little bit of performance could be had by replacing
2969 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2970
2971 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2972
2973 27. Possessive quantifiers such as a++ were previously implemented by turning
2974 them into atomic groups such as ($>a+). Now they have their own opcodes,
2975 which improves performance. This includes the automatically created ones
2976 from 23 above.
2977
2978 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2979 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2980 the first matched character to be a colon. This applied both to named and
2981 numbered groups.
2982
2983 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2984
2985 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2986 building PCRE on Windows easier, so I have included it in the distribution.
2987
2988 31. There is now a check in pcretest against a ridiculously large number being
2989 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2990 loop, the loop is abandoned.
2991
2992 32. Forward references to subpatterns in conditions such as (?(2)...) where
2993 subpattern 2 is defined later cause pcre_compile() to search forwards in
2994 the pattern for the relevant set of parentheses. This search went wrong
2995 when there were unescaped parentheses in a character class, parentheses
2996 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2997
2998 33. "Subroutine" calls and backreferences were previously restricted to
2999 referencing subpatterns earlier in the regex. This restriction has now
3000 been removed.
3001
3002 34. Added a number of extra features that are going to be in Perl 5.10. On the
3003 whole, these are just syntactic alternatives for features that PCRE had
3004 previously implemented using the Python syntax or my own invention. The
3005 other formats are all retained for compatibility.
3006
3007 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3008 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3009 also .NET compatible.
3010
3011 (b) A recursion or subroutine call to a named group can now be defined as
3012 (?&name) as well as (?P>name).
3013
3014 (c) A backreference to a named group can now be defined as \k<name> or
3015 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3016 5.10, are also .NET compatible.
3017
3018 (d) A conditional reference to a named group can now use the syntax
3019 (?(<name>) or (?('name') as well as (?(name).
3020
3021 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3022 groups (named and numbered) that are never evaluated inline, but can be
3023 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3024 is always false. There may be only one alternative in such a group.
3025
3026 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3027 as the simple (?(R). The condition is true only if the most recent
3028 recursion is that of the given number or name. It does not search out
3029 through the entire recursion stack.
3030
3031 (g) The escape \gN or \g{N} has been added, where N is a positive or
3032 negative number, specifying an absolute or relative reference.
3033
3034 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3035 some "unreachable code" warnings.
3036
3037 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3038 things, this adds five new scripts.
3039
3040 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3041 There were also incompatibilities regarding the handling of \Q..\E inside
3042 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3043 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3044
3045 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3046 matches an empty string, and forcibly breaks the loop. There were bugs in
3047 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3048 against aaaa the result was just "a" rather than "aaaa", for example. Two
3049 separate and independent bugs (that affected different cases) have been
3050 fixed.
3051
3052 39. Refactored the code to abolish the use of different opcodes for small
3053 capturing bracket numbers. This is a tidy that I avoided doing when I
3054 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3055 The new approach is not only tidier, it makes it possible to reduce the
3056 memory needed to fix the previous bug (38).
3057
3058 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3059 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3060 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3061 mode.
3062
3063 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3064 report.
3065
3066 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3067 copy construction and assignment in the C++ wrapper.
3068
3069 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3070 couple of bugs that could have given wrong results in the "--newline=crlf"
3071 case.
3072
3073 44. Added a number of casts and did some reorganization of signed/unsigned int
3074 variables following suggestions from Dair Grant. Also renamed the variable
3075 "this" as "item" because it is a C++ keyword.
3076
3077 45. Arranged for dftables to add
3078
3079 #include "pcre_internal.h"
3080
3081 to pcre_chartables.c because without it, gcc 4.x may remove the array
3082 definition from the final binary if PCRE is built into a static library and
3083 dead code stripping is activated.
3084
3085 46. For an unanchored pattern, if a match attempt fails at the start of a
3086 newline sequence, and the newline setting is CRLF or ANY, and the next two
3087 characters are CRLF, advance by two characters instead of one.
3088
3089
3090 Version 6.7 04-Jul-06
3091 ---------------------
3092
3093 1. In order to handle tests when input lines are enormously long, pcretest has
3094 been re-factored so that it automatically extends its buffers when
3095 necessary. The code is crude, but this _is_ just a test program. The
3096 default size has been increased from 32K to 50K.
3097
3098 2. The code in pcre_study() was using the value of the re argument before
3099 testing it for NULL. (Of course, in any sensible call of the function, it
3100 won't be NULL.)
3101
3102 3. The memmove() emulation function in pcre_internal.h, which is used on
3103 systems that lack both memmove() and bcopy() - that is, hardly ever -
3104 was missing a "static" storage class specifier.
3105
3106 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3107 containing an extended class (one that cannot be represented by a bitmap
3108 because it contains high-valued characters or Unicode property items, e.g.
3109 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3110 pattern, but PCRE should not loop if you do not (it no longer does).
3111 [Detail: two cases were found: (a) a repeated subpattern containing an
3112 extended class; (b) a recursive reference to a subpattern that followed a
3113 previous extended class. It wasn't skipping over the extended class
3114 correctly when UTF-8 mode was not set.]
3115
3116 5. A negated single-character class was not being recognized as fixed-length
3117 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3118 compile error "lookbehind assertion is not fixed length".
3119
3120 6. The RunPerlTest auxiliary script was showing an unexpected difference
3121 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3122 write a Perl script that can interpret lines of an input file either as
3123 byte characters or as UTF-8, which is what "perltest" was being required to
3124 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3125 can't do is switch easily at run time between having the "use utf8;" pragma
3126 or not. In the end, I fudged it by using the RunPerlTest script to insert
3127 "use utf8;" explicitly for the UTF-8 tests.
3128
3129 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3130 the end of the subject string, contrary to the documentation and to what
3131 Perl does. This was true of both matching functions. Now it matches only at
3132 the start of the subject and immediately after *internal* newlines.
3133
3134 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3135 a pointer to an int instead of a pointer to an unsigned long int. This
3136 caused problems on 64-bit systems.
3137
3138 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3139 instance of the 'standard' template library not being so standard".
3140
3141 10. There was no check on the number of named subpatterns nor the maximum
3142 length of a subpattern name. The product of these values is used to compute
3143 the size of the memory block for a compiled pattern. By supplying a very
3144 long subpattern name and a large number of named subpatterns, the size
3145 computation could be caused to overflow. This is now prevented by limiting
3146 the length of names to 32 characters, and the number of named subpatterns
3147 to 10,000.
3148
3149 11. Subpatterns that are repeated with specific counts have to be replicated in
3150 the compiled pattern. The size of memory for this was computed from the
3151 length of the subpattern and the repeat count. The latter is limited to
3152 65535, but there was no limit on the former, meaning that integer overflow
3153 could in principle occur. The compiled length of a repeated subpattern is
3154 now limited to 30,000 bytes in order to prevent this.
3155
3156 12. Added the optional facility to have named substrings with the same name.
3157
3158 13. Added the ability to use a named substring as a condition, using the
3159 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3160 are numbers (not recommended). Forward references are permitted.
3161
3162 14. Added forward references in named backreferences (if you see what I mean).
3163
3164 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3165 pattern could run off the end of the subject. For example, the pattern
3166 "(?s)(.{1,5})"8 did this with the subject "ab".
3167
3168 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3169 PCRE_CASELESS was set when matching characters that were quantified with ?
3170 or *.
3171
3172 17. A character class other than a single negated character that had a minimum
3173 but no maximum quantifier - for example [ab]{6,} - was not handled
3174 correctly by pce_dfa_exec(). It would match only one character.
3175
3176 18. A valid (though odd) pattern that looked like a POSIX character
3177 class but used an invalid character after [ (for example [[,abc,]]) caused
3178 pcre_compile() to give the error "Failed: internal error: code overflow" or
3179 in some cases to crash with a glibc free() error. This could even happen if
3180 the pattern terminated after [[ but there just happened to be a sequence of
3181 letters, a binary zero, and a closing ] in the memory that followed.
3182
3183 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3184 over the years. Originally (before any Unicode support), just the bottom 8
3185 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3186 output from "man perlunicode" includes this:
3187
3188 The regular expression compiler produces polymorphic opcodes. That
3189 is, the pattern adapts to the data and automatically switches to
3190 the Unicode character scheme when presented with Unicode data--or
3191 instead uses a traditional byte scheme when presented with byte
3192 data.
3193
3194 Sadly, a wide octal escape does not cause a switch, and in a string with
3195 no other multibyte characters, these octal escapes are treated as before.
3196 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3197 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3198 Unicode string.
3199
3200 I have not perpetrated such confusion in PCRE. Up till now, it took just
3201 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3202 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3203 translate to the appropriate multibyte character.
3204
3205 29. Applied some refactoring to reduce the number of warnings from Microsoft
3206 and Borland compilers. This has included removing the fudge introduced
3207 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3208 a warning about an unused variable.
3209
3210 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3211 characters since release 4.0, because Perl (from release 5.004) does not.
3212 [Or at least, is documented not to: some releases seem to be in conflict
3213 with the documentation.] However, when a pattern was studied with
3214 pcre_study() and all its branches started with \s, PCRE still included VT
3215 as a possible starting character. Of course, this did no harm; it just
3216 caused an unnecessary match attempt.
3217
3218 22. Removed a now-redundant internal flag bit that recorded the fact that case
3219 dependency changed within the pattern. This was once needed for "required
3220 byte" processing, but is no longer used. This recovers a now-scarce options
3221 bit. Also moved the least significant internal flag bit to the most-
3222 significant bit of the word, which was not previously used (hangover from
3223 the days when it was an int rather than a uint) to free up another bit for
3224 the future.
3225
3226 23. Added support for CRLF line endings as well as CR and LF. As well as the
3227 default being selectable at build time, it can now be changed at runtime
3228 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3229 specify that it is scanning data with non-default line endings.
3230
3231 24. Changed the definition of CXXLINK to make it agree with the definition of
3232 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3233
3234 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3235 recursions. This makes a big different to stack usage for some patterns.
3236
3237 26. If a subpattern containing a named recursion or subroutine reference such
3238 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3239 the space required for the compiled pattern went wrong and gave too small a
3240 value. Depending on the environment, this could lead to "Failed: internal
3241 error: code overflow at offset 49" or "glibc detected double free or
3242 corruption" errors.
