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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1438 - (show annotations)
Fri Jan 10 16:13:10 2014 UTC (5 years, 10 months ago) by ph10
File size: 261001 byte(s)
Fix caseless character class bug for characters within a range that have more 
than one other case.
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 2. The auto-possessification of character sets were improved: a normal
12 and an extended character set can be compared now. Furthermore
13 the JIT compiler optimizes more character set checks.
14
15 3. Got rid of some compiler warnings for potentially uninitialized variables
16 that show up only when compiled with -O2.
17
18 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
19 of a match later then the end of the match. The pcretest program was not
20 handling the case sensibly - it was outputting from the start to the next
21 binary zero. It now reports this situation in a message, and outputs the
22 text from the end to the start.
23
24 5. Fast forward search is improved in JIT. Instead of the first three
25 characters, any three characters with fixed position can be searched.
26 Search order: first, last, middle.
27
28 6. Improve character range checks in JIT. Characters are read by an inprecise
29 function now, which returns with an unknown value if the character code is
30 above a certain treshold (e.g: 256). The only limitation is that the value
31 must be bigger than the treshold as well. This function is useful, when
32 the characters above the treshold are handled in the same way.
33
34 7. The macros whose names start with RAWUCHAR are placeholders for a future
35 mode in which only the bottom 21 bits of 32-bit data items are used. To
36 make this more memorable for those maintaining the code, the names have
37 been changed to start with UCHAR21, and an extensive comment has been added
38 to their definition.
39
40 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
41 to the export list in Makefile.am (they were accidentally omitted from the
42 8.34 tarball).
43
44 9. The informational output from pcretest used the phrase "starting byte set"
45 which is inappropriate for the 16-bit and 32-bit libraries. As the output
46 for "first char" and "need char" really means "non-UTF-char", I've changed
47 "byte" to "char", and slightly reworded the output. The documentation about
48 these values has also been (I hope) clarified.
49
50 10. Another JIT related optimization: use table jumps for selecting the correct
51 backtracking path, when more than four alternatives are present inside a
52 bracket.
53
54 11. Empty match is not possible, when the minimum length is greater than zero,
55 and there is no \K in the pattern. Remove these unnecessary checks form JIT.
56
57 12. In a caseless character class with UCP support, when a character with more
58 than one alternative case was not the first character of a range, not all
59 the alternative cases were added to the class. For example, s and \x{17f}
60 are both alternative cases for S: the class [RST] was handled correctly,
61 but [R-T] was not.
62
63
64 Version 8.34 15-December-2013
65 -----------------------------
66
67 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
68 executable memory. Patch inspired by Carsten Klein.
69
70 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
71 this macro is never tested and has no effect, because the work to support
72 coverage involves only compiling and linking options and special targets in
73 the Makefile. The comment in config.h implied that defining the macro would
74 enable coverage support, which is totally false. There was also support for
75 setting this macro in the CMake files (my fault, I just copied it from
76 configure). SUPPORT_GCOV has now been removed.
77
78 3. Make a small performance improvement in strlen16() and strlen32() in
79 pcretest.
80
81 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
82 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
83 these cases). There was also one in pcretest.c.
84
85 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
86
87 6. In UTF mode, the code for checking whether a group could match an empty
88 string (which is used for indefinitely repeated groups to allow for
89 breaking an infinite loop) was broken when the group contained a repeated
90 negated single-character class with a character that occupied more than one
91 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
92 UTF-8 mode). The effect was undefined: the group might or might not be
93 deemed as matching an empty string, or the program might have crashed.
94
95 7. The code for checking whether a group could match an empty string was not
96 recognizing that \h, \H, \v, \V, and \R must match a character.
97
98 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
99 an empty string. If it can, pcretest shows this in its information output.
100
101 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
102 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
103 matched by pcre_exec() without using JIT:
104
105 (a) If the rest of the pattern did not match after a maximal run of
106 grapheme clusters, the code for backing up to try with fewer of them
107 did not always back up over a full grapheme when characters that do not
108 have the modifier quality were involved, e.g. Hangul syllables.
109
110 (b) If the match point in a subject started with modifier character, and
111 there was no match, the code could incorrectly back up beyond the match
112 point, and potentially beyond the first character in the subject,
113 leading to a segfault or an incorrect match result.
114
115 10. A conditional group with an assertion condition could lead to PCRE
116 recording an incorrect first data item for a match if no other first data
117 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
118 first data item, and therefore matched "ca" after "c" instead of at the
119 start.
120
121 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
122 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
123
124 12. The source of pcregrep now includes z/OS-specific code so that it can be
125 compiled for z/OS as part of the special z/OS distribution.
126
127 13. Added the -T and -TM options to pcretest.
128
129 14. The code in pcre_compile.c for creating the table of named capturing groups
130 has been refactored. Instead of creating the table dynamically during the
131 actual compiling pass, the information is remembered during the pre-compile
132 pass (on the stack unless there are more than 20 named groups, in which
133 case malloc() is used) and the whole table is created before the actual
134 compile happens. This has simplified the code (it is now nearly 150 lines
135 shorter) and prepared the way for better handling of references to groups
136 with duplicate names.
137
138 15. A back reference to a named subpattern when there is more than one of the
139 same name now checks them in the order in which they appear in the pattern.
140 The first one that is set is used for the reference. Previously only the
141 first one was inspected. This change makes PCRE more compatible with Perl.
142
143 16. Unicode character properties were updated from Unicode 6.3.0.
144
145 17. The compile-time code for auto-possessification has been refactored, based
146 on a patch by Zoltan Herczeg. It now happens after instead of during
147 compilation. The code is cleaner, and more cases are handled. The option
148 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
149 options in pcretest are provided to set it. It can also be set by
150 (*NO_AUTO_POSSESS) at the start of a pattern.
151
152 18. The character VT has been added to the default ("C" locale) set of
153 characters that match \s and are generally treated as white space,
154 following this same change in Perl 5.18. There is now no difference between
155 "Perl space" and "POSIX space". Whether VT is treated as white space in
156 other locales depends on the locale.
157
158 19. The code for checking named groups as conditions, either for being set or
159 for being recursed, has been refactored (this is related to 14 and 15
160 above). Processing unduplicated named groups should now be as fast at
161 numerical groups, and processing duplicated groups should be faster than
162 before.
163
164 20. Two patches to the CMake build system, by Alexander Barkov:
165
166 (1) Replace the "source" command by "." in CMakeLists.txt because
167 "source" is a bash-ism.
168
169 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
170 without these the CMake build does not work on Solaris.
171
172 21. Perl has changed its handling of \8 and \9. If there is no previously
173 encountered capturing group of those numbers, they are treated as the
174 literal characters 8 and 9 instead of a binary zero followed by the
175 literals. PCRE now does the same.
176
177 22. Following Perl, added \o{} to specify codepoints in octal, making it
178 possible to specify values greater than 0777 and also making them
179 unambiguous.
180
181 23. Perl now gives an error for missing closing braces after \x{... instead of
182 treating the string as literal. PCRE now does the same.
183
184 24. RunTest used to grumble if an inappropriate test was selected explicitly,
185 but just skip it when running all tests. This make it awkward to run ranges
186 of tests when one of them was inappropriate. Now it just skips any
187 inappropriate tests, as it always did when running all tests.
188
189 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
190 character types such as \d or \w, too many callouts were inserted, and the
191 data that they returned was rubbish.
192
193 26. In UCP mode, \s was not matching two of the characters that Perl matches,
194 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
195 were matched by \h. The code has now been refactored so that the lists of
196 the horizontal and vertical whitespace characters used for \h and \v (which
197 are defined only in one place) are now also used for \s.
198
199 27. Add JIT support for the 64 bit TileGX architecture.
200 Patch by Jiong Wang (Tilera Corporation).
201
202 28. Possessive quantifiers for classes (both explicit and automatically
203 generated) now use special opcodes instead of wrapping in ONCE brackets.
204
205 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
206 (because it's meaningless), this was not happening when PCRE_CASELESS was
207 set. Not wrong, but inefficient.
208
209 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
210 properties for \w, \d, etc) is present in a test regex. Otherwise if the
211 test contains no characters greater than 255, Perl doesn't realise it
212 should be using Unicode semantics.
213
214 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
215 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
216 does in Unicode mode.
217
218 32. Added the "forbid" facility to pcretest so that putting tests into the
219 wrong test files can sometimes be quickly detected.
220
221 33. There is now a limit (default 250) on the depth of nesting of parentheses.
222 This limit is imposed to control the amount of system stack used at compile
223 time. It can be changed at build time by --with-parens-nest-limit=xxx or
224 the equivalent in CMake.
225
226 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
227 errors. Perl warns for these when in warning mode, but PCRE has no facility
228 for giving warnings.
229
230 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
231 However, this was not working for (?!) because it is optimized to (*FAIL),
232 for which PCRE does not allow quantifiers. The optimization is now disabled
233 when a quantifier follows (?!). I can't see any use for this, but it makes
234 things uniform.
235
236 36. Perl no longer allows group names to start with digits, so I have made this
237 change also in PCRE. It simplifies the code a bit.
238
239 37. In extended mode, Perl ignores spaces before a + that indicates a
240 possessive quantifier. PCRE allowed a space before the quantifier, but not
241 before the possessive +. It now does.
242
243 38. The use of \K (reset reported match start) within a repeated possessive
244 group such as (a\Kb)*+ was not working.
245
246 40. Document that the same character tables must be used at compile time and
247 run time, and that the facility to pass tables to pcre_exec() and
248 pcre_dfa_exec() is for use only with saved/restored patterns.
249
250 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
251 features for Builds with MSVC:
252
253 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
254 builds). The .rc files can be used to set FileDescription and many other
255 attributes.
256
257 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
258 This allows higher-level build scripts which want .pdb files to avoid
259 hard-coding the exact files needed."
260
261 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
262 mean "start of word" and "end of word", respectively, as a transition aid.
263
264 43. A minimizing repeat of a class containing codepoints greater than 255 in
265 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
266 compiled to use the heap for recursion.
267
268 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
269 is configured.
270
271
272 Version 8.33 28-May-2013
273 ------------------------
274
275 1. Added 'U' to some constants that are compared to unsigned integers, to
276 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
277 variables that are added to signed variables, to ensure the result is
278 signed and can be negated.
279
280 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
281 CMake config files.
282
283 3. Revise the creation of config.h.generic so that all boolean macros are
284 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
285 overriding via -D on the command line possible.
286
287 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
288 to unsigned int is reported to make a quite noticeable speed difference in
289 a specific Windows environment. Testing on Linux did also appear to show
290 some benefit (and it is clearly not harmful). Also fixed the definition of
291 Xop which should be unsigned.
292
293 5. Related to (4), changing the definition of the intermediate variable cc
294 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
295 improvements.
296
297 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
298 unnecessary spaces.
299
300 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
301 and later.
302
303 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
304
305 9. Optimizing fast_forward_start_bits in JIT.
306
307 10. Adding support for callouts in JIT, and fixing some issues revealed
308 during this work. Namely:
309
310 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
311
312 (b) Minimum length was not checked before the matching is started.
313
314 11. The value of capture_last that is passed to callouts was incorrect in some
315 cases when there was a capture on one path that was subsequently abandoned
316 after a backtrack. Also, the capture_last value is now reset after a
317 recursion, since all captures are also reset in this case.
318
319 12. The interpreter no longer returns the "too many substrings" error in the
320 case when an overflowing capture is in a branch that is subsequently
321 abandoned after a backtrack.
322
323 13. In the pathological case when an offset vector of size 2 is used, pcretest
324 now prints out the matched string after a yield of 0 or 1.
325
326 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
327 Only supported by the JIT compiler at the moment.
328
329 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
330
331 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
332 offset of the starting point of the matching process, provided the offsets
333 vector is large enough.
334
335 17. The \A escape now records a lookbehind value of 1, though its execution
336 does not actually inspect the previous character. This is to ensure that,
337 in partial multi-segment matching, at least one character from the old
338 segment is retained when a new segment is processed. Otherwise, if there
339 are no lookbehinds in the pattern, \A might match incorrectly at the start
340 of a new segment.
341
342 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
343
344 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
345 gives some modest performance improvement in 8-bit mode.
346
347 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
348 be expressed in certain programming languages using Universal Character
349 Names.
350
351 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
352 which points out that "non characters" are not "characters that may not
353 appear in Unicode strings" but rather "characters that are reserved for
354 internal use and have only local meaning".
355
356 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
357 there was a conditional group that depended on an assertion, if the
358 assertion was false, the callout that immediately followed the alternation
359 in the condition was skipped when pcre_exec() was used for matching.
360
361 23. Allow an explicit callout to be inserted before an assertion that is the
362 condition for a conditional group, for compatibility with automatic
363 callouts, which always insert a callout at this point.
364
365 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
366 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
367
368 25. (*PRUNE) is now supported by the JIT compiler.
369
370 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
371
372 27. Fix the case where there are two or more SKIPs with arguments that may be
373 ignored.
374
375 28. (*SKIP) is now supported by the JIT compiler.
376
377 29. (*THEN) is now supported by the JIT compiler.
378
379 30. Update RunTest with additional test selector options.
380
381 31. The way PCRE handles backtracking verbs has been changed in two ways.
382
383 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
384 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
385 backtracking. In some cases this makes it more Perl-compatible, but Perl's
386 rather obscure rules do not always do the same thing.
387
388 (2) Previously, backtracking verbs were confined within assertions. This is
389 no longer the case for positive assertions, except for (*ACCEPT). Again,
390 this sometimes improves Perl compatibility, and sometimes does not.
391
392 32. A number of tests that were in test 2 because Perl did things differently
393 have been moved to test 1, because either Perl or PCRE has changed, and
394 these tests are now compatible.
395
396 32. Backtracking control verbs are now handled in the same way in JIT and
397 interpreter.
398
399 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
400 contained a forward subroutine reference caused a compile error.
401
402 34. Auto-detect and optimize limited repetitions in JIT.
403
404 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
405 blocking (*UTF) etc.
406
407 36. In the interpreter, maximizing pattern repetitions for characters and
408 character types now use tail recursion, which reduces stack usage.
409
410 37. The value of the max lookbehind was not correctly preserved if a compiled
411 and saved regex was reloaded on a host of different endianness.
412
413 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
414 of the compiled pattern block, expand the flags field from 16 to 32 bits
415 because it was almost full.
416
417 39. Try madvise first before posix_madvise.
418
419 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
420 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
421 original change locked out all matches of empty strings. This has been
422 changed so that one match of an empty string per line is recognized.
423 Subsequent searches on the same line (for colouring or for --only-matching,
424 for example) do not recognize empty strings.
425
426 41. Applied a user patch to fix a number of spelling mistakes in comments.
427
428 42. Data lines longer than 65536 caused pcretest to crash.