3243
3244 27. Applied patches from Google (a) to support the new newline modes and (b) to
3245 advance over multibyte UTF-8 characters in GlobalReplace.
3246
3247 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3248 difference for some implementation of PCRE in some Windows version.
3249
3250 29. Added some extra testing facilities to pcretest:
3251
3252 \q<number> in a data line sets the "match limit" value
3253 \Q<number> in a data line sets the "match recursion limt" value
3254 -S <number> sets the stack size, where <number> is in megabytes
3255
3256 The -S option isn't available for Windows.
3257
3258
3259 Version 6.6 06-Feb-06
3260 ---------------------
3261
3262 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3263 in pcreposix.h. I have copied the definition from pcre.h.
3264
3265 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3266 because pcre.h is no longer a built file.
3267
3268 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3269 not normally included in the compiled code.
3270
3271
3272 Version 6.5 01-Feb-06
3273 ---------------------
3274
3275 1. When using the partial match feature with pcre_dfa_exec(), it was not
3276 anchoring the second and subsequent partial matches at the new starting
3277 point. This could lead to incorrect results. For example, with the pattern
3278 /1234/, partially matching against "123" and then "a4" gave a match.
3279
3280 2. Changes to pcregrep:
3281
3282 (a) All non-match returns from pcre_exec() were being treated as failures
3283 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3284 error message is output. Some extra information is given for the
3285 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3286 probably the only errors that are likely to be caused by users (by
3287 specifying a regex that has nested indefinite repeats, for instance).
3288 If there are more than 20 of these errors, pcregrep is abandoned.
3289
3290 (b) A binary zero was treated as data while matching, but terminated the
3291 output line if it was written out. This has been fixed: binary zeroes
3292 are now no different to any other data bytes.
3293
3294 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3295 used to set a locale for matching. The --locale=xxxx long option has
3296 been added (no short equivalent) to specify a locale explicitly on the
3297 pcregrep command, overriding the environment variables.
3298
3299 (d) When -B was used with -n, some line numbers in the output were one less
3300 than they should have been.
3301
3302 (e) Added the -o (--only-matching) option.
3303
3304 (f) If -A or -C was used with -c (count only), some lines of context were
3305 accidentally printed for the final match.
3306
3307 (g) Added the -H (--with-filename) option.
3308
3309 (h) The combination of options -rh failed to suppress file names for files
3310 that were found from directory arguments.
3311
3312 (i) Added the -D (--devices) and -d (--directories) options.
3313
3314 (j) Added the -F (--fixed-strings) option.
3315
3316 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3317
3318 (l) Added the --colo(u)r option.
3319
3320 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3321 is not present by default.
3322
3323 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3324 items such as (?R) or (?1), when the recursion could match a number of
3325 alternatives. If it matched one of the alternatives, but subsequently,
3326 outside the recursion, there was a failure, the code tried to back up into
3327 the recursion. However, because of the way PCRE is implemented, this is not
3328 possible, and the result was an incorrect result from the match.
3329
3330 In order to prevent this happening, the specification of recursion has
3331 been changed so that all such subpatterns are automatically treated as
3332 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3333
3334 4. I had overlooked the fact that, in some locales, there are characters for
3335 which isalpha() is true but neither isupper() nor islower() are true. In
3336 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3337 and ordfeminine) are like this. This affected the treatment of \w and \W
3338 when they appeared in character classes, but not when they appeared outside
3339 a character class. The bit map for "word" characters is now created
3340 separately from the results of isalnum() instead of just taking it from the
3341 upper, lower, and digit maps. (Plus the underscore character, of course.)
3342
3343 5. The above bug also affected the handling of POSIX character classes such as
3344 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3345 permanent tables. Instead, the bit maps for such a class were previously
3346 created as the appropriate unions of the upper, lower, and digit bitmaps.
3347 Now they are created by subtraction from the [[:word:]] class, which has
3348 its own bitmap.
3349
3350 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3351 It is created by subtracting the vertical space characters (\x09, \x0a,
3352 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3353 subtraction was done in the overall bitmap for a character class, meaning
3354 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3355 be recognized. This bug has been fixed.
3356
3357 7. Patches from the folks at Google:
3358
3359 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3360 real life, but is still worth protecting against".
3361
3362 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3363 regular expressions".
3364
3365 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3366 have it.
3367
3368 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3369 "configure" and the latter not, in order to fix a problem somebody had
3370 with compiling the Arg class on HP-UX.
3371
3372 (e) Improve the error-handling of the C++ wrapper a little bit.
3373
3374 (f) New tests for checking recursion limiting.
3375
3376 8. The pcre_memmove() function, which is used only if the environment does not
3377 have a standard memmove() function (and is therefore rarely compiled),
3378 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3379 returning a result (though PCRE never actually uses the result).
3380
3381 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3382 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3383 returned instead of calling malloc() with an overflowing number that would
3384 most likely cause subsequent chaos.
3385
3386 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3387
3388 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3389 with this option is matched, the nmatch and pmatch options of regexec() are
3390 ignored.
3391
3392 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3393 provided in case anyone wants to the the POSIX interface with UTF-8
3394 strings.
3395
3396 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3397 C++ linking (needed for some HP-UX environments).
3398
3399 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3400 (unused parameter) and in the pcre_printint() function (omitted "default"
3401 switch label when the default is to do nothing).
3402
3403 15. Added some code to make it possible, when PCRE is compiled as a C++
3404 library, to replace subject pointers for pcre_exec() with a smart pointer
3405 class, thus making it possible to process discontinuous strings.
3406
3407 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3408 much the same function. They were added by different people who were trying
3409 to make PCRE easy to compile on non-Unix systems. It has been suggested
3410 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3411 for compiling on Windows systems. I have therefore replaced it with
3412 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3413 defaults to "extern" for C or "extern C" for C++, which works fine on
3414 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3415 SCOPE with something explicit in config.h. In addition:
3416
3417 (a) pcreposix.h still had just "extern" instead of either of these macros;
3418 I have replaced it with PCRE_DATA_SCOPE.
3419
3420 (b) Functions such as _pcre_xclass(), which are internal to the library,
3421 but external in the C sense, all had PCRE_EXPORT in their definitions.
3422 This is apparently wrong for the Windows case, so I have removed it.
3423 (It makes no difference on Unix-like systems.)
3424
3425 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3426 of recursive calls to match(). This is different to MATCH_LIMIT because
3427 that limits the total number of calls to match(), not all of which increase
3428 the depth of recursion. Limiting the recursion depth limits the amount of
3429 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3430 when PCRE is compiled, and changed at run time. A patch from Google adds
3431 this functionality to the C++ interface.
3432
3433 18. Changes to the handling of Unicode character properties:
3434
3435 (a) Updated the table to Unicode 4.1.0.
3436
3437 (b) Recognize characters that are not in the table as "Cn" (undefined).
3438
3439 (c) I revised the way the table is implemented to a much improved format
3440 which includes recognition of ranges. It now supports the ranges that
3441 are defined in UnicodeData.txt, and it also amalgamates other
3442 characters into ranges. This has reduced the number of entries in the
3443 table from around 16,000 to around 3,000, thus reducing its size
3444 considerably. I realized I did not need to use a tree structure after
3445 all - a binary chop search is just as efficient. Having reduced the
3446 number of entries, I extended their size from 6 bytes to 8 bytes to
3447 allow for more data.
3448
3449 (d) Added support for Unicode script names via properties such as \p{Han}.
3450
3451 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3452 matching that character.
3453
3454 20. When matching a repeated Unicode property with a minimum greater than zero,
3455 (for example \pL{2,}), PCRE could look past the end of the subject if it
3456 reached it while seeking the minimum number of characters. This could
3457 happen only if some of the characters were more than one byte long, because
3458 there is a check for at least the minimum number of bytes.
3459
3460 21. Refactored the implementation of \p and \P so as to be more general, to
3461 allow for more different types of property in future. This has changed the
3462 compiled form incompatibly. Anybody with saved compiled patterns that use
3463 \p or \P will have to recompile them.
3464
3465 22. Added "Any" and "L&" to the supported property types.
3466
3467 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3468 but give a compile time error if the value is greater than 0xff.
3469
3470 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3471 accidentally not being installed or uninstalled.
3472
3473 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3474 made were to insert the current release number. This seemed silly, because
3475 it made things harder for people building PCRE on systems that don't run
3476 "configure". I have turned pcre.h into a distributed file, no longer built
3477 by "configure", with the version identification directly included. There is
3478 no longer a pcre.h.in file.
3479
3480 However, this change necessitated a change to the pcre-config script as
3481 well. It is built from pcre-config.in, and one of the substitutions was the
3482 release number. I have updated configure.ac so that ./configure now finds
3483 the release number by grepping pcre.h.
3484
3485 26. Added the ability to run the tests under valgrind.
3486
3487
3488 Version 6.4 05-Sep-05
3489 ---------------------
3490
3491 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3492 "--" to be printed when multiple files were scanned, even when none of the
3493 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3494 consider it to be a bug, and have restored the previous behaviour.
3495
3496 2. A couple of code tidies to get rid of compiler warnings.
3497
3498 3. The pcretest program used to cheat by referring to symbols in the library
3499 whose names begin with _pcre_. These are internal symbols that are not
3500 really supposed to be visible externally, and in some environments it is
3501 possible to suppress them. The cheating is now confined to including
3502 certain files from the library's source, which is a bit cleaner.
3503
3504 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3505 file's purpose clearer.
3506
3507 5. Reorganized pcre_ucp_findchar().
3508
3509
3510 Version 6.3 15-Aug-05
3511 ---------------------
3512
3513 1. The file libpcre.pc.in did not have general read permission in the tarball.
3514
3515 2. There were some problems when building without C++ support:
3516
3517 (a) If C++ support was not built, "make install" and "make test" still
3518 tried to test it.
3519
3520 (b) There were problems when the value of CXX was explicitly set. Some
3521 changes have been made to try to fix these, and ...
3522
3523 (c) --disable-cpp can now be used to explicitly disable C++ support.
3524
3525 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3526 backslash in a target when C++ was disabled. This confuses some
3527 versions of "make", apparently. Using an intermediate variable solves
3528 this. (Same for CPP_LOBJ.)