429
430 43. Clarified the data type for length and startoffset arguments for pcre_exec
431 and pcre_dfa_exec in the function-specific man pages, where they were
432 explicitly stated to be in bytes, never having been updated. I also added
433 some clarification to the pcreapi man page.
434
435 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
436 a segmentation fault.
437
438
439 Version 8.32 30-November-2012
440 -----------------------------
441
442 1. Improved JIT compiler optimizations for first character search and single
443 character iterators.
444
445 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
446 Patch by Daniel Richard G.
447
448 3. Single character iterator optimizations in the JIT compiler.
449
450 4. Improved JIT compiler optimizations for character ranges.
451
452 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
453 Reported by Giuseppe D'Angelo.
454
455 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
456 of a line, was being set incorrectly in cases where .* appeared inside
457 atomic brackets at the start of a pattern, or where there was a subsequent
458 *PRUNE or *SKIP.
459
460 7. Improved instruction cache flush for POWER/PowerPC.
461 Patch by Daniel Richard G.
462
463 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
464 grep:
465
466 (a) There is now no limit to the number of patterns to be matched.
467
468 (b) An error is given if a pattern is too long.
469
470 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
471 are now supported.
472
473 (d) --exclude-from and --include-from (multiple use) have been added.
474
475 (e) Exclusions and inclusions now apply to all files and directories, not
476 just to those obtained from scanning a directory recursively.
477
478 (f) Multiple uses of -f and --file-list are now supported.
479
480 (g) In a Windows environment, the default for -d has been changed from
481 "read" (the GNU grep default) to "skip", because otherwise the presence
482 of a directory in the file list provokes an error.
483
484 (h) The documentation has been revised and clarified in places.
485
486 9. Improve the matching speed of capturing brackets.
487
488 10. Changed the meaning of \X so that it now matches a Unicode extended
489 grapheme cluster.
490
491 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
492 out POSIX threads when JIT support is configured.
493
494 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
495
496 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
497 structure could go wrong in environments where size_t is not the same size
498 as int.
499
500 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
501 set.
502
503 15. The EBCDIC support had decayed; later updates to the code had included
504 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
505 general tidy up of EBCDIC-related issues, and the documentation was also
506 not quite right. There is now a test that can be run on ASCII systems to
507 check some of the EBCDIC-related things (but is it not a full test).
508
509 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
510 in a small tidy to the code.
511
512 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
513
514 18. If the --only-matching (-o) option in pcregrep is specified multiple
515 times, each one causes appropriate output. For example, -o1 -o2 outputs the
516 substrings matched by the 1st and 2nd capturing parentheses. A separating
517 string can be specified by --om-separator (default empty).
518
519 19. Improving the first n character searches.
520
521 20. Turn case lists for horizontal and vertical white space into macros so that
522 they are defined only once.
523
524 21. This set of changes together give more compatible Unicode case-folding
525 behaviour for characters that have more than one other case when UCP
526 support is available.
527
528 (a) The Unicode property table now has offsets into a new table of sets of
529 three or more characters that are case-equivalent. The MultiStage2.py
530 script that generates these tables (the pcre_ucd.c file) now scans
531 CaseFolding.txt instead of UnicodeData.txt for character case
532 information.
533
534 (b) The code for adding characters or ranges of characters to a character
535 class has been abstracted into a generalized function that also handles
536 case-independence. In UTF-mode with UCP support, this uses the new data
537 to handle characters with more than one other case.
538
539 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
540 whose other case is greater than 256 are now correctly matched
541 caselessly. Previously, the high codepoint matched the low one, but not
542 vice versa.
543
544 (d) The processing of \h, \H, \v, and \ in character classes now makes use
545 of the new class addition function, using character lists defined as
546 macros alongside the case definitions of 20 above.
547
548 (e) Caseless back references now work with characters that have more than
549 one other case.
550
551 (f) General caseless matching of characters with more than one other case
552 is supported.
553
554 22. Unicode character properties were updated from Unicode 6.2.0
555
556 23. Improved CMake support under Windows. Patch by Daniel Richard G.
557
558 24. Add support for 32-bit character strings, and UTF-32
559
560 25. Major JIT compiler update (code refactoring and bugfixing).
561 Experimental Sparc 32 support is added.
562
563 26. Applied a modified version of Daniel Richard G's patch to create
564 pcre.h.generic and config.h.generic by "make" instead of in the
565 PrepareRelease script.
566
567 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
568 pcre_compile.c when checking for a zero character.
569
570 28. Introducing a native interface for JIT. Through this interface, the compiled
571 machine code can be directly executed. The purpose of this interface is to
572 provide fast pattern matching, so several sanity checks are not performed.
573 However, feature tests are still performed. The new interface provides
574 1.4x speedup compared to the old one.
575
576 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
577 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
578 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
579
580 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
581 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
582 the "old" RFC 2279). Instead, it ended up passing a negative length to
583 pcre_exec().
584
585 31. Add support for GCC's visibility feature to hide internal functions.
586
587 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
588 "unknown -C option" after outputting 0 or 1.
589
590 33. There is now support for generating a code coverage report for the test
591 suite in environments where gcc is the compiler and lcov is installed. This
592 is mainly for the benefit of the developers.
593
594 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
595 unaddressable using valgrind annotations, allowing valgrind to detect
596 invalid memory accesses. This is mainly for the benefit of the developers.
597
598 25. (*UTF) can now be used to start a pattern in any of the three libraries.
599
600 26. Give configure error if --enable-cpp but no C++ compiler found.
601
602
603 Version 8.31 06-July-2012
604 -------------------------
605
606 1. Fixing a wrong JIT test case and some compiler warnings.
607
608 2. Removed a bashism from the RunTest script.
609
610 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
611 to unsigned type, result still unsigned" that was given by an MS compiler
612 on encountering the code "-sizeof(xxx)".
613
614 4. Partial matching support is added to the JIT compiler.
615
616 5. Fixed several bugs concerned with partial matching of items that consist
617 of more than one character:
618
619 (a) /^(..)\1/ did not partially match "aba" because checking references was
620 done on an "all or nothing" basis. This also applied to repeated
621 references.
622
623 (b) \R did not give a hard partial match if \r was found at the end of the
624 subject.
625
626 (c) \X did not give a hard partial match after matching one or more
627 characters at the end of the subject.
628
629 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
630 a partial match for the string "\r".
631
632 (e) When newline was set to CRLF, the metacharacter "." did not recognize
633 a partial match for a CR character at the end of the subject string.
634
635 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
636 running pcretest, the text "(JIT)" added to the output whenever JIT is
637 actually used to run the match.
638
639 7. Individual JIT compile options can be set in pcretest by following -s+[+]
640 or /S+[+] with a digit between 1 and 7.
641
642 8. OP_NOT now supports any UTF character not just single-byte ones.
643
644 9. (*MARK) control verb is now supported by the JIT compiler.
645
646 10. The command "./RunTest list" lists the available tests without actually
647 running any of them. (Because I keep forgetting what they all are.)
648
649 11. Add PCRE_INFO_MAXLOOKBEHIND.
650
651 12. Applied a (slightly modified) user-supplied patch that improves performance
652 when the heap is used for recursion (compiled with --disable-stack-for-
653 recursion). Instead of malloc and free for each heap frame each time a
654 logical recursion happens, frames are retained on a chain and re-used where
655 possible. This sometimes gives as much as 30% improvement.
656
657 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
658 call.
659
660 14. As documented, (*COMMIT) is now confined to within a positive assertion.
661
662 15. It is now possible to link pcretest with libedit as an alternative to
663 libreadline.
664
665 16. (*COMMIT) control verb is now supported by the JIT compiler.
666
667 17. The Unicode data tables have been updated to Unicode 6.1.0.
668
669 18. Added --file-list option to pcregrep.
670
671 19. Added binary file support to pcregrep, including the -a, --binary-files,
672 -I, and --text options.
673
674 20. The madvise function is renamed for posix_madvise for QNX compatibility
675 reasons. Fixed by Giuseppe D'Angelo.
676
677 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
678
679 22. Changed the option for creating symbolic links for 16-bit man pages from
680 -s to -sf so that re-installing does not cause issues.
681
682 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
683
684 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
685 matches in certain environments (the workspace was not being correctly
686 retained). Also added to pcre_dfa_exec() a simple plausibility check on
687 some of the workspace data at the beginning of a restart.
688
689 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
690 was not doing so when it should - probably a typo introduced by SVN 528
691 (change 8.10/14).
692
693 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
694 \w+ when the character tables indicated that \x{c4} was a word character.
695 There were several related cases, all because the tests for doing a table
696 lookup were testing for characters less than 127 instead of 255.
697
698 27. If a pattern contains capturing parentheses that are not used in a match,
699 their slots in the ovector are set to -1. For those that are higher than
700 any matched groups, this happens at the end of processing. In the case when
701 there were back references that the ovector was too small to contain
702 (causing temporary malloc'd memory to be used during matching), and the
703 highest capturing number was not used, memory off the end of the ovector
704 was incorrectly being set to -1. (It was using the size of the temporary
705 memory instead of the true size.)
706
707 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
708 ovector size, it uses memory at the end of the block that it has got.
709
710 29. Check for an overlong MARK name and give an error at compile time. The
711 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
712
713 30. JIT compiler update.
714
715 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
716 Rill for the patch.
717
718 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
719
720 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
721
722 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
723 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
724 was enabled.
725
726 35. Improve JIT code generation for greedy plus quantifier.
727
728 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
729 1 to "aa" instead of to an empty string. The bug affected repeated groups
730 that could potentially match an empty string.
731
732 37. Optimizing single character iterators in JIT.
733
734 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
735 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
736 codepoints that are too big for the mode are faulted, and in a UTF mode,
737 disallowed codepoints are also faulted.
738
739 39. If PCRE was compiled with UTF support, in three places in the DFA
740 matcher there was code that should only have been obeyed in UTF mode, but
741 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
742 processing when bytes with values greater than 127 were present. In 16-bit
743 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
744 both cases the values are those that cannot be the first data item in a UTF
745 character. The three items that might have provoked this were recursions,
746 possessively repeated groups, and atomic groups.
747
748 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
749 and pcregrep, because some OS require shared objects to be explicitly
750 passed to ld, causing the link step to fail if they are not.
751
752 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
753 mode, patterns that started with \h* or \R* might be incorrectly matched.
754
755
756 Version 8.30 04-February-2012
757 -----------------------------
758
759 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
760 name is defined in ctype.h.
761
762 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
763 only in quite long subpatterns.
764
765 3. Removed the function pcre_info(), which has been obsolete and deprecated
766 since it was replaced by pcre_fullinfo() in February 2000.
767
768 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
769 match a (*MARK), and the match failed at the start of the subject, a
770 reference to memory before the start of the subject could occur. This bug
771 was introduced by fix 17 of release 8.21.
772
773 5. A reference to an unset group with zero minimum repetition was giving
774 totally wrong answers (in non-JavaScript-compatibility mode). For example,
775 /(another)?(\1?)test/ matched against "hello world test". This bug was
776 introduced in release 8.13.
777
778 6. Add support for 16-bit character strings (a large amount of work involving
779 many changes and refactorings).
780
781 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
782 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
783 from a file.
784
785 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
786 rounding is not applied in this particular case).
787
788 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
789 if they appear, or are escaped, in patterns.
790
791 10. Get rid of a number of -Wunused-but-set-variable warnings.
792
793 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
794 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
795 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
796 also returns the mark "x". This bug applied to capturing parentheses,
797 non-capturing parentheses, and atomic parentheses. It also applied to some
798 assertions.
799
800 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
801 information out of configure.ac instead of relying on pcre.h.generic, which
802 is not stored in the repository.
803
804 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
805 -lreadline.
806
807 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
808
809 15. Applied Graycode's patch to put the top-level frame on the stack rather
810 than the heap when not using the stack for recursion. This gives a
811 performance improvement in many cases when recursion is not deep.
812
813 16. Experimental code added to "pcretest -C" to output the stack frame size.
814
815
816 Version 8.21 12-Dec-2011
817 ------------------------
818
819 1. Updating the JIT compiler.
820
821 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
822 are added as well.
823
824 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
825 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
826 calling _pcre_jit_exec. Some extra comments are added.
827
828 4. (*MARK) settings inside atomic groups that do not contain any capturing
829 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
830 was introduced by change 18 for 8.20.
831
832 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
833 ECMA-262 standard.
834
835 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
836 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
837 This bug was probably introduced by change 9 of 8.13.
838
839 7. While fixing 6 above, I noticed that a number of other items were being
840 incorrectly rejected as "not fixed length". This arose partly because newer
841 opcodes had not been added to the fixed-length checking code. I have (a)
842 corrected the bug and added tests for these items, and (b) arranged for an
843 error to occur if an unknown opcode is encountered while checking for fixed
844 length instead of just assuming "not fixed length". The items that were
845 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
846 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
847 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
848
849 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
850 being incorrectly compiled and would have given unpredicatble results.
851
852 9. A possessively repeated subpattern with minimum repeat count greater than
853 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
854 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
855 the first (A) could occur when it should not.
856
857 10. Add a cast and remove a redundant test from the code.
858
859 11. JIT should use pcre_malloc/pcre_free for allocation.
860
861 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
862 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
863 is anything other than /usr, -L is still shown).
864
865 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
866
867 14. Perl does not support \N without a following name in a [] class; PCRE now
868 also gives an error.
869
870 15. If a forward reference was repeated with an upper limit of around 2000,
871 it caused the error "internal error: overran compiling workspace". The
872 maximum number of forward references (including repeats) was limited by the
873 internal workspace, and dependent on the LINK_SIZE. The code has been
874 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
875 the default depends on LINK_SIZE. There is a new upper limit (for safety)
876 of around 200,000 forward references. While doing this, I also speeded up
877 the filling in of repeated forward references.
878
879 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
880 incorrectly expecting the subject to contain another "a" after the start.
881
882 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
883 in the match, the SKIP should be ignored. This was not happening; instead
884 the SKIP was being treated as NOMATCH. For patterns such as
885 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
886 tested.
887
888 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
889 now much more compatible with Perl, in particular in cases where the result
890 is a non-match for a non-anchored pattern. For example, if
891 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
892 "m", where previously it did not return a name. A side effect of this
893 change is that for partial matches, the last encountered mark name is
894 returned, as for non matches. A number of tests that were previously not
895 Perl-compatible have been moved into the Perl-compatible test files. The
896 refactoring has had the pleasing side effect of removing one argument from
897 the match() function, thus reducing its stack requirements.
898
899 19. If the /S+ option was used in pcretest to study a pattern using JIT,
900 subsequent uses of /S (without +) incorrectly behaved like /S+.
901
902 21. Retrieve executable code size support for the JIT compiler and fixing
903 some warnings.
904
905 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
906 not work when the shorter character appeared right at the end of the
907 subject string.
908
909 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
910 systems.
911
912 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
913 output it when the /M option is used in pcretest.
914
915 25. The CheckMan script was not being included in the distribution. Also, added
916 an explicit "perl" to run Perl scripts from the PrepareRelease script
917 because this is reportedly needed in Windows.
918
919 26. If study data was being save in a file and studying had not found a set of
920 "starts with" bytes for the pattern, the data written to the file (though
921 never used) was taken from uninitialized memory and so caused valgrind to
922 complain.
923
924 27. Updated RunTest.bat as provided by Sheri Pierce.
925
926 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
927
928 29. Computation of memory usage for the table of capturing group names was
929 giving an unnecessarily large value.
930
931
932 Version 8.20 21-Oct-2011
933 ------------------------
934
935 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
936 a POSIX class. After further experiments with Perl, which convinced me that
937 Perl has bugs and confusions, a closing square bracket is no longer allowed
938 in a POSIX name. This bug also affected patterns with classes that started
939 with full stops.
940
941 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
942 captured substring, but while checking the failing first alternative,
943 substring 1 is temporarily captured. If the output vector supplied to
944 pcre_exec() was not big enough for this capture, the yield of the function
945 was still zero ("insufficient space for captured substrings"). This cannot
946 be totally fixed without adding another stack variable, which seems a lot
947 of expense for a edge case. However, I have improved the situation in cases
948 such as /(a)(b)x|abc/ matched against "abc", where the return code
949 indicates that fewer than the maximum number of slots in the ovector have
950 been set.
951
952 3. Related to (2) above: when there are more back references in a pattern than
953 slots in the output vector, pcre_exec() uses temporary memory during
954 matching, and copies in the captures as far as possible afterwards. It was
955 using the entire output vector, but this conflicts with the specification
956 that only 2/3 is used for passing back captured substrings. Now it uses
957 only the first 2/3, for compatibility. This is, of course, another edge
958 case.
959
960 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
961 main code base, and can be used by building with --enable-jit. When this is
962 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
963 runtime --no-jit option is given.
964
965 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
966 ovector, the return from the function was zero, implying that there were
967 other matches that did not fit. The correct "exactly full" value is now
968 returned.
969
970 6. If a subpattern that was called recursively or as a subroutine contained
971 (*PRUNE) or any other control that caused it to give a non-standard return,
972 invalid errors such as "Error -26 (nested recursion at the same subject
973 position)" or even infinite loops could occur.
974
975 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
976 computing the minimum length on reaching *ACCEPT, and so ended up with the
977 wrong value of 1 rather than 0. Further investigation indicates that
978 computing a minimum subject length in the presence of *ACCEPT is difficult
979 (think back references, subroutine calls), and so I have changed the code
980 so that no minimum is registered for a pattern that contains *ACCEPT.
981
982 8. If (*THEN) was present in the first (true) branch of a conditional group,
983 it was not handled as intended. [But see 16 below.]
984
985 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
986 Sheri Pierce.
987
988 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
989 the first byte in a match must be "a".
990
991 11. Change 17 for 8.13 increased the recursion depth for patterns like
992 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
993 pattern contains any instances of (*THEN). If it does not, the old
994 optimizations are restored. It would be nice to do this on a per-group
995 basis, but at the moment that is not feasible.
996
997 12. In some environments, the output of pcretest -C is CRLF terminated. This
998 broke RunTest's code that checks for the link size. A single white space
999 character after the value is now allowed for.
1000
1001 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1002 For "fr", it uses the Windows-specific input and output files.
1003
1004 14. If (*THEN) appeared in a group that was called recursively or as a
1005 subroutine, it did not work as intended. [But see next item.]
1006
1007 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1008 pattern fragments (but not containing any | characters). If A and B are
1009 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1010 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1011 D. In other words, Perl considers parentheses that do not contain any |
1012 characters to be part of a surrounding alternative, whereas PCRE was
1013 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1014 differently. PCRE now behaves in the same way as Perl, except in the case
1015 of subroutine/recursion calls such as (?1) which have in any case always
1016 been different (but PCRE had them first :-).
1017
1018 16. Related to 15 above: Perl does not treat the | in a conditional group as
1019 creating alternatives. Such a group is treated in the same way as an
1020 ordinary group without any | characters when processing (*THEN). PCRE has
1021 been changed to match Perl's behaviour.
1022
1023 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1024 RunGrepTest script failed.
1025
1026 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1027 inevitable for groups that contain captures, but it can lead to a lot of
1028 stack use in large patterns. The old behaviour has been restored for atomic
1029 groups that do not contain any capturing parentheses.
1030
1031 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1032 suppress the check for a minimum subject length at run time. (If it was
1033 given to pcre_exec() or pcre_dfa_exec() it did work.)
1034
1035 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1036 fail to work when decoding hex characters in data strings in EBCDIC
1037 environments.
1038
1039 21. It appears that in at least one Mac OS environment, the isxdigit() function
1040 is implemented as a macro that evaluates to its argument more than once,
1041 contravening the C 90 Standard (I haven't checked a later standard). There
1042 was an instance in pcretest which caused it to go wrong when processing
1043 \x{...} escapes in subject strings. The has been rewritten to avoid using
1044 things like p++ in the argument of isxdigit().
1045
1046
1047 Version 8.13 16-Aug-2011
1048 ------------------------
1049
1050 1. The Unicode data tables have been updated to Unicode 6.0.0.
1051
1052 2. Two minor typos in pcre_internal.h have been fixed.
1053
1054 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1055 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1056 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1057
1058 4. There were a number of related bugs in the code for matching backrefences
1059 caselessly in UTF-8 mode when codes for the characters concerned were
1060 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1061 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1062 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1063 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1064 2-byte code at the end of the subject (it thought there wasn't enough data
1065 left).
1066
1067 5. Comprehensive information about what went wrong is now returned by
1068 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1069 as the output vector has at least 2 elements. The offset of the start of
1070 the failing character and a reason code are placed in the vector.
1071
1072 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1073 now returned is for the first byte of the failing character, instead of the
1074 last byte inspected. This is an incompatible change, but I hope it is small
1075 enough not to be a problem. It makes the returned offset consistent with
1076 pcre_exec() and pcre_dfa_exec().
1077
1078 7. pcretest now gives a text phrase as well as the error number when
1079 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1080 failure, the offset and reason code are output.
1081
1082 8. When \R was used with a maximizing quantifier it failed to skip backwards
1083 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1084 back over a single character (\n). This seems wrong (because it treated the
1085 two characters as a single entity when going forwards), conflicts with the
1086 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1087 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1088 has been changed.
1089
1090 9. Some internal refactoring has changed the processing so that the handling
1091 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1092 time (the PCRE_DOTALL option was changed this way some time ago: version
1093 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1094 which was always a bit of a fudge. It also means that there is one less
1095 argument for the match() function, which reduces its stack requirements
1096 slightly. This change also fixes an incompatibility with Perl: the pattern
1097 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1098
1099 10. More internal refactoring has drastically reduced the number of recursive
1100 calls to match() for possessively repeated groups such as (abc)++ when
1101 using pcre_exec().
1102
1103 11. While implementing 10, a number of bugs in the handling of groups were
1104 discovered and fixed:
1105
1106 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1107 (a|)*(?1) gave a compile-time internal error.
1108 ((a|)+)+ did not notice that the outer group could match an empty string.
1109 (^a|^)+ was not marked as anchored.
1110 (.*a|.*)+ was not marked as matching at start or after a newline.
1111
1112 12. Yet more internal refactoring has removed another argument from the match()
1113 function. Special calls to this function are now indicated by setting a
1114 value in a variable in the "match data" data block.
1115
1116 13. Be more explicit in pcre_study() instead of relying on "default" for
1117 opcodes that mean there is no starting character; this means that when new
1118 ones are added and accidentally left out of pcre_study(), testing should
1119 pick them up.
1120
1121 14. The -s option of pcretest has been documented for ages as being an old
1122 synonym of -m (show memory usage). I have changed it to mean "force study
1123 for every regex", that is, assume /S for every regex. This is similar to -i
1124 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1125 using it. It makes it easier to run collections of tests with and without
1126 study enabled, and thereby test pcre_study() more easily. All the standard
1127 tests are now run with and without -s (but some patterns can be marked as
1128 "never study" - see 20 below).
1129
1130 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1131 restoration of the capturing data to the outer values was not happening
1132 correctly.
1133
1134 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1135 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1136 pattern had matched an empty string, and so incorrectly returned a no
1137 match.
1138
1139 17. There was optimizing code for the last branch of non-capturing parentheses,
1140 and also for the obeyed branch of a conditional subexpression, which used
1141 tail recursion to cut down on stack usage. Unfortunately, now that there is
1142 the possibility of (*THEN) occurring in these branches, tail recursion is
1143 no longer possible because the return has to be checked for (*THEN). These
1144 two optimizations have therefore been removed. [But see 8.20/11 above.]
1145
1146 18. If a pattern containing \R was studied, it was assumed that \R always
1147 matched two bytes, thus causing the minimum subject length to be
1148 incorrectly computed because \R can also match just one byte.
1149
1150 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1151 was incorrectly computed.
1152
1153 20. If /S is present twice on a test pattern in pcretest input, it now
1154 *disables* studying, thereby overriding the use of -s on the command line
1155 (see 14 above). This is necessary for one or two tests to keep the output
1156 identical in both cases.
1157
1158 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1159 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1160
1161 22. When an atomic group that contained a capturing parenthesis was
1162 successfully matched, but the branch in which it appeared failed, the
1163 capturing was not being forgotten if a higher numbered group was later
1164 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1165 group 1 to "a", when in fact it should be unset. This applied to multi-
1166 branched capturing and non-capturing groups, repeated or not, and also to
1167 positive assertions (capturing in negative assertions does not happen
1168 in PCRE) and also to nested atomic groups.
1169
1170 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1171 subject after a captured substring, to make it easier to tell which of a
1172 number of identical substrings has been captured.
1173
1174 24. The way atomic groups are processed by pcre_exec() has been changed so that
1175 if they are repeated, backtracking one repetition now resets captured
1176 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1177 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1178 "aaa". Previously, it would have been "a". As part of this code
1179 refactoring, the way recursive calls are handled has also been changed.
1180
1181 25. If an assertion condition captured any substrings, they were not passed
1182 back unless some other capturing happened later. For example, if
1183 (?(?=(a))a) was matched against "a", no capturing was returned.
1184
1185 26. When studying a pattern that contained subroutine calls or assertions,
1186 the code for finding the minimum length of a possible match was handling
1187 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1188 group 1 called group 2 while simultaneously a separate group 2 called group
1189 1). A stack overflow occurred in this case. I have fixed this by limiting
1190 the recursion depth to 10.
1191
1192 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1193 Fortmann. This supports explicit test numbers on the command line, and has
1194 argument validation and error reporting.
1195
1196 28. An instance of \X with an unlimited repeat could fail if at any point the
1197 first character it looked at was a mark character.
1198
1199 29. Some minor code refactoring concerning Unicode properties and scripts
1200 should reduce the stack requirement of match() slightly.
1201
1202 30. Added the '=' option to pcretest to check the setting of unused capturing
1203 slots at the end of the pattern, which are documented as being -1, but are
1204 not included in the return count.
1205
1206 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1207 compiled something random. Now it gives a compile-time error (as does
1208 Perl).
1209
1210 32. A *MARK encountered during the processing of a positive assertion is now
1211 recorded and passed back (compatible with Perl).
1212
1213 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1214 had alternative anchored branches, the search for a second match in a line
1215 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1216 matched the line "0102" twice. The same bug affected patterns that started
1217 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1218 twice.
1219
1220 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1221 does, and because of capturing effects, quantifying parenthesized
1222 assertions may at times be useful. Quantifiers are now allowed for
1223 parenthesized assertions.
1224
1225 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1226
1227 36. \g was being checked for fancy things in a character class, when it should
1228 just be a literal "g".
1229
1230 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1231 appearance of a nested POSIX class supersedes an apparent external class.
1232 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1233 unescaped square brackets may also appear as part of class names. For
1234 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1235 more like Perl. (But see 8.20/1 above.)
1236
1237 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1238 was because it thought it was \N{name}, which is not supported).
1239
1240 39. Add minix to OS list not supporting the -S option in pcretest.
1241
1242 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1243 cannot analyze patterns in sufficient detail to catch mutual recursions
1244 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1245 subgroup is called recursively as a subpattern for a second time at the
1246 same position in the subject string. In previous releases this might have
1247 been caught by the recursion limit, or it might have run out of stack.
1248
1249 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1250 happen only once. PCRE was, however incorrectly giving a compile time error
1251 "recursive call could loop indefinitely" because it cannot analyze the
1252 pattern in sufficient detail. The compile time test no longer happens when
1253 PCRE is compiling a conditional subpattern, but actual runaway loops are
1254 now caught at runtime (see 40 above).
1255
1256 42. It seems that Perl allows any characters other than a closing parenthesis
1257 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1258 has been changed to be the same.
1259
1260 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1261 as not to get warnings when autogen.sh is called. Also changed
1262 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1263
1264 44. To help people who use pcregrep to scan files containing exceedingly long
1265 lines, the following changes have been made:
1266
1267 (a) The default value of the buffer size parameter has been increased from
1268 8K to 20K. (The actual buffer used is three times this size.)
1269
1270 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1271 PCRE is built.
1272
1273 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1274 to be set at run time.
1275
1276 (d) Numerical values in pcregrep options can be followed by K or M, for
1277 example --buffer-size=50K.
1278
1279 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1280 given and the return code is set to 2.
1281
1282 45. Add a pointer to the latest mark to the callout data block.
1283
1284 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1285 partial match of an empty string instead of no match. This was specific to
1286 the use of ".".
1287
1288 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1289 complete match instead of a partial match. This bug was dependent on both
1290 the PCRE_UTF8 and PCRE_DOTALL options being set.
1291
1292 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1293 starting byte set, because \b was not being ignored.
1294
1295
1296 Version 8.12 15-Jan-2011
1297 ------------------------
1298
1299 1. Fixed some typos in the markup of the man pages, and wrote a script that
1300 checks for such things as part of the documentation building process.
1301
1302 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1303 --match-limit and --recursion-limit options (added for 8.11). In
1304 particular, this made one of the standard tests fail. (The integer value
1305 went into the wrong half of a long int.)
1306
1307 3. If the --colour option was given to pcregrep with -v (invert match), it
1308 did strange things, either producing crazy output, or crashing. It should,
1309 of course, ignore a request for colour when reporting lines that do not
1310 match.
1311
1312 4. Another pcregrep bug caused similar problems if --colour was specified with
1313 -M (multiline) and the pattern match finished with a line ending.
1314
1315 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1316 matched in multiline mode, the following line was shown as part of the
1317 match. This seems wrong, so I have changed it.
1318
1319 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1320 the check for further matches in the same line (so they could be coloured)
1321 to overrun the end of the current line. If another match was found, it was
1322 incorrectly shown (and then shown again when found in the next line).