3529
3530 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3531 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3532 necessary on certain architectures.
3533
3534 4. Added a setting of -export-symbols-regex to the link command to remove
3535 those symbols that are exported in the C sense, but actually are local
3536 within the library, and not documented. Their names all begin with
3537 "_pcre_". This is not a perfect job, because (a) we have to except some
3538 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3539 available (and never for static libraries). I have made a note to try to
3540 find a way round (a) in the future.
3541
3542
3543 Version 6.2 01-Aug-05
3544 ---------------------
3545
3546 1. There was no test for integer overflow of quantifier values. A construction
3547 such as {1111111111111111} would give undefined results. What is worse, if
3548 a minimum quantifier for a parenthesized subpattern overflowed and became
3549 negative, the calculation of the memory size went wrong. This could have
3550 led to memory overwriting.
3551
3552 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3553
3554 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3555 operating environments where this matters.
3556
3557 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3558 PCRE options from within the C++ wrapper.
3559
3560 5. Named capturing subpatterns were not being correctly counted when a pattern
3561 was compiled. This caused two problems: (a) If there were more than 100
3562 such subpatterns, the calculation of the memory needed for the whole
3563 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3564 back references of the form \12, where the number was greater than 9, were
3565 not recognized as back references, even though there were sufficient
3566 previous subpatterns.
3567
3568 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3569 versions of gcc, e.g. 2.95.4.
3570
3571
3572 Version 6.1 21-Jun-05
3573 ---------------------
3574
3575 1. There was one reference to the variable "posix" in pcretest.c that was not
3576 surrounded by "#if !defined NOPOSIX".
3577
3578 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3579 the cross-check on the old pcre_info() function, for the benefit of the
3580 cut-down version of PCRE that is currently imported into Exim.
3581
3582 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3583 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3584 patterns that start (?i) but I don't think that matters. The use of (?i) is
3585 just an example; this all applies to the other options as well.
3586
3587 4. Since libtool seems to echo the compile commands it is issuing, the output
3588 from "make" can be reduced a bit by putting "@" in front of each libtool
3589 compile command.
3590
3591 5. Patch from the folks at Google for configure.in to be a bit more thorough
3592 in checking for a suitable C++ installation before trying to compile the
3593 C++ stuff. This should fix a reported problem when a compiler was present,
3594 but no suitable headers.
3595
3596 6. The man pages all had just "PCRE" as their title. I have changed them to
3597 be the relevant file name. I have also arranged that these names are
3598 retained in the file doc/pcre.txt, which is a concatenation in text format
3599 of all the man pages except the little individual ones for each function.
3600
3601 7. The NON-UNIX-USE file had not been updated for the different set of source
3602 files that come with release 6. I also added a few comments about the C++
3603 wrapper.
3604
3605
3606 Version 6.0 07-Jun-05
3607 ---------------------
3608
3609 1. Some minor internal re-organization to help with my DFA experiments.
3610
3611 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3612 didn't matter for the library itself when fully configured, but did matter
3613 when compiling without UCP support, or within Exim, where the ucp files are
3614 not imported.
3615
3616 3. Refactoring of the library code to split up the various functions into
3617 different source modules. The addition of the new DFA matching code (see
3618 below) to a single monolithic source would have made it really too
3619 unwieldy, quite apart from causing all the code to be include in a
3620 statically linked application, when only some functions are used. This is
3621 relevant even without the DFA addition now that patterns can be compiled in
3622 one application and matched in another.
3623
3624 The downside of splitting up is that there have to be some external
3625 functions and data tables that are used internally in different modules of
3626 the library but which are not part of the API. These have all had their
3627 names changed to start with "_pcre_" so that they are unlikely to clash
3628 with other external names.
3629
3630 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3631 a different (DFA) algorithm. Although it is slower than the original
3632 function, it does have some advantages for certain types of matching
3633 problem.
3634
3635 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3636 including restarting after a partial match.
3637
3638 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3639 defined when compiling for Windows was sent to me. I have put it into the
3640 code, though I have no means of testing or verifying it.
3641
3642 7. Added the pcre_refcount() auxiliary function.
3643
3644 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3645 match before or at the first newline in the subject string. In pcretest,
3646 the /f option on a pattern can be used to set this.
3647
3648 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3649 would behave wrongly. This has been present in PCRE since release 4.0.
3650
3651 10. A number of changes to the pcregrep command:
3652
3653 (a) Refactored how -x works; insert ^(...)$ instead of setting
3654 PCRE_ANCHORED and checking the length, in preparation for adding
3655 something similar for -w.
3656
3657 (b) Added the -w (match as a word) option.
3658
3659 (c) Refactored the way lines are read and buffered so as to have more
3660 than one at a time available.
3661
3662 (d) Implemented a pcregrep test script.
3663
3664 (e) Added the -M (multiline match) option. This allows patterns to match
3665 over several lines of the subject. The buffering ensures that at least
3666 8K, or the rest of the document (whichever is the shorter) is available
3667 for matching (and similarly the previous 8K for lookbehind assertions).
3668
3669 (f) Changed the --help output so that it now says
3670
3671 -w, --word-regex(p)
3672
3673 instead of two lines, one with "regex" and the other with "regexp"
3674 because that confused at least one person since the short forms are the
3675 same. (This required a bit of code, as the output is generated
3676 automatically from a table. It wasn't just a text change.)
3677
3678 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3679 option but starts with a hyphen. Could be a pattern or a path name
3680 starting with a hyphen, for instance.
3681
3682 (h) "-" can be given as a file name to represent stdin.
3683
3684 (i) When file names are being printed, "(standard input)" is used for
3685 the standard input, for compatibility with GNU grep. Previously
3686 "<stdin>" was used.
3687
3688 (j) The option --label=xxx can be used to supply a name to be used for
3689 stdin when file names are being printed. There is no short form.
3690
3691 (k) Re-factored the options decoding logic because we are going to add
3692 two more options that take data. Such options can now be given in four
3693 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3694
3695 (l) Added the -A, -B, and -C options for requesting that lines of context
3696 around matches be printed.
3697
3698 (m) Added the -L option to print the names of files that do not contain
3699 any matching lines, that is, the complement of -l.
3700
3701 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3702 continue to scan other files.
3703
3704 (o) The -s option was incorrectly implemented. For compatibility with other
3705 greps, it now suppresses the error message for a non-existent or non-
3706 accessible file (but not the return code). There is a new option called
3707 -q that suppresses the output of matching lines, which was what -s was
3708 previously doing.
3709
3710 (p) Added --include and --exclude options to specify files for inclusion
3711 and exclusion when recursing.
3712
3713 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3714 Hopefully, it now does.
3715
3716 12. Missing cast in pcre_study().
3717
3718 13. Added an "uninstall" target to the makefile.
3719
3720 14. Replaced "extern" in the function prototypes in Makefile.in with
3721 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3722 world, but is set differently for Windows.
3723
3724 15. Added a second compiling function called pcre_compile2(). The only
3725 difference is that it has an extra argument, which is a pointer to an
3726 integer error code. When there is a compile-time failure, this is set
3727 non-zero, in addition to the error test pointer being set to point to an
3728 error message. The new argument may be NULL if no error number is required
3729 (but then you may as well call pcre_compile(), which is now just a
3730 wrapper). This facility is provided because some applications need a
3731 numeric error indication, but it has also enabled me to tidy up the way
3732 compile-time errors are handled in the POSIX wrapper.
3733
3734 16. Added VPATH=.libs to the makefile; this should help when building with one
3735 prefix path and installing with another. (Or so I'm told by someone who
3736 knows more about this stuff than I do.)
3737
3738 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3739 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3740 match everything, including newlines. This is not POSIX-compatible, but
3741 somebody wanted the feature. From pcretest it can be activated by using
3742 both the P and the s flags.
3743
3744 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3745
3746 19. libpcre.pc was being incorrectly installed as executable.
3747
3748 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3749 it now also looks for '\r' so that it will work unmodified on Windows.
3750
3751 21. Added Google's contributed C++ wrapper to the distribution.
3752
3753 22. Added some untidy missing memory free() calls in pcretest, to keep
3754 Electric Fence happy when testing.
3755
3756
3757
3758 Version 5.0 13-Sep-04
3759 ---------------------
3760
3761 1. Internal change: literal characters are no longer packed up into items
3762 containing multiple characters in a single byte-string. Each character
3763 is now matched using a separate opcode. However, there may be more than one
3764 byte in the character in UTF-8 mode.
3765
3766 2. The pcre_callout_block structure has two new fields: pattern_position and
3767 next_item_length. These contain the offset in the pattern to the next match
3768 item, and its length, respectively.
3769
3770 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3771 insertion of callouts before each pattern item. Added the /C option to
3772 pcretest to make use of this.
3773
3774 4. On the advice of a Windows user, the lines
3775
3776 #if defined(_WIN32) || defined(WIN32)
3777 _setmode( _fileno( stdout ), 0x8000 );
3778 #endif /* defined(_WIN32) || defined(WIN32) */
3779
3780 have been added to the source of pcretest. This apparently does useful
3781 magic in relation to line terminators.
3782
3783 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3784 for the benefit of those environments where the "b" makes a difference.
3785
3786 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3787 to know about it. I have put a hack into configure.in that adds in code
3788 to set GCC=yes if CC=icc. This seems to end up at a point in the
3789 generated configure script that is early enough to affect the setting of
3790 compiler options, which is what is needed, but I have no means of testing
3791 whether it really works. (The user who reported this had patched the
3792 generated configure script, which of course I cannot do.)
3793
3794 LATER: After change 22 below (new libtool files), the configure script
3795 seems to know about icc (and also ecc). Therefore, I have commented out
3796 this hack in configure.in.
3797
3798 7. Added support for pkg-config (2 patches were sent in).
3799
3800 8. Negated POSIX character classes that used a combination of internal tables
3801 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3802 [[:^ascii]]. Typically, they would match almost any characters. The other
3803 POSIX classes were not broken in this way.
3804
3805 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3806 to find the match, as PCRE was deluded into thinking that the match had to
3807 start at the start point or following a newline. The same bug applied to
3808 patterns with negative forward assertions or any backward assertions
3809 preceding ".*" at the start, unless the pattern required a fixed first
3810 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3811
3812 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3813 starting at the last subject character, bytes beyond the end of the subject
3814 string were read.