1323
1324 7. If pcregrep was compiled under Windows, there was a reference to the
1325 function pcregrep_exit() before it was defined. I am assuming this was
1326 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1327 reported by a user. I've moved the definition above the reference.
1328
1329
1330 Version 8.11 10-Dec-2010
1331 ------------------------
1332
1333 1. (*THEN) was not working properly if there were untried alternatives prior
1334 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1335 backtracked to try for "b" instead of moving to the next alternative branch
1336 at the same level (in this case, to look for "c"). The Perl documentation
1337 is clear that when (*THEN) is backtracked onto, it goes to the "next
1338 alternative in the innermost enclosing group".
1339
1340 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1341 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1342 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1343 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1344 (*THEN).
1345
1346 3. If \s appeared in a character class, it removed the VT character from
1347 the class, even if it had been included by some previous item, for example
1348 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1349 of \s, but is part of the POSIX "space" class.)
1350
1351 4. A partial match never returns an empty string (because you can always
1352 match an empty string at the end of the subject); however the checking for
1353 an empty string was starting at the "start of match" point. This has been
1354 changed to the "earliest inspected character" point, because the returned
1355 data for a partial match starts at this character. This means that, for
1356 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1357 (previously it gave "no match").
1358
1359 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1360 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1361 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1362 has an implication that the given string is incomplete (because a partial
1363 match is preferred over a full match). For this reason, these items now
1364 give a partial match in this situation. [Aside: previously, the one case
1365 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1366 match rather than a full match, which was wrong by the old rules, but is
1367 now correct.]
1368
1369 6. There was a bug in the handling of #-introduced comments, recognized when
1370 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1371 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1372 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1373 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1374 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1375 places in pcre_compile().
1376
1377 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1378 comments when looking ahead for named forward references to subpatterns,
1379 the only newline sequence it recognized was NL. It now handles newlines
1380 according to the set newline convention.
1381
1382 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1383 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1384 cater for a lack of strerror(). These oversights have been fixed.
1385
1386 9. Added --match-limit and --recursion-limit to pcregrep.
1387
1388 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1389
1390 11. When the -o option was used, pcregrep was setting a return code of 1, even
1391 when matches were found, and --line-buffered was not being honoured.
1392
1393 12. Added an optional parentheses number to the -o and --only-matching options
1394 of pcregrep.
1395
1396 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1397 can match an empty string. The code to do it in pcretest and pcredemo
1398 needed fixing:
1399
1400 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1401 only one byte after an empty string match just before CRLF (this case
1402 just got forgotten; "any" and "anycrlf" were OK).
1403
1404 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1405 mode when an empty string match preceded an ASCII character followed by
1406 a non-ASCII character. (The code for advancing by one character rather
1407 than one byte was nonsense.)
1408
1409 (c) The pcredemo.c sample program did not have any code at all to handle
1410 the cases when CRLF is a valid newline sequence.
1411
1412 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1413 as a starting offset was within the subject string. There is now a new
1414 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1415 negative or greater than the length of the string. In order to test this,
1416 pcretest is extended to allow the setting of negative starting offsets.
1417
1418 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1419 starting offset points to the beginning of a UTF-8 character was
1420 unnecessarily clumsy. I tidied it up.
1421
1422 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1423 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1424
1425 17. Nobody had reported that the --include_dir option, which was added in
1426 release 7.7 should have been called --include-dir (hyphen, not underscore)
1427 for compatibility with GNU grep. I have changed it to --include-dir, but
1428 left --include_dir as an undocumented synonym, and the same for
1429 --exclude-dir, though that is not available in GNU grep, at least as of
1430 release 2.5.4.
1431
1432 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1433 characters from a string of bytes) have been redefined so as not to use
1434 loops, in order to improve performance in some environments. At the same
1435 time, I abstracted some of the common code into auxiliary macros to save
1436 repetition (this should not affect the compiled code).
1437
1438 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1439 compile-time error is now given if \c is not followed by an ASCII
1440 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1441 different, and any byte value is allowed.)
1442
1443 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1444 START_OPTIMIZE option, which is now allowed at compile time - but just
1445 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1446 to pcregrep and other applications that have no direct access to PCRE
1447 options. The new /Y option in pcretest sets this option when calling
1448 pcre_compile().
1449
1450 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1451 back references. Groups containing recursive back references were forced to
1452 be atomic by that change, but in the case of named groups, the amount of
1453 memory required was incorrectly computed, leading to "Failed: internal
1454 error: code overflow". This has been fixed.
1455
1456 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1457 pcretest.c, to avoid build problems in some Borland environments.
1458
1459
1460 Version 8.10 25-Jun-2010
1461 ------------------------
1462
1463 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1464 THEN.
1465
1466 2. (*ACCEPT) was not working when inside an atomic group.
1467
1468 3. Inside a character class, \B is treated as a literal by default, but
1469 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1470 causes the error). The code is unchanged, but I tidied the documentation.
1471
1472 4. Inside a character class, PCRE always treated \R and \X as literals,
1473 whereas Perl faults them if its -w option is set. I have changed PCRE so
1474 that it faults them when PCRE_EXTRA is set.
1475
1476 5. Added support for \N, which always matches any character other than
1477 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1478
1479 6. When compiling pcregrep with newer versions of gcc which may have
1480 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1481 declared with attribute warn_unused_result" were given. Just casting the
1482 result to (void) does not stop the warnings; a more elaborate fudge is
1483 needed. I've used a macro to implement this.
1484
1485 7. Minor change to pcretest.c to avoid a compiler warning.
1486
1487 8. Added four artifical Unicode properties to help with an option to make
1488 \s etc use properties (see next item). The new properties are: Xan
1489 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1490
1491 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1492 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1493 this option. Modified pcretest to add /W to test this facility. Added
1494 REG_UCP to make it available via the POSIX interface.
1495
1496 10. Added --line-buffered to pcregrep.
1497
1498 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1499 studied, and the match started with a letter with a code point greater than
1500 127 whose first byte was different to the first byte of the other case of
1501 the letter, the other case of this starting letter was not recognized
1502 (#976).
1503
1504 12. If a pattern that was studied started with a repeated Unicode property
1505 test, for example, \p{Nd}+, there was the theoretical possibility of
1506 setting up an incorrect bitmap of starting bytes, but fortunately it could
1507 not have actually happened in practice until change 8 above was made (it
1508 added property types that matched character-matching opcodes).
1509
1510 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1511 possible starting bytes for non-anchored patterns.
1512
1513 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1514 \R, and also a number of cases that involve Unicode properties, both
1515 explicit and implicit when PCRE_UCP is set.
1516
1517 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1518 input, it could crash or give wrong results if characters with values
1519 greater than 0xc0 were present in the subject string. (Detail: it assumed
1520 UTF-8 input when processing these items.)
1521
1522 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1523 size_t is 64-bit (#991).
1524
1525 17. Added a check for running out of memory when PCRE is compiled with
1526 --disable-stack-for-recursion (#990).
1527
1528 18. If the last data line in a file for pcretest does not have a newline on
1529 the end, a newline was missing in the output.
1530
1531 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1532 less than 128) in its various bitmaps. However, there is a facility for
1533 generating tables according to the current locale when PCRE is compiled. It
1534 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1535 characters, are recognized by isspace() and therefore were getting set in
1536 these tables, and indeed these tables seem to approximate to ISO 8859. This
1537 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1538 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1539 which of course cannot start UTF-8 characters. I have changed the code so
1540 that only real ASCII characters (less than 128) and the correct starting
1541 bytes for UTF-8 encodings are set for characters greater than 127 when in
1542 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1543 altogether.)
1544
1545 20. Added the /T option to pcretest so as to be able to run tests with non-
1546 standard character tables, thus making it possible to include the tests
1547 used for 19 above in the standard set of tests.
1548
1549 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1550 reference to a subpattern the other side of a comment that contains an
1551 opening parenthesis caused either an internal compiling error, or a
1552 reference to the wrong subpattern.
1553
1554
1555 Version 8.02 19-Mar-2010
1556 ------------------------
1557
1558 1. The Unicode data tables have been updated to Unicode 5.2.0.
1559
1560 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1561 configured.
1562
1563 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1564 original author of that file, following a query about its status.
1565
1566 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1567 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1568
1569 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1570 quantifier applied to a forward-referencing subroutine call, could compile
1571 incorrect code or give the error "internal error: previously-checked
1572 referenced subpattern not found".
1573
1574 6. Both MS Visual Studio and Symbian OS have problems with initializing
1575 variables to point to external functions. For these systems, therefore,
1576 pcre_malloc etc. are now initialized to local functions that call the
1577 relevant global functions.
1578
1579 7. There were two entries missing in the vectors called coptable and poptable
1580 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1581 I've fixed the data, and added a kludgy way of testing at compile time that
1582 the lengths are correct (equal to the number of opcodes).
1583
1584 8. Following on from 7, I added a similar kludge to check the length of the
1585 eint vector in pcreposix.c.
1586
1587 9. Error texts for pcre_compile() are held as one long string to avoid too
1588 much relocation at load time. To find a text, the string is searched,
1589 counting zeros. There was no check for running off the end of the string,
1590 which could happen if a new error number was added without updating the
1591 string.
1592
1593 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1594
1595 11. \K was not working if it appeared in an atomic group or in a group that
1596 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1597 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1598 the assertion is positive, but not if it is negative.
1599
1600 12. Change 11 fortuitously reduced the size of the stack frame used in the
1601 "match()" function of pcre_exec.c by one pointer. Forthcoming
1602 implementation of support for (*MARK) will need an extra pointer on the
1603 stack; I have reserved it now, so that the stack frame size does not
1604 decrease.
1605
1606 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1607 item in branch that calls a recursion is a subroutine call - as in the
1608 second branch in the above example - was incorrectly given the compile-
1609 time error "recursive call could loop indefinitely" because pcre_compile()
1610 was not correctly checking the subroutine for matching a non-empty string.
1611
1612 14. The checks for overrunning compiling workspace could trigger after an
1613 overrun had occurred. This is a "should never occur" error, but it can be
1614 triggered by pathological patterns such as hundreds of nested parentheses.
1615 The checks now trigger 100 bytes before the end of the workspace.
1616
1617 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1618
1619
1620 Version 8.01 19-Jan-2010
1621 ------------------------
1622
1623 1. If a pattern contained a conditional subpattern with only one branch (in
1624 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1625 computed the wrong minimum data length (which is of course zero for such
1626 subpatterns). This could cause incorrect "no match" results.
1627
1628 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1629 the pattern is reset in the first branch, pcre_compile() failed with
1630 "internal error: code overflow at offset...". This happened only when
1631 the reset was to the original external option setting. (An optimization
1632 abstracts leading options settings into an external setting, which was the
1633 cause of this.)
1634
1635 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1636 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1637 assertion pattern did not match (meaning that the assertion was true), it
1638 was incorrectly treated as false if the SKIP had been reached during the
1639 matching. This also applied to assertions used as conditions.
1640
1641 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1642 assertion subpattern, including such a pattern used as a condition,
1643 unpredictable results occurred, instead of the error return
1644 PCRE_ERROR_DFA_UITEM.
1645
1646 5. The C++ GlobalReplace function was not working like Perl for the special
1647 situation when an empty string is matched. It now does the fancy magic
1648 stuff that is necessary.
1649
1650 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1651 removed. (These were left over from very, very early versions of PCRE.)
1652
1653 7. Some cosmetic changes to the code to make life easier when compiling it
1654 as part of something else:
1655
1656 (a) Change DEBUG to PCRE_DEBUG.
1657
1658 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1659 called "current" as "current_branch", to prevent a collision with the
1660 Linux macro when compiled as a kernel module.
1661
1662 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1663 prevent a collision with the Linux macro when compiled as a kernel
1664 module.
1665
1666 8. In pcre_compile() there are some checks for integer overflows that used to
1667 cast potentially large values to (double). This has been changed to that
1668 when building, a check for int64_t is made, and if it is found, it is used
1669 instead, thus avoiding the use of floating point arithmetic. (There is no
1670 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1671 double.
1672
1673 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1674 2005 (difference between two addresses compared to an unsigned value).
1675
1676 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1677 custom one, because of the following reported problem in Windows:
1678
1679 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1680 under Win32.
1681 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1682 therefore missing the function definition.
1683 - The compiler thus generates a "C" signature for the test function.
1684 - The linker fails to find the "C" function.
1685 - PCRE fails to configure if asked to do so against libbz2.
1686
1687 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1688 messages were output:
1689
1690 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1691 rerunning libtoolize, to keep the correct libtool macros in-tree.
1692 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1693
1694 I have done both of these things.
1695
1696 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1697 most of the time, it *can* run out if it is given a pattern that contains a
1698 runaway infinite recursion. I updated the discussion in the pcrestack man
1699 page.
1700
1701 13. Now that we have gone to the x.xx style of version numbers, the minor
1702 version may start with zero. Using 08 or 09 is a bad idea because users
1703 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1704 interpreted as invalid octal numbers. I've updated the previous comment in
1705 configure.ac, and also added a check that gives an error if 08 or 09 are
1706 used.
1707
1708 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1709 causing partial matching to fail when the end of the subject matched \W
1710 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1711
1712 15. There were some discrepancies between the declarations in pcre_internal.h
1713 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1714 their definitions. The declarations used "const uschar *" and the
1715 definitions used USPTR. Even though USPTR is normally defined as "const
1716 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1717 reported that: "This difference in casting confuses some C++ compilers, for
1718 example, SunCC recognizes above declarations as different functions and
1719 generates broken code for hbpcre." I have changed the declarations to use
1720 USPTR.
1721
1722 16. GNU libtool is named differently on some systems. The autogen.sh script now
1723 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1724 (FreeBSD).
1725
1726 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1727 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1728 comment: "Figure out how to create a longlong from a string: strtoll and
1729 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1730 instance, but it only takes 2 args instead of 3!"
1731
1732 18. A subtle bug concerned with back references has been fixed by a change of
1733 specification, with a corresponding code fix. A pattern such as
1734 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1735 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1736 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1737 same bug. Such groups have to be quantified to be useful, or contained
1738 inside another quantified group. (If there's no repetition, the reference
1739 can never match.) The problem arises because, having left the group and
1740 moved on to the rest of the pattern, a later failure that backtracks into
1741 the group uses the captured value from the final iteration of the group
1742 rather than the correct earlier one. I have fixed this in PCRE by forcing
1743 any group that contains a reference to itself to be an atomic group; that
1744 is, there cannot be any backtracking into it once it has completed. This is
1745 similar to recursive and subroutine calls.
1746
1747
1748 Version 8.00 19-Oct-09
1749 ----------------------
1750
1751 1. The table for translating pcre_compile() error codes into POSIX error codes
1752 was out-of-date, and there was no check on the pcre_compile() error code
1753 being within the table. This could lead to an OK return being given in
1754 error.
1755
1756 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1757 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1758 in a Windows environment.
1759
1760 3. The pcregrep --count option prints the count for each file even when it is
1761 zero, as does GNU grep. However, pcregrep was also printing all files when
1762 --files-with-matches was added. Now, when both options are given, it prints
1763 counts only for those files that have at least one match. (GNU grep just
1764 prints the file name in this circumstance, but including the count seems
1765 more useful - otherwise, why use --count?) Also ensured that the
1766 combination -clh just lists non-zero counts, with no names.
1767
1768 4. The long form of the pcregrep -F option was incorrectly implemented as
1769 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1770 but it seems right to fix it, and I didn't think it was worth preserving
1771 the old behaviour.
1772
1773 5. The command line items --regex=pattern and --regexp=pattern were not
1774 recognized by pcregrep, which required --regex pattern or --regexp pattern
1775 (with a space rather than an '='). The man page documented the '=' forms,
1776 which are compatible with GNU grep; these now work.
1777
1778 6. No libpcreposix.pc file was created for pkg-config; there was just
1779 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1780
1781 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1782 when UCP support is not needed, by modifying the Python script that
1783 generates it from Unicode data files. This should not matter if the module
1784 is correctly used as a library, but I received one complaint about 50K of
1785 unwanted data. My guess is that the person linked everything into his
1786 program rather than using a library. Anyway, it does no harm.
1787
1788 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1789 was a minimum greater than 1 for a wide character in a possessive
1790 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1791 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1792 character. Chaos in the form of incorrect output or a compiling loop could
1793 result.
1794
1795 9. The restrictions on what a pattern can contain when partial matching is
1796 requested for pcre_exec() have been removed. All patterns can now be
1797 partially matched by this function. In addition, if there are at least two
1798 slots in the offset vector, the offset of the earliest inspected character
1799 for the match and the offset of the end of the subject are set in them when
1800 PCRE_ERROR_PARTIAL is returned.
1801
1802 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1803 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1804 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1805 and may be more useful for multi-segment matching.
1806
1807 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1808 used to be given if ever the end of the subject was reached; now it is
1809 given only if matching could not proceed because another character was
1810 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1811 string "Z", which now yields "no match" instead of "partial match". In the
1812 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1813 final character ended with (*FAIL).
1814
1815 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1816 if the pattern had a "must contain" character that was already found in the
1817 earlier partial match, unless partial matching was again requested. For
1818 example, with the pattern /dog.(body)?/, the "must contain" character is
1819 "g". If the first part-match was for the string "dog", restarting with
1820 "sbody" failed. This bug has been fixed.
1821
1822 13. The string returned by pcre_dfa_exec() after a partial match has been
1823 changed so that it starts at the first inspected character rather than the
1824 first character of the match. This makes a difference only if the pattern
1825 starts with a lookbehind assertion or \b or \B (\K is not supported by
1826 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1827 matching functions compatible, and I think it's the right thing to do.
1828
1829 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1830 so that the demonstration program is easily available in environments where
1831 PCRE has not been installed from source.
1832
1833 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1834 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1835 library.
1836
1837 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1838 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1839 is not the first non-POSIX option to be added. Clearly some people find
1840 these options useful.
1841
1842 17. If a caller to the POSIX matching function regexec() passes a non-zero
1843 value for nmatch with a NULL value for pmatch, the value of
1844 nmatch is forced to zero.
1845
1846 18. RunGrepTest did not have a test for the availability of the -u option of
1847 the diff command, as RunTest does. It now checks in the same way as
1848 RunTest, and also checks for the -b option.
1849
1850 19. If an odd number of negated classes containing just a single character
1851 interposed, within parentheses, between a forward reference to a named
1852 subpattern and the definition of the subpattern, compilation crashed with
1853 an internal error, complaining that it could not find the referenced
1854 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1855 [The bug was that it was starting one character too far in when skipping
1856 over the character class, thus treating the ] as data rather than
1857 terminating the class. This meant it could skip too much.]
1858
1859 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1860 /g option in pcretest when the pattern contains \K, which makes it possible
1861 to have an empty string match not at the start, even when the pattern is
1862 anchored. Updated pcretest and pcredemo to use this option.
1863
1864 21. If the maximum number of capturing subpatterns in a recursion was greater
1865 than the maximum at the outer level, the higher number was returned, but
1866 with unset values at the outer level. The correct (outer level) value is
1867 now given.
1868
1869 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1870 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1871 make it do so. The string so far is captured, making this feature
1872 compatible with Perl.
1873
1874 23. The tests have been re-organized, adding tests 11 and 12, to make it
1875 possible to check the Perl 5.10 features against Perl 5.10.
1876
1877 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1878 pattern matches a fixed length string. PCRE did not allow this; now it
1879 does. Neither allows recursion.
1880
1881 25. I finally figured out how to implement a request to provide the minimum
1882 length of subject string that was needed in order to match a given pattern.
1883 (It was back references and recursion that I had previously got hung up
1884 on.) This code has now been added to pcre_study(); it finds a lower bound
1885 to the length of subject needed. It is not necessarily the greatest lower
1886 bound, but using it to avoid searching strings that are too short does give
1887 some useful speed-ups. The value is available to calling programs via
1888 pcre_fullinfo().
1889
1890 26. While implementing 25, I discovered to my embarrassment that pcretest had
1891 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1892 study optimizations had never been tested with that matching function.
1893 Oops. What is worse, even when it was passed study data, there was a bug in
1894 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1895 were also very few tests of studied patterns with pcre_dfa_exec().
1896
1897 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1898 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1899 on the other side of the coin, they are no longer allowed to have different
1900 names, because these cannot be distinguished in PCRE, and this has caused
1901 confusion. (This is a difference from Perl.)
1902
1903 28. When duplicate subpattern names are present (necessarily with different
1904 numbers, as required by 27 above), and a test is made by name in a
1905 conditional pattern, either for a subpattern having been matched, or for
1906 recursion in such a pattern, all the associated numbered subpatterns are
1907 tested, and the overall condition is true if the condition is true for any
1908 one of them. This is the way Perl works, and is also more like the way
1909 testing by number works.
1910
1911
1912 Version 7.9 11-Apr-09
1913 ---------------------
1914
1915 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1916 (pcretest), all targets were linked against these libraries. This included
1917 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1918 libraries. This caused unwanted dependencies to be created. This problem
1919 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1920 pcretest is linked with readline.
1921
1922 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1923 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1924 moved outside it again, because FALSE and TRUE are already defined in AIX,
1925 but BOOL is not.
1926
1927 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1928 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1929
1930 4. The pcregrep documentation said spaces were inserted as well as colons (or
1931 hyphens) following file names and line numbers when outputting matching
1932 lines. This is not true; no spaces are inserted. I have also clarified the
1933 wording for the --colour (or --color) option.
1934
1935 5. In pcregrep, when --colour was used with -o, the list of matching strings
1936 was not coloured; this is different to GNU grep, so I have changed it to be
1937 the same.
1938
1939 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1940 each matching line was coloured. Now it goes on to look for further matches
1941 of any of the test patterns, which is the same behaviour as GNU grep.
1942
1943 7. A pattern that could match an empty string could cause pcregrep to loop; it
1944 doesn't make sense to accept an empty string match in pcregrep, so I have
1945 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1946 seems to be how GNU grep behaves. [But see later change 40 for release
1947 8.33.]
1948
1949 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1950 start or after a newline", because the conditional assertion was not being
1951 correctly handled. The rule now is that both the assertion and what follows
1952 in the first alternative must satisfy the test.
1953
1954 9. If auto-callout was enabled in a pattern with a conditional group whose
1955 condition was an assertion, PCRE could crash during matching, both with
1956 pcre_exec() and pcre_dfa_exec().
1957
1958 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1959 used for matching.
1960
1961 11. Unicode property support in character classes was not working for
1962 characters (bytes) greater than 127 when not in UTF-8 mode.
1963
1964 12. Added the -M command line option to pcretest.
1965
1966 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1967
1968 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1969
1970 16. Added comments and documentation about mis-use of no_arg in the C++
1971 wrapper.
1972
1973 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1974 from Martin Jerabek that uses macro names for all relevant character and
1975 string constants.
1976
1977 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1978 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1979 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1980 these, but not everybody uses configure.
1981
1982 19. A conditional group that had only one branch was not being correctly
1983 recognized as an item that could match an empty string. This meant that an
1984 enclosing group might also not be so recognized, causing infinite looping
1985 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1986 with the subject "ab", where knowledge that the repeated group can match
1987 nothing is needed in order to break the loop.
1988
1989 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1990 exec(), but without supplying a callout function, matching went wrong.
1991
1992 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1993 leak if the size of the offset vector was greater than 30. When the vector
1994 is smaller, the saved offsets during recursion go onto a local stack
1995 vector, but for larger vectors malloc() is used. It was failing to free
1996 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1997 error, in fact).
1998
1999 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2000 heapframe that is used only when UTF-8 support is enabled. This caused no
2001 problem, but was untidy.
2002
2003 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2004 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2005 included within another project.
2006
2007 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2008 slightly modified by me:
2009
2010 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2011 not building pcregrep.
2012
2013 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2014 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2015
2016 25. Forward references, both numeric and by name, in patterns that made use of
2017 duplicate group numbers, could behave incorrectly or give incorrect errors,
2018 because when scanning forward to find the reference group, PCRE was not
2019 taking into account the duplicate group numbers. A pattern such as
2020 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2021
2022 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2023 the feature of a custom pointer more persuasive (as requested by a user).
2024
2025 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2026 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2027 pre-defined.
2028
2029 28. Added support for (*UTF8) at the start of a pattern.
2030
2031 29. Arrange for flags added by the "release type" setting in CMake to be shown
2032 in the configuration summary.
2033
2034
2035 Version 7.8 05-Sep-08
2036 ---------------------
2037
2038 1. Replaced UCP searching code with optimized version as implemented for Ad
2039 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2040 stage table and inline lookup instead of a function, giving speed ups of 2
2041 to 5 times on some simple patterns that I tested. Permission was given to
2042 distribute the MultiStage2.py script that generates the tables (it's not in
2043 the tarball, but is in the Subversion repository).
2044
2045 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2046 scripts.
2047
2048 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2049 a group with a zero qualifier. The result of the study could be incorrect,
2050 or the function might crash, depending on the pattern.
2051
2052 4. Caseless matching was not working for non-ASCII characters in back
2053 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2054 It now works when Unicode Property Support is available.
2055
2056 5. In pcretest, an escape such as \x{de} in the data was always generating
2057 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2058 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2059 truncation.
2060
2061 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2062
2063 7. Added two (int) casts to pcregrep when printing the difference of two
2064 pointers, in case they are 64-bit values.
2065
2066 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2067 test 2 if it fails.
2068
2069 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2070 and a #define of that name to empty if it is not externally set. This is to
2071 allow users of MSVC to set it if necessary.
2072
2073 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2074 the convenience functions in the pcre_get.c source file.
2075
2076 11. An option change at the start of a pattern that had top-level alternatives
2077 could cause overwriting and/or a crash. This command provoked a crash in
2078 some environments:
2079
2080 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2081
2082 This potential security problem was recorded as CVE-2008-2371.
2083
2084 12. For a pattern where the match had to start at the beginning or immediately
2085 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2086 pcre_dfa_exec() could read past the end of the passed subject if there was
2087 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2088 pcretest so that it places the subject at the end of its malloc-ed buffer.
2089
2090 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2091 exec() might read past the end of the data buffer in UTF-8 mode.
2092
2093 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2094 the data contained the byte 0x85 as part of a UTF-8 character within its
2095 first line. This applied both to normal and DFA matching.
2096
2097 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2098 /^[^d]*?$/8 failed to match "abc".
2099
2100 16. Added a missing copyright notice to pcrecpp_internal.h.
2101
2102 17. Make it more clear in the documentation that values returned from
2103 pcre_exec() in ovector are byte offsets, not character counts.
2104
2105 18. Tidied a few places to stop certain compilers from issuing warnings.
2106
2107 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2108 supplied by Stefan Weber. I made a further small update for 7.8 because
2109 there is a change of source arrangements: the pcre_searchfuncs.c module is
2110 replaced by pcre_ucd.c.
2111
2112
2113 Version 7.7 07-May-08
2114 ---------------------
2115
2116 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2117 a string to a long long, pretend we don't even have a long long." This is
2118 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2119
2120 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2121 pre-7.6 versions, which defined a global no_arg variable instead of putting
2122 it in the RE class. (See also #8 below.)
2123
2124 3. Remove a line of dead code, identified by coverity and reported by Nuno
2125 Lopes.
2126
2127 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2128
2129 (1) The include/exclude patterns were being applied to the whole pathnames
2130 of files, instead of just to the final components.
2131
2132 (2) If there was more than one level of directory, the subdirectories were
2133 skipped unless they satisfied the include/exclude conditions. This is
2134 inconsistent with GNU grep (and could even be seen as contrary to the
2135 pcregrep specification - which I improved to make it absolutely clear).
2136 The action now is always to scan all levels of directory, and just
2137 apply the include/exclude patterns to regular files.
2138
2139 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2140 --exclude_dir in the tests to avoid scanning .svn directories.
2141
2142 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2143 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2144 doesn't support NULs in patterns.
2145
2146 7. Added some missing "const"s to declarations of static tables in
2147 pcre_compile.c and pcre_dfa_exec.c.
2148
2149 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2150 caused by fix #2 above. (Subsequently also a second patch to fix the
2151 first patch. And a third patch - this was a messy problem.)
2152
2153 9. Applied Craig's patch to remove the use of push_back().
2154
2155 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2156 matching function regexec().
2157
2158 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2159 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2160 references. PCRE supports relative numbers with this syntax (I don't think
2161 Oniguruma does).
2162
2163 12. Previously, a group with a zero repeat such as (...){0} was completely
2164 omitted from the compiled regex. However, this means that if the group
2165 was called as a subroutine from elsewhere in the pattern, things went wrong
2166 (an internal error was given). Such groups are now left in the compiled
2167 pattern, with a new opcode that causes them to be skipped at execution
2168 time.
2169
2170 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2171 to the way PCRE behaves:
2172
2173 (a) A lone ] character is dis-allowed (Perl treats it as data).
2174
2175 (b) A back reference to an unmatched subpattern matches an empty string
2176 (Perl fails the current match path).
2177
2178 (c) A data ] in a character class must be notated as \] because if the
2179 first data character in a class is ], it defines an empty class. (In
2180 Perl it is not possible to have an empty class.) The empty class []
2181 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2182 The negative empty class [^] matches any one character, independently
2183 of the DOTALL setting.