3815
3816 11. Renamed the variable "class" as "classbits" to make life easier for C++
3817 users. (Previously there was a macro definition, but it apparently wasn't
3818 enough.)
3819
3820 12. Added the new field "tables" to the extra data so that tables can be passed
3821 in at exec time, or the internal tables can be re-selected. This allows
3822 a compiled regex to be saved and re-used at a later time by a different
3823 program that might have everything at different addresses.
3824
3825 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3826 -R library as well as a -L library.
3827
3828 14. The debugging options of pcretest (-d on the command line or D on a
3829 pattern) showed incorrect output for anything following an extended class
3830 that contained multibyte characters and which was followed by a quantifier.
3831
3832 15. Added optional support for general category Unicode character properties
3833 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3834 support. It adds about 90K to the size of the library. The meanings of the
3835 inbuilt class escapes such as \d and \s have NOT been changed.
3836
3837 16. Updated pcredemo.c to include calls to free() to release the memory for the
3838 compiled pattern.
3839
3840 17. The generated file chartables.c was being created in the source directory
3841 instead of in the building directory. This caused the build to fail if the
3842 source directory was different from the building directory, and was
3843 read-only.
3844
3845 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3846 file. No doubt somebody will tell me if they don't make sense... Also added
3847 Dan Mooney's comments about building on OpenVMS.
3848
3849 19. Added support for partial matching via the PCRE_PARTIAL option for
3850 pcre_exec() and the \P data escape in pcretest.
3851
3852 20. Extended pcretest with 3 new pattern features:
3853
3854 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3855 write the compiled pattern to the file whose name is "rest-of-line".
3856 This is a straight binary dump of the data, with the saved pointer to
3857 the character tables forced to be NULL. The study data, if any, is
3858 written too. After writing, pcretest reads a new pattern.
3859
3860 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3861 compiled pattern from the given file. There must not be any
3862 occurrences of "<" in the file name (pretty unlikely); if there are,
3863 pcretest will instead treat the initial "<" as a pattern delimiter.
3864 After reading in the pattern, pcretest goes on to read data lines as
3865 usual.
3866
3867 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3868 and 16-bit fields in a compiled pattern, to simulate a pattern that
3869 was compiled on a host of opposite endianness.
3870
3871 21. The pcre-exec() function can now cope with patterns that were compiled on
3872 hosts of opposite endianness, with this restriction:
3873
3874 As for any compiled expression that is saved and used later, the tables
3875 pointer field cannot be preserved; the extra_data field in the arguments
3876 to pcre_exec() should be used to pass in a tables address if a value
3877 other than the default internal tables were used at compile time.
3878
3879 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3880 now diagnosed as an error. Previously, most of the time, a negative number
3881 would have been treated as zero, but if in addition "ovector" was passed as
3882 NULL, a crash could occur.
3883
3884 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3885 new versions from the libtool 1.5 distribution (the last one is a copy of
3886 a file called libtool.m4). This seems to have fixed the need to patch
3887 "configure" to support Darwin 1.3 (which I used to do). However, I still
3888 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3889 workstation).
3890
3891 24. Changed the PCRE licence to be the more standard "BSD" licence.
3892
3893
3894 Version 4.5 01-Dec-03
3895 ---------------------
3896
3897 1. There has been some re-arrangement of the code for the match() function so
3898 that it can be compiled in a version that does not call itself recursively.
3899 Instead, it keeps those local variables that need separate instances for
3900 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3901 needs to "recurse". Keeping track of where control must go is done by means
3902 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3903 hide most of the details from the main code, and operates only if
3904 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3905 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3906 operating.
3907
3908 To make it easier for callers to provide specially tailored get/free
3909 functions for this usage, two new functions, pcre_stack_malloc, and
3910 pcre_stack_free, are used. They are always called in strict stacking order,
3911 and the size of block requested is always the same.
3912
3913 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3914 PCRE has been compiled to use the stack or the heap for recursion. The
3915 -C option of pcretest uses this to show which version is compiled.
3916
3917 A new data escape \S, is added to pcretest; it causes the amounts of store
3918 obtained and freed by both kinds of malloc/free at match time to be added
3919 to the output.
3920
3921 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3922 what's available on my current Linux desktop machine.
3923
3924 3. When matching a UTF-8 string, the test for a valid string at the start has
3925 been extended. If start_offset is not zero, PCRE now checks that it points
3926 to a byte that is the start of a UTF-8 character. If not, it returns
3927 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3928 this is necessary because there may be backward assertions in the pattern.
3929 When matching the same subject several times, it may save resources to use
3930 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3931
3932 4. The code for checking the validity of UTF-8 strings has been tightened so
3933 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3934 containing "overlong sequences".
3935
3936 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3937 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3938 should have been "&", but it just so happened that all the cases this let
3939 through by mistake were picked up later in the function.
3940
3941 6. I had used a variable called "isblank" - this is a C99 function, causing
3942 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3943
3944 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3945 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3946 so the version is shown; (c) stop "make test" echoing "./RunTest".
3947
3948 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3949
3950 9. The prototype for memmove() for systems that don't have it was using
3951 size_t, but the inclusion of the header that defines size_t was later. I've
3952 moved the #includes for the C headers earlier to avoid this.
3953
3954 10. Added some adjustments to the code to make it easier to compiler on certain
3955 special systems:
3956
3957 (a) Some "const" qualifiers were missing.
3958 (b) Added the macro EXPORT before all exported functions; by default this
3959 is defined to be empty.
3960 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3961 that it reads its output file name as an argument instead of writing
3962 to the standard output and assuming this can be redirected.
3963
3964 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3965 class containing characters with values greater than 255, PCRE compilation
3966 went into a loop.
3967
3968 12. A recursive reference to a subpattern that was within another subpattern
3969 that had a minimum quantifier of zero caused PCRE to crash. For example,
3970 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3971 recursion. If the recursively-called subpattern itself had a zero repeat,
3972 that was OK.
3973
3974 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3975 buffer into which it was copied (for escape processing) was still set at
3976 1024, so long lines caused crashes.
3977
3978 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3979 "internal error: code overflow...". This applied to any character class
3980 that was followed by a possessive quantifier.
3981
3982 15. Modified the Makefile to add libpcre.la as a prerequisite for
3983 libpcreposix.la because I was told this is needed for a parallel build to
3984 work.
3985
3986 16. If a pattern that contained .* following optional items at the start was
3987 studied, the wrong optimizing data was generated, leading to matching
3988 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3989 matching string must start with a or b or c. The correct conclusion for
3990 this pattern is that a match can start with any character.
3991
3992
3993 Version 4.4 13-Aug-03
3994 ---------------------
3995
3996 1. In UTF-8 mode, a character class containing characters with values between
3997 127 and 255 was not handled correctly if the compiled pattern was studied.
3998 In fixing this, I have also improved the studying algorithm for such
3999 classes (slightly).
4000
4001 2. Three internal functions had redundant arguments passed to them. Removal
4002 might give a very teeny performance improvement.
4003
4004 3. Documentation bug: the value of the capture_top field in a callout is *one
4005 more than* the number of the hightest numbered captured substring.
4006
4007 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4008 in incorrectly linking with a previously installed version. They now link
4009 explicitly with libpcre.la.
4010
4011 5. configure.in no longer needs to recognize Cygwin specially.
4012
4013 6. A problem in pcre.in for Windows platforms is fixed.
4014
4015 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4016 pcretest, it used to include the size of the study block as part of its
4017 output. Unfortunately, the structure contains a field that has a different
4018 size on different hardware architectures. This meant that the tests that
4019 showed this size failed. As the block is currently always of a fixed size,
4020 this information isn't actually particularly useful in pcretest output, so
4021 I have just removed it.
4022
4023 8. Three pre-processor statements accidentally did not start in column 1.
4024 Sadly, there are *still* compilers around that complain, even though
4025 standard C has not required this for well over a decade. Sigh.
4026
4027 9. In pcretest, the code for checking callouts passed small integers in the
4028 callout_data field, which is a void * field. However, some picky compilers
4029 complained about the casts involved for this on 64-bit systems. Now
4030 pcretest passes the address of the small integer instead, which should get
4031 rid of the warnings.
4032
4033 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4034 both compile and run time, and gives an error if an invalid UTF-8 sequence
4035 is found. There is a option for disabling this check in cases where the
4036 string is known to be correct and/or the maximum performance is wanted.
4037
4038 11. In response to a bug report, I changed one line in Makefile.in from
4039
4040 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4041 to
4042 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4043
4044 to look similar to other lines, but I have no way of telling whether this
4045 is the right thing to do, as I do not use Windows. No doubt I'll get told
4046 if it's wrong...
4047
4048
4049 Version 4.3 21-May-03
4050 ---------------------
4051
4052 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4053 Makefile.
4054
4055 2. Some refactoring to improve the quality of the code:
4056
4057 (i) The utf8_table... variables are now declared "const".
4058
4059 (ii) The code for \cx, which used the "case flipping" table to upper case
4060 lower case letters, now just substracts 32. This is ASCII-specific,
4061 but the whole concept of \cx is ASCII-specific, so it seems
4062 reasonable.
4063
4064 (iii) PCRE was using its character types table to recognize decimal and
4065 hexadecimal digits in the pattern. This is silly, because it handles
4066 only 0-9, a-f, and A-F, but the character types table is locale-
4067 specific, which means strange things might happen. A private
4068 table is now used for this - though it costs 256 bytes, a table is
4069 much faster than multiple explicit tests. Of course, the standard
4070 character types table is still used for matching digits in subject
4071 strings against \d.
4072
4073 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4074 ending in _t are). So I've renamed it as ESC_tee.
4075
4076 3. The first argument for regexec() in the POSIX wrapper should have been
4077 defined as "const".
4078
4079 4. Changed pcretest to use malloc() for its buffers so that they can be
4080 Electric Fenced for debugging.
4081
4082 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4083 to read one or more bytes before the start of the subject string. Often this
4084 had no effect on PCRE's behaviour, but in some circumstances it could
4085 provoke a segmentation fault.
4086
4087 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4088 to try to read one or more bytes before the start of the subject string.
4089
4090 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4091 UTF-8 support could misbehave in various ways if the subject string
4092 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4093 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4094 back over UTF-8 characters.)
4095
4096
4097 Version 4.2 14-Apr-03
4098 ---------------------
4099
4100 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4101
4102 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4103 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4104 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4105 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4106 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4107 and BUILD_EXEEXT
4108 Note: automatic setting of the BUILD variables is not yet working
4109 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4110 compile-time but not at link-time
4111 [LINK]: use for linking executables only
4112 make different versions for Windows and non-Windows
4113 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4114 libraries
4115 [LINK_FOR_BUILD]: new variable
4116 [OBJEXT]: use throughout
4117 [EXEEXT]: use throughout
4118 <winshared>: new target
4119 <wininstall>: new target
4120 <dftables.o>: use native compiler
4121 <dftables>: use native linker
4122 <install>: handle Windows platform correctly
4123 <clean>: ditto
4124 <check>: ditto
4125 copy DLL to top builddir before testing
4126
4127 As part of these changes, -no-undefined was removed again. This was reported
4128 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4129 in any case.
4130
4131 3. Some tidies to get rid of compiler warnings:
4132
4133 . In the match_data structure, match_limit was an unsigned long int, whereas
4134 match_call_count was an int. I've made them both unsigned long ints.
4135
4136 . In pcretest the fact that a const uschar * doesn't automatically cast to
4137 a void * provoked a warning.
4138
4139 . Turning on some more compiler warnings threw up some "shadow" variables
4140 and a few more missing casts.
4141
4142 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4143 option, a class that contained a single character with a value between 128
4144 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4145
4146 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4147 option, a class that contained several characters, but with at least one
4148 whose value was between 128 and 255 caused PCRE to crash.
4149
4150
4151 Version 4.1 12-Mar-03
4152 ---------------------
4153
4154 1. Compiling with gcc -pedantic found a couple of places where casts were
4155 needed, and a string in dftables.c that was longer than standard compilers are
4156 required to support.
4157
4158 2. Compiling with Sun's compiler found a few more places where the code could
4159 be tidied up in order to avoid warnings.
4160
4161 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4162 first of these names is deprecated in the latest Autoconf in favour of the name
4163 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4164 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4165 analogy I have changed it to CFLAGS_FOR_BUILD.
4166
4167 4. Added -no-undefined to the linking command in the Makefile, because this is
4168 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4169 linking step for the pcreposix library.
4170
4171 5. PCRE was failing to diagnose the case of two named groups with the same
4172 name.
4173
4174 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4175 literal character that is needed in the subject for a match, and scans along to
4176 ensure that it is present before embarking on the full matching process. This
4177 saves time in cases of nested unlimited repeats that are never going to match.
4178 Problem: the scan can take a lot of time if the subject is very long (e.g.
4179 megabytes), thus penalizing straightforward matches. It is now done only if the
4180 amount of subject to be scanned is less than 1000 bytes.
4181
4182 7. A lesser problem with the same optimization is that it was recording the
4183 first character of an anchored pattern as "needed", thus provoking a search
4184 right along the subject, even when the first match of the pattern was going to
4185 fail. The "needed" character is now not set for anchored patterns, unless it
4186 follows something in the pattern that is of non-fixed length. Thus, it still
4187 fulfils its original purpose of finding quick non-matches in cases of nested
4188 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4189
4190
4191 Version 4.0 17-Feb-03
4192 ---------------------
4193
4194 1. If a comment in an extended regex that started immediately after a meta-item
4195 extended to the end of string, PCRE compiled incorrect data. This could lead to
4196 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4197
4198 2. Moved to autoconf 2.53 and libtool 1.4.2.
4199
4200 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4201 the special perltest8 script is no longer needed - all the tests can be run
4202 from a single perltest script.
4203
4204 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4205 by \s. It has now been removed in PCRE. This means it isn't recognized as
4206 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4207 class [:space:] *does* include VT, thereby creating a mess.
4208
4209 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4210 space and tab.
4211
4212 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4213 its new features into the main test script, reducing the number of scripts.
4214
4215 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4216 were backward compatible, and made the (?i) apply to the whole pattern, as if
4217 /i were given. Now it behaves more logically, and applies the option setting
4218 only to what follows. PCRE has been changed to follow suit. However, if it
4219 finds options settings right at the start of the pattern, it extracts them into
4220 the global options, as before. Thus, they show up in the info data.
4221
4222 8. Added support for the \Q...\E escape sequence. Characters in between are
4223 treated as literals. This is slightly different from Perl in that $ and @ are
4224 also handled as literals inside the quotes. In Perl, they will cause variable
4225 interpolation. Note the following examples:
4226
4227 Pattern PCRE matches Perl matches
4228
4229 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4230 \Qabc\$xyz\E abc\$xyz abc\$xyz
4231 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4232
4233 For compatibility with Perl, \Q...\E sequences are recognized inside character
4234 classes as well as outside them.
4235
4236 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4237 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4238 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4239 signed/unsigned warnings.
4240
4241 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4242 option for pcretest, so I've replaced it by a simple function that does just
4243 that job.
4244
4245 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4246 "pcregrep -".
4247
4248 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4249 Java package. This provides some syntactic sugar for simple cases of what my
4250 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4251 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4252 item, you can use this simplified notation. Note that only makes sense with
4253 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4254 greediness, whatever the setting of the PCRE_UNGREEDY option.
4255
4256 13. A change of greediness default within a pattern was not taking effect at
4257 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4258 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4259 was abstracted outside.
4260
4261 14. PCRE now supports the \G assertion. It is true when the current matching
4262 position is at the start point of the match. This differs from \A when the
4263 starting offset is non-zero. Used with the /g option of pcretest (or similar
4264 code), it works in the same way as it does for Perl's /g option. If all
4265 alternatives of a regex begin with \G, the expression is anchored to the start
4266 match position, and the "anchored" flag is set in the compiled expression.
4267
4268 15. Some bugs concerning the handling of certain option changes within patterns
4269 have been fixed. These applied to options other than (?ims). For example,
4270 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4271 been the other way round. Some of this was related to change 7 above.
4272
4273 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4274 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4275 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4276 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4277
4278 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4279 mode. Unlike ".", it always matches newline, whatever the setting of
4280 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4281 assertions. Perl allows it, but it doesn't (in general) work because it can't
4282 calculate the length of the lookbehind. At least, that's the case for Perl
4283 5.8.0 - I've been told they are going to document that it doesn't work in
4284 future.
4285
4286 18. Added an error diagnosis for escapes that PCRE does not support: these are
4287 \L, \l, \N, \P, \p, \U, \u, and \X.
4288
4289 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4290 reading past the end of the pattern in cases such as /[abcd/.
4291
4292 20. PCRE was getting more memory than necessary for patterns with classes that
4293 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4294
4295 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4296 compiling PCRE for use with Virtual Pascal.
4297
4298 22. Small fix to the Makefile to make it work properly if the build is done
4299 outside the source tree.
4300
4301 23. Added a new extension: a condition to go with recursion. If a conditional
4302 subpattern starts with (?(R) the "true" branch is used if recursion has
4303 happened, whereas the "false" branch is used only at the top level.
4304
4305 24. When there was a very long string of literal characters (over 255 bytes
4306 without UTF support, over 250 bytes with UTF support), the computation of how
4307 much memory was required could be incorrect, leading to segfaults or other
4308 strange effects.
4309
4310 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4311 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4312 there was a subsequent back reference to those brackets. This meant that, for
4313 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4314 possible to check for precisely this case. All we can do is abandon the
4315 optimization if .* occurs inside capturing brackets when there are any back
4316 references whatsoever. (See below for a better fix that came later.)
4317
4318 26. The handling of the optimization for finding the first character of a
4319 non-anchored pattern, and for finding a character that is required later in the
4320 match were failing in some cases. This didn't break the matching; it just
4321 failed to optimize when it could. The way this is done has been re-implemented.
4322
4323 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4324
4325 28. Added a new feature that provides some of the functionality that Perl
4326 provides with (?{...}). The facility is termed a "callout". The way it is done
4327 in PCRE is for the caller to provide an optional function, by setting
4328 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4329 global variable. By default it is unset, which disables all calling out. To get
4330 the function called, the regex must include (?C) at appropriate points. This
4331 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4332 This provides a means of identifying different callout points. When PCRE
4333 reaches such a point in the regex, if pcre_callout has been set, the external
4334 function is called. It is provided with data in a structure called
4335 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4336 matching continues; if it returns a non-zero value, the match at the current
4337 point fails. However, backtracking will occur if possible. [This was changed
4338 later and other features added - see item 49 below.]
4339
4340 29. pcretest is upgraded to test the callout functionality. It provides a
4341 callout function that displays information. By default, it shows the start of
4342 the match and the current position in the text. There are some new data escapes
4343 to vary what happens:
4344
4345 \C+ in addition, show current contents of captured substrings
4346 \C- do not supply a callout function
4347 \C!n return 1 when callout number n is reached
4348 \C!n!m return 1 when callout number n is reached for the mth time
4349
4350 30. If pcregrep was called with the -l option and just a single file name, it
4351 output "<stdin>" if a match was found, instead of the file name.
4352
4353 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4354 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4355 pcre_exec(). This saves a malloc/free per call. The default value of
4356 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4357 when configuring.
4358
4359 32. The default maximum size of a compiled pattern is 64K. There have been a
4360 few cases of people hitting this limit. The code now uses macros to handle the
4361 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4362 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4363 configuring. Tests 2 and 5 work only with 2-byte links because they output
4364 debugging information about compiled patterns.
4365
4366 33. Internal code re-arrangements:
4367
4368 (a) Moved the debugging function for printing out a compiled regex into
4369 its own source file (printint.c) and used #include to pull it into
4370 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4371 separate copies.
4372
4373 (b) Defined the list of op-code names for debugging as a macro in
4374 internal.h so that it is next to the definition of the opcodes.
4375
4376 (c) Defined a table of op-code lengths for simpler skipping along compiled
4377 code. This is again a macro in internal.h so that it is next to the
4378 definition of the opcodes.
4379
4380 34. Added support for recursive calls to individual subpatterns, along the
4381 lines of Robin Houston's patch (but implemented somewhat differently).