2184
2185 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2186 non-existent subpattern following a character class starting with ']' and
2187 containing () gave an internal compiling error instead of "reference to
2188 non-existent subpattern". Fortunately, when the pattern did exist, the
2189 compiled code was correct. (When scanning forwards to check for the
2190 existence of the subpattern, it was treating the data ']' as terminating
2191 the class, so got the count wrong. When actually compiling, the reference
2192 was subsequently set up correctly.)
2193
2194 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2195 it was being rejected as not supported by pcre_dfa_exec(), even though
2196 other assertions are supported. I have made pcre_dfa_exec() support
2197 (*FAIL).
2198
2199 16. The implementation of 13c above involved the invention of a new opcode,
2200 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2201 cannot be changed at match time, I realized I could make a small
2202 improvement to matching performance by compiling OP_ALLANY instead of
2203 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2204 on the OP_ANY path.
2205
2206 17. Compiling pcretest on Windows with readline support failed without the
2207 following two fixes: (1) Make the unistd.h include conditional on
2208 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2209
2210 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2211 ncurses library to be included for pcretest when ReadLine support is
2212 requested, but also to allow for it to be overridden. This patch came from
2213 Daniel Bergström.
2214
2215 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2216 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2217 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2218 spotting this.
2219
2220
2221 Version 7.6 28-Jan-08
2222 ---------------------
2223
2224 1. A character class containing a very large number of characters with
2225 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2226 overflow.
2227
2228 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2229 HAVE_LONG_LONG is not defined.
2230
2231 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2232 bring them up to date and include new features. This patch includes:
2233
2234 - Fixed PH's badly added libz and libbz2 support.
2235 - Fixed a problem with static linking.
2236 - Added pcredemo. [But later removed - see 7 below.]
2237 - Fixed dftables problem and added an option.
2238 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2239 HAVE_LONG_LONG.
2240 - Added readline support for pcretest.
2241 - Added an listing of the option settings after cmake has run.
2242
2243 4. A user submitted a patch to Makefile that makes it easy to create
2244 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2245 Makefile.am that cause it to include this special target, without
2246 affecting anything else. Note that the same mingw target plus all
2247 the other distribution libraries and programs are now supported
2248 when configuring with CMake (see 6 below) instead of with
2249 Configure/Make.
2250
2251 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2252 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2253 exported in the Windows port". It has not yet been confirmed that the patch
2254 solves the problem, but it does no harm.
2255
2256 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2257 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2258 with CMake, and also correct the comment about stack recursion.
2259
2260 7. Remove the automatic building of pcredemo from the ./configure system and
2261 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2262 of a program that users should build themselves after PCRE is installed, so
2263 building it automatically is not really right. What is more, it gave
2264 trouble in some build environments.
2265
2266 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2267
2268
2269 Version 7.5 10-Jan-08
2270 ---------------------
2271
2272 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2273 values in parens when parsing an RE using the C++ wrapper."
2274
2275 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2276 Characters greater than 255 were excluded from the class instead of being
2277 included.
2278
2279 3. The same bug as (2) above applied to negated POSIX classes such as
2280 [:^space:].
2281
2282 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2283 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2284 I have changed it.
2285
2286 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2287 first named subpattern) and a construct such as (?&a) would reference the
2288 first named subpattern whose name started with "a" (in other words, the
2289 length check was missing). Both these problems are fixed. "Subpattern name
2290 expected" is now given for (?&) (a zero-length name), and this patch also
2291 makes it give the same error for \k'' (previously it complained that that
2292 was a reference to a non-existent subpattern).
2293
2294 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2295 this is right because (?- can be followed by option settings as well as by
2296 digits. I have, however, made the messages clearer.
2297
2298 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2299 than the number used in the conditional) now cause a compile-time error.
2300 This is actually not compatible with Perl, which accepts such patterns, but
2301 treats the conditional as always being FALSE (as PCRE used to), but it
2302 seems to me that giving a diagnostic is better.
2303
2304 8. Change "alphameric" to the more common word "alphanumeric" in comments
2305 and messages.
2306
2307 9. Fix two occurrences of "backslash" in comments that should have been
2308 "backspace".
2309
2310 10. Remove two redundant lines of code that can never be obeyed (their function
2311 was moved elsewhere).
2312
2313 11. The program that makes PCRE's Unicode character property table had a bug
2314 which caused it to generate incorrect table entries for sequences of
2315 characters that have the same character type, but are in different scripts.
2316 It amalgamated them into a single range, with the script of the first of
2317 them. In other words, some characters were in the wrong script. There were
2318 thirteen such cases, affecting characters in the following ranges:
2319
2320 U+002b0 - U+002c1
2321 U+0060c - U+0060d
2322 U+0061e - U+00612
2323 U+0064b - U+0065e
2324 U+0074d - U+0076d
2325 U+01800 - U+01805
2326 U+01d00 - U+01d77
2327 U+01d9b - U+01dbf
2328 U+0200b - U+0200f
2329 U+030fc - U+030fe
2330 U+03260 - U+0327f
2331 U+0fb46 - U+0fbb1
2332 U+10450 - U+1049d
2333
2334 12. The -o option (show only the matching part of a line) for pcregrep was not
2335 compatible with GNU grep in that, if there was more than one match in a
2336 line, it showed only the first of them. It now behaves in the same way as
2337 GNU grep.
2338
2339 13. If the -o and -v options were combined for pcregrep, it printed a blank
2340 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2341 does the same. The return code can be used to tell if there were any
2342 non-matching lines.
2343
2344 14. Added --file-offsets and --line-offsets to pcregrep.
2345
2346 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2347 infinitely looping recursion. The bug was that positive lookaheads were not
2348 being skipped when checking for a possible empty match (negative lookaheads
2349 and both kinds of lookbehind were skipped).
2350
2351 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2352 inclusion of <windows.h> to before rather than after the definition of
2353 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2354
2355 17. Specifying a possessive quantifier with a specific limit for a Unicode
2356 character property caused pcre_compile() to compile bad code, which led at
2357 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2358 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2359 caused the error; without that there was no problem.
2360
2361 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2362
2363 19. Added --enable-pcretest-libreadline.
2364
2365 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2366 RE::GlobalReplace(). As a result, the number of replacements returned was
2367 double what it should be. I removed one of the increments, but Craig sent a
2368 later patch that removed the other one (the right fix) and added unit tests
2369 that check the return values (which was not done before).
2370
2371 21. Several CMake things:
2372
2373 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2374 the names libpcre and libpcreposix, not just pcre and pcreposix.
2375
2376 (2) The above change means that pcretest and pcregrep are now correctly
2377 linked with the newly-built libraries, not previously installed ones.
2378
2379 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2380
2381 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2382 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2383 UTF-8 newline character). The key issue is that the pattern starts .*;
2384 this means that the match must be either at the beginning, or after a
2385 newline. The bug was in the code for advancing after a failed match and
2386 checking that the new position followed a newline. It was not taking
2387 account of UTF-8 characters correctly.
2388
2389 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2390 character classes. PCRE was not treating the sequence [:...:] as a
2391 character class unless the ... were all letters. Perl, however, seems to
2392 allow any characters between [: and :], though of course it rejects as
2393 unknown any "names" that contain non-letters, because all the known class
2394 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2395 for example, whereas PCRE did not - it did not recognize a POSIX character
2396 class. This seemed a bit dangerous, so the code has been changed to be
2397 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2398 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2399 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2400 Perl does, and where it didn't before.
2401
2402 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2403 Windows environments %n is disabled by default.
2404
2405
2406 Version 7.4 21-Sep-07
2407 ---------------------
2408
2409 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2410 means that a class such as [\s] counted as "explicit reference to CR or
2411 LF". That isn't really right - the whole point of the change was to try to
2412 help when there was an actual mention of one of the two characters. So now
2413 the change happens only if \r or \n (or a literal CR or LF) character is
2414 encountered.
2415
2416 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2417 of both had grown to the point where there were only 3 bits left.
2418 Fortunately, there was spare space in the data structure, and so I have
2419 moved the internal flags into a new 16-bit field to free up more option
2420 bits.
2421
2422 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2423 but did not set the internal JCHANGED flag - either of these is enough to
2424 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2425 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2426 start sets both bits.
2427
2428 4. Added options (at build time, compile time, exec time) to change \R from
2429 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2430
2431 5. doc/pcresyntax.html was missing from the distribution.
2432
2433 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2434 compatibility, even though it is no longer used.
2435
2436 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2437 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2438 windows.h file is present (where different names are used). [This was
2439 reversed later after testing - see 16 below.]
2440
2441 8. Changed all #include <config.h> to #include "config.h". There were also
2442 some further <pcre.h> cases that I changed to "pcre.h".
2443
2444 9. When pcregrep was used with the --colour option, it missed the line ending
2445 sequence off the lines that it output.
2446
2447 10. It was pointed out to me that arrays of string pointers cause lots of
2448 relocations when a shared library is dynamically loaded. A technique of
2449 using a single long string with a table of offsets can drastically reduce
2450 these. I have refactored PCRE in four places to do this. The result is
2451 dramatic:
2452
2453 Originally: 290
2454 After changing UCP table: 187
2455 After changing error message table: 43
2456 After changing table of "verbs" 36
2457 After changing table of Posix names 22
2458
2459 Thanks to the folks working on Gregex for glib for this insight.
2460
2461 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2462 unicode-properties was also set.
2463
2464 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2465
2466 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2467 checked only for CRLF.
2468
2469 14. Added casts to pcretest.c to avoid compiler warnings.
2470
2471 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2472
2473 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2474 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2475 entirely. This removes changes made in 7 above.
2476
2477 17. The CMake files have been updated, and there is now more information about
2478 building with CMake in the NON-UNIX-USE document.
2479
2480
2481 Version 7.3 28-Aug-07
2482 ---------------------
2483
2484 1. In the rejigging of the build system that eventually resulted in 7.1, the
2485 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2486 brackets there is not right, since it causes compilers to look for an
2487 installed pcre.h, not the version that is in the source that is being
2488 compiled (which of course may be different). I have changed it back to:
2489
2490 #include "pcre.h"
2491
2492 I have a vague recollection that the change was concerned with compiling in
2493 different directories, but in the new build system, that is taken care of
2494 by the VPATH setting the Makefile.
2495
2496 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2497 when the subject happened to end in the byte 0x85 (e.g. if the last
2498 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2499 characters but of course it shouldn't be taken as a newline when it is part
2500 of another character. The bug was that, for an unlimited repeat of . in
2501 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2502 characters when looking for a newline.
2503
2504 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2505
2506 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2507 in debug output.
2508
2509 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2510 long printing in the pcrecpp unittest when running under MinGW.
2511
2512 6. ESC_K was left out of the EBCDIC table.
2513
2514 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2515 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2516 limit also applies to "virtual nesting" when a pattern is recursive, and in
2517 this case 1000 isn't so big. I have been able to remove this limit at the
2518 expense of backing off one optimization in certain circumstances. Normally,
2519 when pcre_exec() would call its internal match() function recursively and
2520 immediately return the result unconditionally, it uses a "tail recursion"
2521 feature to save stack. However, when a subpattern that can match an empty
2522 string has an unlimited repetition quantifier, it no longer makes this
2523 optimization. That gives it a stack frame in which to save the data for
2524 checking that an empty string has been matched. Previously this was taken
2525 from the 1000-entry workspace that had been reserved. So now there is no
2526 explicit limit, but more stack is used.
2527
2528 8. Applied Daniel's patches to solve problems with the import/export magic
2529 syntax that is required for Windows, and which was going wrong for the
2530 pcreposix and pcrecpp parts of the library. These were overlooked when this
2531 problem was solved for the main library.
2532
2533 9. There were some crude static tests to avoid integer overflow when computing
2534 the size of patterns that contain repeated groups with explicit upper
2535 limits. As the maximum quantifier is 65535, the maximum group length was
2536 set at 30,000 so that the product of these two numbers did not overflow a
2537 32-bit integer. However, it turns out that people want to use groups that
2538 are longer than 30,000 bytes (though not repeat them that many times).
2539 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2540 made it possible to implement the integer overflow checks in a much more
2541 dynamic way, which I have now done. The artificial limitation on group
2542 length has been removed - we now have only the limit on the total length of
2543 the compiled pattern, which depends on the LINK_SIZE setting.
2544
2545 10. Fixed a bug in the documentation for get/copy named substring when
2546 duplicate names are permitted. If none of the named substrings are set, the
2547 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2548 empty string.
2549
2550 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2551 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2552 because the ] is interpreted as the first data character and the
2553 terminating ] is not found. PCRE has been made compatible with Perl in this
2554 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2555 cause memory overwriting.
2556
2557 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2558 string has been matched (to stop an infinite loop). It was not recognizing
2559 a conditional subpattern that could match an empty string if that
2560 subpattern was within another subpattern. For example, it looped when
2561 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2562 condition was not nested. This bug has been fixed.
2563
2564 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2565 past the start of the subject in the presence of bytes with the top bit
2566 set, for example "\x8aBCD".
2567
2568 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2569 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2570
2571 14. Optimized (?!) to (*FAIL).
2572
2573 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2574 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2575 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2576 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2577 does: it's just the validity check that is more restrictive.
2578
2579 16. Inserted checks for integer overflows during escape sequence (backslash)
2580 processing, and also fixed erroneous offset values for syntax errors during
2581 backslash processing.
2582
2583 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2584 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2585
2586 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2587 caused an overrun.
2588
2589 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2590 something other than just ASCII characters) inside a group that had an
2591 unlimited repeat caused a loop at compile time (while checking to see
2592 whether the group could match an empty string).
2593
2594 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2595 [\P{Any}] did so. (Error in the code for printing property names.)
2596
2597 21. An orphan \E inside a character class could cause a crash.
2598
2599 22. A repeated capturing bracket such as (A)? could cause a wild memory
2600 reference during compilation.
2601
2602 23. There are several functions in pcre_compile() that scan along a compiled
2603 expression for various reasons (e.g. to see if it's fixed length for look
2604 behind). There were bugs in these functions when a repeated \p or \P was
2605 present in the pattern. These operators have additional parameters compared
2606 with \d, etc, and these were not being taken into account when moving along
2607 the compiled data. Specifically:
2608
2609 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2610 length.
2611
2612 (b) An item such as \pL+ within a repeated group could cause crashes or
2613 loops.
2614
2615 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2616 "reference to non-existent subpattern" error.
2617
2618 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2619
2620 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2621 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2622
2623 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2624
2625 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2626 character were causing crashes (broken optimization).
2627
2628 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2629 \p or \P) caused a compile-time loop.
2630
2631 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2632 break. For example, the unstudied pattern [\r\n]A does not match the string
2633 "\r\nA" because change 7.0/46 below moves the current point on by two
2634 characters after failing to match at the start. However, the pattern \nA
2635 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2636 the same is true. There doesn't seem any very clean way out of this, but
2637 what I have chosen to do makes the common cases work: PCRE now takes note
2638 of whether there can be an explicit match for \r or \n anywhere in the
2639 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2640 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2641 pattern has explicit CR or LF references.