4382
4383 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4384 allow it to read and process whole directories in Win32. This code was
4385 contributed by Lionel Fourquaux; it has not been tested by me.
4386
4387 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4388 used to name a group. Names consist of alphanumerics and underscores, and must
4389 be unique. Back references use the syntax (?P=name) and recursive calls use
4390 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4391 numbers. The function pcre_fullinfo() can be used after compilation to extract
4392 a name/number map. There are three relevant calls:
4393
4394 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4395 PCRE_INFO_NAMECOUNT yields the number of entries
4396 PCRE_INFO_NAMETABLE yields a pointer to the map.
4397
4398 The map is a vector of fixed-size entries. The size of each entry depends on
4399 the length of the longest name used. The first two bytes of each entry are the
4400 group number, most significant byte first. There follows the corresponding
4401 name, zero terminated. The names are in alphabetical order.
4402
4403 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4404 case instead of 255. Making it the same both with and without UTF-8 support
4405 means that the same test output works with both.
4406
4407 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4408 calling malloc() with a zero argument.
4409
4410 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4411 optimization. I've improved things by keeping a bitmap of backreferences with
4412 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4413 fact referenced, the optimization can be applied. It is unlikely that a
4414 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4415 the match to follow \n) will appear inside brackets with a number greater than
4416 31, but if it does, any back reference > 31 suppresses the optimization.
4417
4418 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4419 of disabling numbered capturing parentheses. Any opening parenthesis that is
4420 not followed by ? behaves as if it were followed by ?: but named parentheses
4421 can still be used for capturing (and they will acquire numbers in the usual
4422 way).
4423
4424 41. Redesigned the return codes from the match() function into yes/no/error so
4425 that errors can be passed back from deep inside the nested calls. A malloc
4426 failure while inside a recursive subpattern call now causes the
4427 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4428
4429 42. It is now possible to set a limit on the number of times the match()
4430 function is called in a call to pcre_exec(). This facility makes it possible to
4431 limit the amount of recursion and backtracking, though not in a directly
4432 obvious way, because the match() function is used in a number of different
4433 circumstances. The count starts from zero for each position in the subject
4434 string (for non-anchored patterns). The default limit is, for compatibility, a
4435 large number, namely 10 000 000. You can change this in two ways:
4436
4437 (a) When configuring PCRE before making, you can use --with-match-limit=n
4438 to set a default value for the compiled library.
4439
4440 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4441 a different value is set. See 45 below.
4442
4443 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4444
4445 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4446 of things that can be changed at compile time. The first argument specifies
4447 what is wanted and the second points to where the information is to be placed.
4448 The current list of available information is:
4449
4450 PCRE_CONFIG_UTF8
4451
4452 The output is an integer that is set to one if UTF-8 support is available;
4453 otherwise it is set to zero.
4454
4455 PCRE_CONFIG_NEWLINE
4456
4457 The output is an integer that it set to the value of the code that is used for
4458 newline. It is either LF (10) or CR (13).
4459
4460 PCRE_CONFIG_LINK_SIZE
4461
4462 The output is an integer that contains the number of bytes used for internal
4463 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4464
4465 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4466
4467 The output is an integer that contains the threshold above which the POSIX
4468 interface uses malloc() for output vectors. See item 31 above.
4469
4470 PCRE_CONFIG_MATCH_LIMIT
4471
4472 The output is an unsigned integer that contains the default limit of the number
4473 of match() calls in a pcre_exec() execution. See 42 above.
4474
4475 44. pcretest has been upgraded by the addition of the -C option. This causes it
4476 to extract all the available output from the new pcre_config() function, and to
4477 output it. The program then exits immediately.
4478
4479 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4480 order to support additional features. One way would have been to define
4481 pcre_exec2() (for example) with extra arguments, but this would not have been
4482 extensible, and would also have required all calls to the original function to
4483 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4484 is used for passing in "extra" data from pcre_study().
4485
4486 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4487 contains the following fields:
4488
4489 flags a bitmap indicating which of the following fields are set
4490 study_data opaque data from pcre_study()
4491 match_limit a way of specifying a limit on match() calls for a specific
4492 call to pcre_exec()
4493 callout_data data for callouts (see 49 below)
4494
4495 The flag bits are also defined in pcre.h, and are
4496
4497 PCRE_EXTRA_STUDY_DATA
4498 PCRE_EXTRA_MATCH_LIMIT
4499 PCRE_EXTRA_CALLOUT_DATA
4500
4501 The pcre_study() function now returns one of these new pcre_extra blocks, with
4502 the actual study data pointed to by the study_data field, and the
4503 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4504 before. That is, this change is entirely upwards-compatible and requires no
4505 change to existing code.
4506
4507 If you want to pass in additional data to pcre_exec(), you can either place it
4508 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4509 block.
4510
4511 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4512 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4513 times with different match limits, until it finds the minimum value needed for
4514 pcre_exec() to complete. The value is then output. This can be instructive; for
4515 most simple matches the number is quite small, but for pathological cases it
4516 gets very large very quickly.
4517
4518 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4519 returns the size of the data block pointed to by the study_data field in a
4520 pcre_extra block, that is, the value that was passed as the argument to
4521 pcre_malloc() when PCRE was getting memory in which to place the information
4522 created by pcre_study(). The fourth argument should point to a size_t variable.
4523 pcretest has been extended so that this information is shown after a successful
4524 pcre_study() call when information about the compiled regex is being displayed.
4525
4526 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4527 because what follows is always an absolute path. (Later: it turns out that this
4528 is more than cosmetic for MinGW, because it doesn't like empty path
4529 components.)
4530
4531 49. Some changes have been made to the callout feature (see 28 above):
4532
4533 (i) A callout function now has three choices for what it returns:
4534
4535 0 => success, carry on matching
4536 > 0 => failure at this point, but backtrack if possible
4537 < 0 => serious error, return this value from pcre_exec()
4538
4539 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4540 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4541 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4542 use by callout functions. It will never be used by PCRE itself.
4543
4544 (ii) The pcre_extra structure (see 45 above) has a void * field called
4545 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4546 pcre_callout_block structure has a field of the same name. The contents of
4547 the field passed in the pcre_extra structure are passed to the callout
4548 function in the corresponding field in the callout block. This makes it
4549 easier to use the same callout-containing regex from multiple threads. For
4550 testing, the pcretest program has a new data escape
4551
4552 \C*n pass the number n (may be negative) as callout_data
4553
4554 If the callout function in pcretest receives a non-zero value as
4555 callout_data, it returns that value.
4556
4557 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4558 there were some redundant $(CFLAGS) in commands that are now specified as
4559 $(LINK), which already includes $(CFLAGS).
4560
4561 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4562 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4563 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4564 one-byte characters throughout. Note that case-insensitive matching applies
4565 only to characters whose values are less than 256. PCRE doesn't support the
4566 notion of cases for higher-valued characters.
4567
4568 (i) A character class whose characters are all within 0-255 is handled as
4569 a bit map, and the map is inverted for negative classes. Previously, a
4570 character > 255 always failed to match such a class; however it should
4571 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4572
4573 (ii) A negated character class with a single character < 255 is coded as
4574 "not this character" (OP_NOT). This wasn't working properly when the test
4575 character was multibyte, either singly or repeated.
4576
4577 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4578 mode, for example: \x{100}{2,3}.
4579
4580 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4581 singly or repeated) now correctly test multibyte characters. However,
4582 PCRE doesn't recognize any characters with values greater than 255 as
4583 digits, spaces, or word characters. Such characters always match \D, \S,
4584 and \W, and never match \d, \s, or \w.
4585
4586 (v) Classes may now contain characters and character ranges with values
4587 greater than 255. For example: [ab\x{100}-\x{400}].
4588
4589 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4590 PCRE in UTF-8 mode.
4591
4592 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4593 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4594 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4595 value.)
4596
4597 53. The single man page has become too large. I have therefore split it up into
4598 a number of separate man pages. These also give rise to individual HTML pages;
4599 these are now put in a separate directory, and there is an index.html page that
4600 lists them all. Some hyperlinking between the pages has been installed.
4601
4602 54. Added convenience functions for handling named capturing parentheses.
4603
4604 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4605 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4606 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4607 are faulted.
4608
4609 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4610 calling configure. These values are used when compiling the dftables.c program
4611 which is run to generate the source of the default character tables. They
4612 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4613 you will need to set these values.
4614
4615 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4616
4617
4618 Version 3.9 02-Jan-02
4619 ---------------------
4620
4621 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4622
4623 2. If --disable-static was given, the building process failed when trying to
4624 build pcretest and pcregrep. (For some reason it was using libtool to compile
4625 them, which is not right, as they aren't part of the library.)
4626
4627
4628 Version 3.8 18-Dec-01
4629 ---------------------
4630
4631 1. The experimental UTF-8 code was completely screwed up. It was packing the
4632 bytes in the wrong order. How dumb can you get?
4633
4634
4635 Version 3.7 29-Oct-01
4636 ---------------------
4637
4638 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4639 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4640 this didn't happen under Solaris 8, where I normally test things.
4641
4642 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4643 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4644 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4645 this fix an hour or so after the initial 3.7 release.)
4646
4647
4648 Version 3.6 23-Oct-01
4649 ---------------------
4650
4651 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4652 offsets passed as NULL with zero offset count.
4653
4654 2. The config.guess and config.sub files had not been updated when I moved to
4655 the latest autoconf.
4656
4657
4658 Version 3.5 15-Aug-01
4659 ---------------------
4660
4661 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4662 had been forgotten.
4663
4664 2. By using declared but undefined structures, we can avoid using "void"
4665 definitions in pcre.h while keeping the internal definitions of the structures
4666 private.
4667
4668 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4669 user point of view, this means that both static and shared libraries are built
4670 by default, but this can be individually controlled. More of the work of
4671 handling this static/shared cases is now inside libtool instead of PCRE's make
4672 file.
4673
4674 4. The pcretest utility is now installed along with pcregrep because it is
4675 useful for users (to test regexs) and by doing this, it automatically gets
4676 relinked by libtool. The documentation has been turned into a man page, so
4677 there are now .1, .txt, and .html versions in /doc.
4678
4679 5. Upgrades to pcregrep:
4680 (i) Added long-form option names like gnu grep.
4681 (ii) Added --help to list all options with an explanatory phrase.
4682 (iii) Added -r, --recursive to recurse into sub-directories.
4683 (iv) Added -f, --file to read patterns from a file.
4684
4685 6. pcre_exec() was referring to its "code" argument before testing that
4686 argument for NULL (and giving an error if it was NULL).
4687
4688 7. Upgraded Makefile.in to allow for compiling in a different directory from
4689 the source directory.
4690
4691 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4692 options bits, the pointer it was passed was to an int instead of to an unsigned
4693 long int. This mattered only on 64-bit systems.
4694
4695 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4696 generated) instead of pcre.in, which it its source. Also made the same change
4697 in several of the .c files.
4698
4699 10. A new release of gcc defines printf() as a macro, which broke pcretest
4700 because it had an ifdef in the middle of a string argument for printf(). Fixed
4701 by using separate calls to printf().
4702
4703 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4704 script, to force use of CR or LF instead of \n in the source. On non-Unix
4705 systems, the value can be set in config.h.
4706
4707 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4708 absolute limit. Changed the text of the error message to make this clear, and
4709 likewise updated the man page.
4710
4711 13. The limit of 99 on the number of capturing subpatterns has been removed.
4712 The new limit is 65535, which I hope will not be a "real" limit.
4713
4714
4715 Version 3.4 22-Aug-00
4716 ---------------------
4717
4718 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4719
4720 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4721
4722
4723 Version 3.3 01-Aug-00
4724 ---------------------
4725
4726 1. If an octal character was given, but the value was greater than \377, it
4727 was not getting masked to the least significant bits, as documented. This could
4728 lead to crashes in some systems.
4729
4730 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4731 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4732
4733 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4734 These just pass their arguments on to (pcre_free)(), but they are provided
4735 because some uses of PCRE bind it to non-C systems that can call its functions,
4736 but cannot call free() or pcre_free() directly.
4737
4738 4. Add "make test" as a synonym for "make check". Corrected some comments in
4739 the Makefile.
4740
4741 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4742 Makefile.
4743
4744 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4745 command called pgrep for grepping around the active processes.
4746
4747 7. Added the beginnings of support for UTF-8 character strings.
4748
4749 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4750 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4751 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4752 out for the ar command.)
4753
4754
4755 Version 3.2 12-May-00
4756 ---------------------
4757
4758 This is purely a bug fixing release.
4759
4760 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4761 of ZA. This was just one example of several cases that could provoke this bug,
4762 which was introduced by change 9 of version 2.00. The code for breaking
4763 infinite loops after an iteration that matches an empty string was't working
4764 correctly.
4765
4766 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4767 when matched against abbab (for example). After matching an empty string, it
4768 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4769 caused it to match further down the string than it should.
4770
4771 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4772 was there because it doesn't seem to be needed, and it causes trouble on some
4773 systems, as it is not a Standard C header. It has been removed.
4774
4775 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4776 were reported on the Macintosh. The changes were from
4777
4778 while ((c = *(++ptr)) != 0 && c != '\n');
4779 to
4780 while ((c = *(++ptr)) != 0 && c != '\n') ;
4781
4782 Totally extraordinary, but if that's what it takes...
4783
4784 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4785 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4786 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4787 assumes the way PCRE uses memmove() (always moving upwards).
4788
4789 6. PCRE is being used in one environment where strchr() is not available. There
4790 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4791 faster code anyway.
4792
4793
4794 Version 3.1 09-Feb-00
4795 ---------------------
4796
4797 The only change in this release is the fixing of some bugs in Makefile.in for
4798 the "install" target:
4799
4800 (1) It was failing to install pcreposix.h.
4801
4802 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4803
4804
4805 Version 3.0 01-Feb-00
4806 ---------------------
4807
4808 1. Add support for the /+ modifier to perltest (to output $` like it does in
4809 pcretest).
4810
4811 2. Add support for the /g modifier to perltest.
4812
4813 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4814 matches null strings.
4815
4816 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4817 pattern. Perl treats empty patterns specially - it reuses the most recent
4818 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4819 effect.
4820
4821 5. The POSIX interface was broken in that it was just handing over the POSIX
4822 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4823 required a bigger vector, with some working space on the end. This means that
4824 the POSIX wrapper now has to get and free some memory, and copy the results.
4825
4826 6. Added some simple autoconf support, placing the test data and the
4827 documentation in separate directories, re-organizing some of the
4828 information files, and making it build pcre-config (a GNU standard). Also added
4829 libtool support for building PCRE as a shared library, which is now the
4830 default.
4831
4832 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4833 09 are not valid octal constants. Single digits will be used for minor values
4834 less than 10.
4835
4836 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4837 existing programs that set these in the POSIX interface can use PCRE without
4838 modification.
4839
4840 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4841 return all that pcre_info() returns, plus additional data. The pcre_info()
4842 function is retained for compatibility, but is considered to be obsolete.
4843
4844 10. Added experimental recursion feature (?R) to handle one common case that
4845 Perl 5.6 will be able to do with (?p{...}).
4846
4847 11. Added support for POSIX character classes like [:alpha:], which Perl is
4848 adopting.
4849
4850
4851 Version 2.08 31-Aug-99
4852 ----------------------
4853
4854 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4855 trying to match at the startoffset position, but instead was moving forward to
4856 the next newline as if a previous match had failed.
4857
4858 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4859 and could get into a loop if a null string was matched other than at the start
4860 of the subject.
4861
4862 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4863 be distinguished at compile time, and for completeness also added PCRE_DATE.
4864
4865 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4866 in GnuWin32 environments.
4867
4868
4869 Version 2.07 29-Jul-99
4870 ----------------------
4871
4872 1. The documentation is now supplied in plain text form and HTML as well as in
4873 the form of man page sources.
4874
4875 2. C++ compilers don't like assigning (void *) values to other pointer types.
4876 In particular this affects malloc(). Although there is no problem in Standard
4877 C, I've put in casts to keep C++ compilers happy.
4878
4879 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4880 should be (const char *).
4881
4882 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4883 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4884 However, I haven't made this a standard facility. The documentation doesn't
4885 mention it, and the Makefile doesn't support it.
4886
4887 5. The Makefile now contains an "install" target, with editable destinations at
4888 the top of the file. The pcretest program is not installed.
4889
4890 6. pgrep -V now gives the PCRE version number and date.
4891
4892 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4893 causing the entire string to be ignored, instead of just the last character.
4894
4895 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4896 non-matching string, it can take a very, very long time, even for strings of
4897 quite modest length, because of the nested recursion. PCRE now does better in
4898 some of these cases. It does this by remembering the last required literal
4899 character in the pattern, and pre-searching the subject to ensure it is present
4900 before running the real match. In other words, it applies a heuristic to detect
4901 some types of certain failure quickly, and in the above example, if presented
4902 with a string that has no trailing " it gives "no match" very quickly.
4903
4904 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4905 other alternatives are tried instead.
4906
4907
4908 Version 2.06 09-Jun-99
4909 ----------------------
4910
4911 1. Change pcretest's output for amount of store used to show just the code
4912 space, because the remainder (the data block) varies in size between 32-bit and
4913 64-bit systems.
4914
4915 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4916 start matching at. This allows lookbehinds to work when searching for multiple
4917 occurrences in a string.
4918
4919 3. Added additional options to pcretest for testing multiple occurrences:
4920
4921 /+ outputs the rest of the string that follows a match
4922 /g loops for multiple occurrences, using the new startoffset argument
4923 /G loops for multiple occurrences by passing an incremented pointer
4924
4925 4. PCRE wasn't doing the "first character" optimization for patterns starting
4926 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4927 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4928 the letter 'x'. On long subject strings, this gives a significant speed-up.
4929
4930
4931 Version 2.05 21-Apr-99
4932 ----------------------
4933
4934 1. Changed the type of magic_number from int to long int so that it works
4935 properly on 16-bit systems.
4936
4937 2. Fixed a bug which caused patterns starting with .* not to work correctly
4938 when the subject string contained newline characters. PCRE was assuming
4939 anchoring for such patterns in all cases, which is not correct because .* will
4940 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4941 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4942 must be retried after every newline in the subject.
4943
4944
4945 Version 2.04 18-Feb-99
4946 ----------------------
4947
4948 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4949 computation of the store needed to hold the pattern was incorrect (too large).
4950 If such patterns were nested a few deep, this could multiply and become a real
4951 problem.
4952
4953 2. Added /M option to pcretest to show the memory requirement of a specific
4954 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4955
4956 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4957 compiled in such a way that the backtracking after subsequent failure was
4958 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4959 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4960
4961
4962 Version 2.03 02-Feb-99
4963 ----------------------
4964
4965 1. Fixed typo and small mistake in man page.
4966
4967 2. Added 4th condition (GPL supersedes if conflict) and created separate
4968 LICENCE file containing the conditions.
4969
4970 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4971 Perl, that is the internal \ allows the delimiter to be included in the
4972 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4973 the final delimiter, add \ to the end of the pattern (to test the error).
4974
4975 4. Added the convenience functions for extracting substrings after a successful
4976 match. Updated pcretest to make it able to test these functions.
4977
4978
4979 Version 2.02 14-Jan-99
4980 ----------------------
4981
4982 1. Initialized the working variables associated with each extraction so that
4983 their saving and restoring doesn't refer to uninitialized store.
4984
4985 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4986 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4987 fix the problem.
4988
4989 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4990 calls, and wasn't printing the correct value for compiling calls. Increased the
4991 default value of LOOPREPEAT, and the number of significant figures in the
4992 times.
4993
4994 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4995
4996 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4997 a building problem on Windows NT with a FAT file system.
4998
4999
5000 Version 2.01 21-Oct-98
5001 ----------------------
5002
5003 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5004 to character tables built by pcre_maketables() in the current locale. If NULL
5005 is passed, the default tables are used.
5006
5007
5008 Version 2.00 24-Sep-98
5009 ----------------------
5010
5011 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5012 it any more.
5013
5014 2. Allow quantification of (?>) groups, and make it work correctly.
5015
5016 3. The first character computation wasn't working for (?>) groups.