2642
2643 29. Added (*CR) etc for changing newline setting at start of pattern.
2644
2645
2646 Version 7.2 19-Jun-07
2647 ---------------------
2648
2649 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2650 which is apparently normally available under Windows.
2651
2652 2. Re-jig the pcregrep tests with different newline settings in an attempt
2653 to make them independent of the local environment's newline setting.
2654
2655 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2656
2657 4. Some of the "internals" tests were previously cut out when the link size
2658 was not 2, because the output contained actual offsets. The recent new
2659 "Z" feature of pcretest means that these can be cut out, making the tests
2660 usable with all link sizes.
2661
2662 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2663 stack recursion. This gives a massive performance boost under BSD, but just
2664 a small improvement under Linux. However, it saves one field in the frame
2665 in all cases.
2666
2667 6. Added more features from the forthcoming Perl 5.10:
2668
2669 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2670 recursion call. It refers to the nth most recently opened parentheses.
2671
2672 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2673 to be opened parentheses.
2674
2675 (c) Conditions that refer to capturing parentheses can be specified
2676 relatively, for example, (?(-2)... or (?(+3)...
2677
2678 (d) \K resets the start of the current match so that everything before
2679 is not part of it.
2680
2681 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2682
2683 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2684 reference syntax.
2685
2686 (g) (?| introduces a group in which the numbering of parentheses in each
2687 alternative starts with the same number.
2688
2689 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2690
2691 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2692 PCRE_INFO_JCHANGED.
2693
2694 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2695 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2696 for detecting groups that can match an empty string.
2697
2698 9. A pattern with a very large number of alternatives (more than several
2699 hundred) was running out of internal workspace during the pre-compile
2700 phase, where pcre_compile() figures out how much memory will be needed. A
2701 bit of new cunning has reduced the workspace needed for groups with
2702 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2703 workspace instead of running out of the 4096 that are available.
2704
2705 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2706
2707 11. Applied patch from Google to remove an optimization that didn't quite work.
2708 The report of the bug said:
2709
2710 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2711 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2712 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2713
2714 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2715 it matched the wrong number of bytes.
2716
2717
2718 Version 7.1 24-Apr-07
2719 ---------------------
2720
2721 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2722 that is more "standard", making use of automake and other Autotools. There
2723 is some re-arrangement of the files and adjustment of comments consequent
2724 on this.
2725
2726 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2727 for recursive directory scanning broke on some systems because the files
2728 are not scanned in any specific order and on different systems the order
2729 was different. A call to "sort" has been inserted into RunGrepTest for the
2730 approprate test as a short-term fix. In the longer term there may be an
2731 alternative.
2732
2733 3. I had an email from Eric Raymond about problems translating some of PCRE's
2734 man pages to HTML (despite the fact that I distribute HTML pages, some
2735 people do their own conversions for various reasons). The problems
2736 concerned the use of low-level troff macros .br and .in. I have therefore
2737 removed all such uses from the man pages (some were redundant, some could
2738 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2739 HTML has been updated to handle .nf/.fi and to complain if it encounters
2740 .br or .in.
2741
2742 4. Updated comments in configure.ac that get placed in config.h.in and also
2743 arranged for config.h to be included in the distribution, with the name
2744 config.h.generic, for the benefit of those who have to compile without
2745 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2746
2747 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2748 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2749 makevp.bat for the current PCRE, using the additional files
2750 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2751
2752 6. A Windows user reported a minor discrepancy with test 2, which turned out
2753 to be caused by a trailing space on an input line that had got lost in his
2754 copy. The trailing space was an accident, so I've just removed it.
2755
2756 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2757 that is needed.
2758
2759 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2760 as "const" (a) because they are and (b) because it helps the PHP
2761 maintainers who have recently made a script to detect big data structures
2762 in the php code that should be moved to the .rodata section. I remembered
2763 to update Builducptable as well, so it won't revert if ucptable.h is ever
2764 re-created.
2765
2766 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2767 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2768 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2769 support is not required. This saves 1.5-2K of code, which is important in
2770 some applications.
2771
2772 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2773 so as not to refer to the tables, even though these functions will never be
2774 called when UTF-8 support is disabled. Otherwise there are problems with a
2775 shared library.
2776
2777 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2778
2779 (a) It was defining its arguments as char * instead of void *.
2780
2781 (b) It was assuming that all moves were upwards in memory; this was true
2782 a long time ago when I wrote it, but is no longer the case.
2783
2784 The emulated memove() is provided for those environments that have neither
2785 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2786 is clearly not the case, as these two bugs were recently reported.
2787
2788 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2789 and Detrail to create the HTML documentation, the .txt form of the man
2790 pages, and it removes trailing spaces from listed files. It also creates
2791 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2792 case, it wraps all the #defines with #ifndefs. This script should be run
2793 before "make dist".
2794
2795 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2796 with Unicode property support.
2797
2798 (a) For a maximizing quantifier, if the two different cases of the
2799 character were of different lengths in their UTF-8 codings (there are
2800 some cases like this - I found 11), and the matching function had to
2801 back up over a mixture of the two cases, it incorrectly assumed they
2802 were both the same length.
2803
2804 (b) When PCRE was configured to use the heap rather than the stack for
2805 recursion during matching, it was not correctly preserving the data for
2806 the other case of a UTF-8 character when checking ahead for a match
2807 while processing a minimizing repeat. If the check also involved
2808 matching a wide character, but failed, corruption could cause an
2809 erroneous result when trying to check for a repeat of the original
2810 character.
2811
2812 13. Some tidying changes to the testing mechanism:
2813
2814 (a) The RunTest script now detects the internal link size and whether there
2815 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2816 values substituted by "configure". (The RunGrepTest script already did
2817 this for UTF-8.) The configure.ac script no longer substitutes the
2818 relevant variables.
2819
2820 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2821 with length and offset values. This means that the output is different
2822 for different internal link sizes. Test 2 is skipped for link sizes
2823 other than 2 because of this, bypassing the problem. Unfortunately,
2824 there was also a test in test 3 (the locale tests) that used /B and
2825 failed for link sizes other than 2. Rather than cut the whole test out,
2826 I have added a new /Z option to pcretest that replaces the length and
2827 offset values with spaces. This is now used to make test 3 independent
2828 of link size. (Test 2 will be tidied up later.)
2829
2830 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2831 segmentation fault instead of returning the appropriate error message.
2832
2833 15. In multiline mode when the newline sequence was set to "any", the pattern
2834 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2835 This doesn't seem right; it now treats the CRLF combination as the line
2836 ending, and so does not match in that case. It's only a pattern such as ^$
2837 that would hit this one: something like ^ABC$ would have failed after \r
2838 and then tried again after \r\n.
2839
2840 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2841 in an attempt to make files that differ only in their line terminators
2842 compare equal. This works on Linux.
2843
2844 17. Under certain error circumstances pcregrep might try to free random memory
2845 as it exited. This is now fixed, thanks to valgrind.
2846
2847 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2848 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2849 was because its rules for how to advance for /g after matching an empty
2850 string at the end of a line did not allow for this case. They now check for
2851 it specially.
2852
2853 20. pcretest is supposed to handle patterns and data of any length, by
2854 extending its buffers when necessary. It was getting this wrong when the
2855 buffer for a data line had to be extended.
2856
2857 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2858 CRLF as a newline sequence.
2859
2860 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2861 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2862 I have nevertheless tidied it up.
2863
2864 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2865
2866 24. Added a man page for pcre-config.
2867
2868
2869 Version 7.0 19-Dec-06
2870 ---------------------
2871
2872 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2873 moving to gcc 4.1.1.
2874
2875 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2876 sys/time.h, which is documented as needed for this function. It doesn't
2877 seem to matter on Linux, but it showed up on some releases of OS X.
2878
2879 3. It seems that there are systems where bytes whose values are greater than
2880 127 match isprint() in the "C" locale. The "C" locale should be the
2881 default when a C program starts up. In most systems, only ASCII printing
2882 characters match isprint(). This difference caused the output from pcretest
2883 to vary, making some of the tests fail. I have changed pcretest so that:
2884
2885 (a) When it is outputting text in the compiled version of a pattern, bytes
2886 other than 32-126 are always shown as hex escapes.
2887
2888 (b) When it is outputting text that is a matched part of a subject string,
2889 it does the same, unless a different locale has been set for the match
2890 (using the /L modifier). In this case, it uses isprint() to decide.
2891
2892 4. Fixed a major bug that caused incorrect computation of the amount of memory
2893 required for a compiled pattern when options that changed within the
2894 pattern affected the logic of the preliminary scan that determines the
2895 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2896 that the computed length was too small. The symptoms of this bug were
2897 either the PCRE error "internal error: code overflow" from pcre_compile(),
2898 or a glibc crash with a message such as "pcretest: free(): invalid next
2899 size (fast)". Examples of patterns that provoked this bug (shown in
2900 pcretest format) are:
2901
2902 /(?-x: )/x
2903 /(?x)(?-x: \s*#\s*)/
2904 /((?i)[\x{c0}])/8
2905 /(?i:[\x{c0}])/8
2906
2907 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2908 is now done differently.
2909
2910 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2911 wrapper classes; (b) implement a new function in the C++ scanner that is
2912 more efficient than the old way of doing things because it avoids levels of
2913 recursion in the regex matching; (c) add a paragraph to the documentation
2914 for the FullMatch() function.
2915
2916 6. The escape sequence \n was being treated as whatever was defined as
2917 "newline". Not only was this contrary to the documentation, which states
2918 that \n is character 10 (hex 0A), but it also went horribly wrong when
2919 "newline" was defined as CRLF. This has been fixed.
2920
2921 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2922 was being set to -1 for the "end of line" case (supposedly a value that no
2923 character can have). Though this value is never used (the check for end of
2924 line is "zero bytes in current character"), it caused compiler complaints.
2925 I've changed it to 0xffffffff.
2926
2927 8. In pcre_version.c, the version string was being built by a sequence of
2928 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2929 string (as it is for production releases) called a macro with an empty
2930 argument. The C standard says the result of this is undefined. The gcc
2931 compiler treats it as an empty string (which was what was wanted) but it is
2932 reported that Visual C gives an error. The source has been hacked around to
2933 avoid this problem.
2934
2935 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2936 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2937 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2938 of them did).
2939
2940 10. Originally, pcretest opened its input and output without "b"; then I was
2941 told that "b" was needed in some environments, so it was added for release
2942 5.0 to both the input and output. (It makes no difference on Unix-like
2943 systems.) Later I was told that it is wrong for the input on Windows. I've
2944 now abstracted the modes into two macros, to make it easier to fiddle with
2945 them, and removed "b" from the input mode under Windows.
2946
2947 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2948
2949 12. Added -help and --help to pcretest as an official way of being reminded
2950 of the options.
2951
2952 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2953 and pcrecpp.cc because they annoy compilers at high warning levels.
2954
2955 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2956
2957 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2958 scripts are not C programs :-) and which was not noticed because it works
2959 on Linux.
2960
2961 16. pcretest is supposed to handle any length of pattern and data line (as one
2962 line or as a continued sequence of lines) by extending its input buffer if
2963 necessary. This feature was broken for very long pattern lines, leading to
2964 a string of junk being passed to pcre_compile() if the pattern was longer
2965 than about 50K.
2966
2967 17. I have done a major re-factoring of the way pcre_compile() computes the
2968 amount of memory needed for a compiled pattern. Previously, there was code
2969 that made a preliminary scan of the pattern in order to do this. That was
2970 OK when PCRE was new, but as the facilities have expanded, it has become
2971 harder and harder to keep it in step with the real compile phase, and there
2972 have been a number of bugs (see for example, 4 above). I have now found a
2973 cunning way of running the real compile function in a "fake" mode that
2974 enables it to compute how much memory it would need, while actually only
2975 ever using a few hundred bytes of working memory and without too many
2976 tests of the mode. This should make future maintenance and development
2977 easier. A side effect of this work is that the limit of 200 on the nesting
2978 depth of parentheses has been removed (though this was never a serious
2979 limitation, I suspect). However, there is a downside: pcre_compile() now
2980 runs more slowly than before (30% or more, depending on the pattern). I
2981 hope this isn't a big issue. There is no effect on runtime performance.
2982
2983 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2984 newline (only possible for the last line of a file) and it was a
2985 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2986
2987 19. Added additional timing features to pcretest. (1) The -tm option now times
2988 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2989 separate command line item, by a number that specifies the number of
2990 repeats to use when timing. The default is 50000; this gives better
2991 precision, but takes uncomfortably long for very large patterns.
2992
2993 20. Extended pcre_study() to be more clever in cases where a branch of a
2994 subpattern has no definite first character. For example, (a*|b*)[cd] would
2995 previously give no result from pcre_study(). Now it recognizes that the
2996 first character must be a, b, c, or d.
2997
2998 21. There was an incorrect error "recursive call could loop indefinitely" if
2999 a subpattern (or the entire pattern) that was being tested for matching an
3000 empty string contained only one non-empty item after a nested subpattern.
3001 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3002 incorrectly, because the \d was being skipped in the check.
3003
3004 22. The pcretest program now has a new pattern option /B and a command line
3005 option -b, which is equivalent to adding /B to every pattern. This causes
3006 it to show the compiled bytecode, without the additional information that
3007 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3008 is the same as /B/I).
3009
3010 23. A new optimization is now able automatically to treat some sequences such
3011 as a*b as a*+b. More specifically, if something simple (such as a character
3012 or a simple class like \d) has an unlimited quantifier, and is followed by
3013 something that cannot possibly match the quantified thing, the quantifier
3014 is automatically "possessified".
3015
3016 24. A recursive reference to a subpattern whose number was greater than 39
3017 went wrong under certain circumstances in UTF-8 mode. This bug could also
3018 have affected the operation of pcre_study().
3019
3020 25. Realized that a little bit of performance could be had by replacing
3021 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3022
3023 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3024
3025 27. Possessive quantifiers such as a++ were previously implemented by turning
3026 them into atomic groups such as ($>a+). Now they have their own opcodes,
3027 which improves performance. This includes the automatically created ones
3028 from 23 above.
3029
3030 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3031 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3032 the first matched character to be a colon. This applied both to named and
3033 numbered groups.
3034
3035 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3036
3037 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3038 building PCRE on Windows easier, so I have included it in the distribution.
3039
3040 31. There is now a check in pcretest against a ridiculously large number being
3041 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3042 loop, the loop is abandoned.
3043
3044 32. Forward references to subpatterns in conditions such as (?(2)...) where
3045 subpattern 2 is defined later cause pcre_compile() to search forwards in
3046 the pattern for the relevant set of parentheses. This search went wrong
3047 when there were unescaped parentheses in a character class, parentheses
3048 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3049
3050 33. "Subroutine" calls and backreferences were previously restricted to
3051 referencing subpatterns earlier in the regex. This restriction has now
3052 been removed.