5017
5018 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5019 end of the subject) and add 5.005's \z, which really does match only at the
5020 very end of the subject.
5021
5022 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5023
5024 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5025 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5026 localized options. All options to pcre_study() were also removed.
5027
5028 7. Add other new features from 5.005:
5029
5030 $(?<= positive lookbehind
5031 $(?<! negative lookbehind
5032 (?imsx-imsx) added the unsetting capability
5033 such a setting is global if at outer level; local otherwise
5034 (?imsx-imsx:) non-capturing groups with option setting
5035 (?(cond)re|re) conditional pattern matching
5036
5037 A backreference to itself in a repeated group matches the previous
5038 captured string.
5039
5040 8. General tidying up of studying (both automatic and via "study")
5041 consequential on the addition of new assertions.
5042
5043 9. As in 5.005, unlimited repeated groups that could match an empty substring
5044 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5045 runtime if any iteration does actually match an empty substring.
5046
5047 10. Include the RunTest script in the distribution.
5048
5049 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5050 discrepancies, some of which were old and were also with respect to 5.004. They
5051 have now been fixed.
5052
5053
5054 Version 1.09 28-Apr-98
5055 ----------------------
5056
5057 1. A negated single character class followed by a quantifier with a minimum
5058 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5059 program crashes, or just wrong answers. This did not apply to negated classes
5060 containing more than one character, or to minima other than one.
5061
5062
5063 Version 1.08 27-Mar-98
5064 ----------------------
5065
5066 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5067
5068 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5069 latter must appear before anything that relies on it in the pattern.
5070
5071
5072 Version 1.07 16-Feb-98
5073 ----------------------
5074
5075 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5076 repeat of a potentially empty string).
5077
5078
5079 Version 1.06 23-Jan-98
5080 ----------------------
5081
5082 1. Added Markus Oberhumer's little patches for C++.
5083
5084 2. Literal strings longer than 255 characters were broken.
5085
5086
5087 Version 1.05 23-Dec-97
5088 ----------------------
5089
5090 1. Negated character classes containing more than one character were failing if
5091 PCRE_CASELESS was set at run time.
5092
5093
5094 Version 1.04 19-Dec-97
5095 ----------------------
5096
5097 1. Corrected the man page, where some "const" qualifiers had been omitted.
5098
5099 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5100 input syntax.
5101
5102 3. Fixed memory leak which occurred when a regex with back references was
5103 matched with an offsets vector that wasn't big enough. The temporary memory
5104 that is used in this case wasn't being freed if the match failed.
5105
5106 4. Tidied pcretest to ensure it frees memory that it gets.
5107
5108 5. Temporary memory was being obtained in the case where the passed offsets
5109 vector was exactly big enough.
5110
5111 6. Corrected definition of offsetof() from change 5 below.
5112
5113 7. I had screwed up change 6 below and broken the rules for the use of
5114 setjmp(). Now fixed.
5115
5116
5117 Version 1.03 18-Dec-97
5118 ----------------------
5119
5120 1. A erroneous regex with a missing opening parenthesis was correctly
5121 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5122 on some systems.
5123
5124 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5125 it was reported that one broken compiler failed on the former because "code" is
5126 also an independent variable.
5127
5128 3. The erroneous regex a[]b caused an array overrun reference.
5129
5130 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5131 fail on data ending with that character. (It was going on too far, and checking
5132 the next character, typically a binary zero.) This was specific to the
5133 optimized code for single-character negative classes.
5134
5135 5. Added a contributed patch from the TIN world which does the following:
5136
5137 + Add an undef for memmove, in case the the system defines a macro for it.
5138
5139 + Add a definition of offsetof(), in case there isn't one. (I don't know
5140 the reason behind this - offsetof() is part of the ANSI standard - but
5141 it does no harm).
5142
5143 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5144 most of the places where whitespace preceded '#'. I have given up and
5145 allowed the remaining 2 cases to be at the margin.
5146
5147 + Rename some variables in pcre to eliminate shadowing. This seems very
5148 pedantic, but does no harm, of course.
5149
5150 6. Moved the call to setjmp() into its own function, to get rid of warnings
5151 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5152
5153 7. Constructs such as \d{8,} were compiling into the equivalent of
5154 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5155 outcome, but in this particular case used more store than had been allocated,
5156 which caused the bug to be discovered because it threw up an internal error.
5157
5158 8. The debugging code in both pcre and pcretest for outputting the compiled
5159 form of a regex was going wrong in the case of back references followed by
5160 curly-bracketed repeats.
5161
5162
5163 Version 1.02 12-Dec-97
5164 ----------------------
5165
5166 1. Typos in pcre.3 and comments in the source fixed.
5167
5168 2. Applied a contributed patch to get rid of places where it used to remove
5169 'const' from variables, and fixed some signed/unsigned and uninitialized
5170 variable warnings.
5171
5172 3. Added the "runtest" target to Makefile.
5173
5174 4. Set default compiler flag to -O2 rather than just -O.
5175
5176
5177 Version 1.01 19-Nov-97
5178 ----------------------
5179
5180 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5181 like /([ab]*)*/, that is, for classes with more than one character in them.
5182
5183 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5184 as /((?>a*))*/ (a PCRE_EXTRA facility).
5185
5186
5187 Version 1.00 18-Nov-97
5188 ----------------------
5189
5190 1. Added compile-time macros to support systems such as SunOS4 which don't have
5191 memmove() or strerror() but have other things that can be used instead.
5192
5193 2. Arranged that "make clean" removes the executables.
5194
5195
5196 Version 0.99 27-Oct-97
5197 ----------------------
5198
5199 1. Fixed bug in code for optimizing classes with only one character. It was
5200 initializing a 32-byte map regardless, which could cause it to run off the end
5201 of the memory it had got.
5202
5203 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5204
5205
5206 Version 0.98 22-Oct-97
5207 ----------------------
5208
5209 1. Fixed bug in code for handling temporary memory usage when there are more
5210 back references than supplied space in the ovector. This could cause segfaults.
5211
5212
5213 Version 0.97 21-Oct-97
5214 ----------------------
5215
5216 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5217
5218 2. Optimized negated single characters not to use a bit map.
5219
5220 3. Brought error texts together as macro definitions; clarified some of them;
5221 fixed one that was wrong - it said "range out of order" when it meant "invalid
5222 escape sequence".
5223
5224 4. Changed some char * arguments to const char *.
5225
5226 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5227
5228 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5229 pcretest.
5230
5231
5232 Version 0.96 16-Oct-97
5233 ----------------------
5234
5235 1. Added a simple "pgrep" utility to the distribution.
5236
5237 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5238 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5239 where "ddd" means "one or more decimal digits".
5240
5241 3. Fixed serious bug. If a pattern had a back reference, but the call to
5242 pcre_exec() didn't supply a large enough ovector to record the related
5243 identifying subpattern, the match always failed. PCRE now remembers the number
5244 of the largest back reference, and gets some temporary memory in which to save
5245 the offsets during matching if necessary, in order to ensure that
5246 backreferences always work.
5247
5248 4. Increased the compatibility with Perl in a number of ways:
5249
5250 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5251 to request this handling. The option can be set at compile or exec time.
5252
5253 (b) $ matches before a terminating newline by default; an option
5254 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5255 mode). The option can be set at compile or exec time.
5256
5257 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5258 the same as Perl's. If the decimal number it represents is less than 10
5259 or there aren't that many previous left capturing parentheses, an octal
5260 escape is read. Inside a character class, it's always an octal escape,
5261 even if it is a single digit.
5262
5263 (d) An escaped but undefined alphabetic character is taken as a literal,
5264 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5265 escapes.
5266
5267 (e) {0} is now permitted. (The previous item is removed from the compiled
5268 pattern).
5269
5270 5. Changed all the names of code files so that the basic parts are no longer
5271 than 10 characters, and abolished the teeny "globals.c" file.
5272
5273 6. Changed the handling of character classes; they are now done with a 32-byte
5274 bit map always.
5275
5276 7. Added the -d and /D options to pcretest to make it possible to look at the
5277 internals of compilation without having to recompile pcre.
5278
5279
5280 Version 0.95 23-Sep-97
5281 ----------------------
5282
5283 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5284 \x20 at the start of a run of normal characters. These were being treated as
5285 real characters, instead of the source characters being re-checked.
5286
5287
5288 Version 0.94 18-Sep-97
5289 ----------------------
5290
5291 1. The functions are now thread-safe, with the caveat that the global variables
5292 containing pointers to malloc() and free() or alternative functions are the
5293 same for all threads.
5294
5295 2. Get pcre_study() to generate a bitmap of initial characters for non-
5296 anchored patterns when this is possible, and use it if passed to pcre_exec().
5297
5298
5299 Version 0.93 15-Sep-97
5300 ----------------------
5301
5302 1. /(b)|(:+)/ was computing an incorrect first character.
5303
5304 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5305 but not actually doing anything yet.
5306
5307 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5308 as Perl does (e.g. [-az] or [az-]).
5309
5310 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5311 all possible positions.
5312
5313 5. Split up into different modules to avoid including unneeded functions in a
5314 compiled binary. However, compile and exec are still in one module. The "study"
5315 function is split off.
5316
5317 6. The character tables are now in a separate module whose source is generated
5318 by an auxiliary program - but can then be edited by hand if required. There are
5319 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5320 toupper() in the code.
5321
5322 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5323 make them global. Abolish the function for setting them, as the caller can now
5324 set them directly.
5325
5326
5327 Version 0.92 11-Sep-97
5328 ----------------------
5329
5330 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5331 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5332
5333 2. Caseless matching was not working in character classes if the characters in
5334 the pattern were in upper case.
5335
5336 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5337
5338 4. Make PCRE_ANCHORED public and accept as a compile option.
5339
5340 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5341 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5342 pass them.
5343
5344 6. Give an error if bad option bits passed at compile or run time.
5345
5346 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5347 pcretest to cause it to pass that flag.
5348
5349 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5350 options, and the first character, if set.
5351
5352 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5353
5354
5355 Version 0.91 10-Sep-97
5356 ----------------------
5357
5358 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5359 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5360
5361 2. PCRE was looping on encountering an indefinitely repeated back reference to
5362 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5363 Perl does - treats the match as successful.
5364
5365 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5