3053
3054 34. Added a number of extra features that are going to be in Perl 5.10. On the
3055 whole, these are just syntactic alternatives for features that PCRE had
3056 previously implemented using the Python syntax or my own invention. The
3057 other formats are all retained for compatibility.
3058
3059 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3060 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3061 also .NET compatible.
3062
3063 (b) A recursion or subroutine call to a named group can now be defined as
3064 (?&name) as well as (?P>name).
3065
3066 (c) A backreference to a named group can now be defined as \k<name> or
3067 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3068 5.10, are also .NET compatible.
3069
3070 (d) A conditional reference to a named group can now use the syntax
3071 (?(<name>) or (?('name') as well as (?(name).
3072
3073 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3074 groups (named and numbered) that are never evaluated inline, but can be
3075 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3076 is always false. There may be only one alternative in such a group.
3077
3078 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3079 as the simple (?(R). The condition is true only if the most recent
3080 recursion is that of the given number or name. It does not search out
3081 through the entire recursion stack.
3082
3083 (g) The escape \gN or \g{N} has been added, where N is a positive or
3084 negative number, specifying an absolute or relative reference.
3085
3086 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3087 some "unreachable code" warnings.
3088
3089 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3090 things, this adds five new scripts.
3091
3092 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3093 There were also incompatibilities regarding the handling of \Q..\E inside
3094 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3095 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3096
3097 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3098 matches an empty string, and forcibly breaks the loop. There were bugs in
3099 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3100 against aaaa the result was just "a" rather than "aaaa", for example. Two
3101 separate and independent bugs (that affected different cases) have been
3102 fixed.
3103
3104 39. Refactored the code to abolish the use of different opcodes for small
3105 capturing bracket numbers. This is a tidy that I avoided doing when I
3106 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3107 The new approach is not only tidier, it makes it possible to reduce the
3108 memory needed to fix the previous bug (38).
3109
3110 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3111 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3112 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3113 mode.
3114
3115 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3116 report.
3117
3118 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3119 copy construction and assignment in the C++ wrapper.
3120
3121 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3122 couple of bugs that could have given wrong results in the "--newline=crlf"
3123 case.
3124
3125 44. Added a number of casts and did some reorganization of signed/unsigned int
3126 variables following suggestions from Dair Grant. Also renamed the variable
3127 "this" as "item" because it is a C++ keyword.
3128
3129 45. Arranged for dftables to add
3130
3131 #include "pcre_internal.h"
3132
3133 to pcre_chartables.c because without it, gcc 4.x may remove the array
3134 definition from the final binary if PCRE is built into a static library and
3135 dead code stripping is activated.
3136
3137 46. For an unanchored pattern, if a match attempt fails at the start of a
3138 newline sequence, and the newline setting is CRLF or ANY, and the next two
3139 characters are CRLF, advance by two characters instead of one.
3140
3141
3142 Version 6.7 04-Jul-06
3143 ---------------------
3144
3145 1. In order to handle tests when input lines are enormously long, pcretest has
3146 been re-factored so that it automatically extends its buffers when
3147 necessary. The code is crude, but this _is_ just a test program. The
3148 default size has been increased from 32K to 50K.
3149
3150 2. The code in pcre_study() was using the value of the re argument before
3151 testing it for NULL. (Of course, in any sensible call of the function, it
3152 won't be NULL.)
3153
3154 3. The memmove() emulation function in pcre_internal.h, which is used on
3155 systems that lack both memmove() and bcopy() - that is, hardly ever -
3156 was missing a "static" storage class specifier.
3157
3158 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3159 containing an extended class (one that cannot be represented by a bitmap
3160 because it contains high-valued characters or Unicode property items, e.g.
3161 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3162 pattern, but PCRE should not loop if you do not (it no longer does).
3163 [Detail: two cases were found: (a) a repeated subpattern containing an
3164 extended class; (b) a recursive reference to a subpattern that followed a
3165 previous extended class. It wasn't skipping over the extended class
3166 correctly when UTF-8 mode was not set.]
3167
3168 5. A negated single-character class was not being recognized as fixed-length
3169 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3170 compile error "lookbehind assertion is not fixed length".
3171
3172 6. The RunPerlTest auxiliary script was showing an unexpected difference
3173 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3174 write a Perl script that can interpret lines of an input file either as
3175 byte characters or as UTF-8, which is what "perltest" was being required to
3176 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3177 can't do is switch easily at run time between having the "use utf8;" pragma
3178 or not. In the end, I fudged it by using the RunPerlTest script to insert
3179 "use utf8;" explicitly for the UTF-8 tests.
3180
3181 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3182 the end of the subject string, contrary to the documentation and to what
3183 Perl does. This was true of both matching functions. Now it matches only at
3184 the start of the subject and immediately after *internal* newlines.
3185
3186 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3187 a pointer to an int instead of a pointer to an unsigned long int. This
3188 caused problems on 64-bit systems.
3189
3190 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3191 instance of the 'standard' template library not being so standard".
3192
3193 10. There was no check on the number of named subpatterns nor the maximum
3194 length of a subpattern name. The product of these values is used to compute
3195 the size of the memory block for a compiled pattern. By supplying a very
3196 long subpattern name and a large number of named subpatterns, the size
3197 computation could be caused to overflow. This is now prevented by limiting
3198 the length of names to 32 characters, and the number of named subpatterns
3199 to 10,000.
3200
3201 11. Subpatterns that are repeated with specific counts have to be replicated in
3202 the compiled pattern. The size of memory for this was computed from the
3203 length of the subpattern and the repeat count. The latter is limited to
3204 65535, but there was no limit on the former, meaning that integer overflow
3205 could in principle occur. The compiled length of a repeated subpattern is
3206 now limited to 30,000 bytes in order to prevent this.
3207
3208 12. Added the optional facility to have named substrings with the same name.
3209
3210 13. Added the ability to use a named substring as a condition, using the
3211 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3212 are numbers (not recommended). Forward references are permitted.
3213
3214 14. Added forward references in named backreferences (if you see what I mean).
3215
3216 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3217 pattern could run off the end of the subject. For example, the pattern
3218 "(?s)(.{1,5})"8 did this with the subject "ab".
3219
3220 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3221 PCRE_CASELESS was set when matching characters that were quantified with ?
3222 or *.
3223
3224 17. A character class other than a single negated character that had a minimum
3225 but no maximum quantifier - for example [ab]{6,} - was not handled
3226 correctly by pce_dfa_exec(). It would match only one character.
3227
3228 18. A valid (though odd) pattern that looked like a POSIX character
3229 class but used an invalid character after [ (for example [[,abc,]]) caused
3230 pcre_compile() to give the error "Failed: internal error: code overflow" or
3231 in some cases to crash with a glibc free() error. This could even happen if
3232 the pattern terminated after [[ but there just happened to be a sequence of
3233 letters, a binary zero, and a closing ] in the memory that followed.
3234
3235 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3236 over the years. Originally (before any Unicode support), just the bottom 8
3237 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3238 output from "man perlunicode" includes this:
3239
3240 The regular expression compiler produces polymorphic opcodes. That
3241 is, the pattern adapts to the data and automatically switches to
3242 the Unicode character scheme when presented with Unicode data--or
3243 instead uses a traditional byte scheme when presented with byte
3244 data.
3245
3246 Sadly, a wide octal escape does not cause a switch, and in a string with
3247 no other multibyte characters, these octal escapes are treated as before.
3248 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3249 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3250 Unicode string.
3251
3252 I have not perpetrated such confusion in PCRE. Up till now, it took just
3253 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3254 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3255 translate to the appropriate multibyte character.
3256
3257 29. Applied some refactoring to reduce the number of warnings from Microsoft
3258 and Borland compilers. This has included removing the fudge introduced
3259 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3260 a warning about an unused variable.
3261
3262 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3263 characters since release 4.0, because Perl (from release 5.004) does not.
3264 [Or at least, is documented not to: some releases seem to be in conflict
3265 with the documentation.] However, when a pattern was studied with
3266 pcre_study() and all its branches started with \s, PCRE still included VT
3267 as a possible starting character. Of course, this did no harm; it just
3268 caused an unnecessary match attempt.
3269
3270 22. Removed a now-redundant internal flag bit that recorded the fact that case
3271 dependency changed within the pattern. This was once needed for "required
3272 byte" processing, but is no longer used. This recovers a now-scarce options
3273 bit. Also moved the least significant internal flag bit to the most-
3274 significant bit of the word, which was not previously used (hangover from
3275 the days when it was an int rather than a uint) to free up another bit for
3276 the future.
3277
3278 23. Added support for CRLF line endings as well as CR and LF. As well as the
3279 default being selectable at build time, it can now be changed at runtime
3280 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3281 specify that it is scanning data with non-default line endings.
3282
3283 24. Changed the definition of CXXLINK to make it agree with the definition of
3284 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3285
3286 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3287 recursions. This makes a big different to stack usage for some patterns.
3288
3289 26. If a subpattern containing a named recursion or subroutine reference such
3290 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3291 the space required for the compiled pattern went wrong and gave too small a
3292 value. Depending on the environment, this could lead to "Failed: internal
3293 error: code overflow at offset 49" or "glibc detected double free or
3294 corruption" errors.
3295
3296 27. Applied patches from Google (a) to support the new newline modes and (b) to
3297 advance over multibyte UTF-8 characters in GlobalReplace.
3298
3299 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3300 difference for some implementation of PCRE in some Windows version.
3301
3302 29. Added some extra testing facilities to pcretest:
3303
3304 \q<number> in a data line sets the "match limit" value
3305 \Q<number> in a data line sets the "match recursion limt" value
3306 -S <number> sets the stack size, where <number> is in megabytes
3307
3308 The -S option isn't available for Windows.
3309
3310
3311 Version 6.6 06-Feb-06
3312 ---------------------
3313
3314 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3315 in pcreposix.h. I have copied the definition from pcre.h.
3316
3317 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3318 because pcre.h is no longer a built file.
3319
3320 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3321 not normally included in the compiled code.
3322
3323
3324 Version 6.5 01-Feb-06
3325 ---------------------
3326
3327 1. When using the partial match feature with pcre_dfa_exec(), it was not
3328 anchoring the second and subsequent partial matches at the new starting
3329 point. This could lead to incorrect results. For example, with the pattern
3330 /1234/, partially matching against "123" and then "a4" gave a match.
3331
3332 2. Changes to pcregrep:
3333
3334 (a) All non-match returns from pcre_exec() were being treated as failures
3335 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3336 error message is output. Some extra information is given for the
3337 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3338 probably the only errors that are likely to be caused by users (by
3339 specifying a regex that has nested indefinite repeats, for instance).
3340 If there are more than 20 of these errors, pcregrep is abandoned.
3341
3342 (b) A binary zero was treated as data while matching, but terminated the
3343 output line if it was written out. This has been fixed: binary zeroes
3344 are now no different to any other data bytes.
3345
3346 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3347 used to set a locale for matching. The --locale=xxxx long option has
3348 been added (no short equivalent) to specify a locale explicitly on the
3349 pcregrep command, overriding the environment variables.
3350
3351 (d) When -B was used with -n, some line numbers in the output were one less
3352 than they should have been.
3353
3354 (e) Added the -o (--only-matching) option.
3355
3356 (f) If -A or -C was used with -c (count only), some lines of context were
3357 accidentally printed for the final match.
3358
3359 (g) Added the -H (--with-filename) option.
3360
3361 (h) The combination of options -rh failed to suppress file names for files
3362 that were found from directory arguments.
3363
3364 (i) Added the -D (--devices) and -d (--directories) options.
3365
3366 (j) Added the -F (--fixed-strings) option.
3367
3368 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3369
3370 (l) Added the --colo(u)r option.
3371
3372 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3373 is not present by default.
3374
3375 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3376 items such as (?R) or (?1), when the recursion could match a number of
3377 alternatives. If it matched one of the alternatives, but subsequently,
3378 outside the recursion, there was a failure, the code tried to back up into
3379 the recursion. However, because of the way PCRE is implemented, this is not
3380 possible, and the result was an incorrect result from the match.
3381
3382 In order to prevent this happening, the specification of recursion has
3383 been changed so that all such subpatterns are automatically treated as
3384 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3385
3386 4. I had overlooked the fact that, in some locales, there are characters for
3387 which isalpha() is true but neither isupper() nor islower() are true. In
3388 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3389 and ordfeminine) are like this. This affected the treatment of \w and \W
3390 when they appeared in character classes, but not when they appeared outside
3391 a character class. The bit map for "word" characters is now created
3392 separately from the results of isalnum() instead of just taking it from the
3393 upper, lower, and digit maps. (Plus the underscore character, of course.)
3394
3395 5. The above bug also affected the handling of POSIX character classes such as
3396 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3397 permanent tables. Instead, the bit maps for such a class were previously
3398 created as the appropriate unions of the upper, lower, and digit bitmaps.
3399 Now they are created by subtraction from the [[:word:]] class, which has
3400 its own bitmap.
3401
3402 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3403 It is created by subtracting the vertical space characters (\x09, \x0a,
3404 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3405 subtraction was done in the overall bitmap for a character class, meaning
3406 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3407 be recognized. This bug has been fixed.
3408
3409 7. Patches from the folks at Google:
3410
3411 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3412 real life, but is still worth protecting against".
3413
3414 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3415 regular expressions".
3416
3417 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3418 have it.
3419
3420 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3421 "configure" and the latter not, in order to fix a problem somebody had
3422 with compiling the Arg class on HP-UX.
3423
3424 (e) Improve the error-handling of the C++ wrapper a little bit.
3425
3426 (f) New tests for checking recursion limiting.
3427
3428 8. The pcre_memmove() function, which is used only if the environment does not
3429 have a standard memmove() function (and is therefore rarely compiled),
3430 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3431 returning a result (though PCRE never actually uses the result).
3432
3433 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3434 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3435 returned instead of calling malloc() with an overflowing number that would
3436 most likely cause subsequent chaos.
3437
3438 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3439
3440 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3441 with this option is matched, the nmatch and pmatch options of regexec() are
3442 ignored.
3443
3444 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3445 provided in case anyone wants to the the POSIX interface with UTF-8
3446 strings.
3447
3448 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3449 C++ linking (needed for some HP-UX environments).
3450
3451 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3452 (unused parameter) and in the pcre_printint() function (omitted "default"
3453 switch label when the default is to do nothing).
3454
3455 15. Added some code to make it possible, when PCRE is compiled as a C++
3456 library, to replace subject pointers for pcre_exec() with a smart pointer
3457 class, thus making it possible to process discontinuous strings.
3458
3459 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3460 much the same function. They were added by different people who were trying
3461 to make PCRE easy to compile on non-Unix systems. It has been suggested
3462 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3463 for compiling on Windows systems. I have therefore replaced it with
3464 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3465 defaults to "extern" for C or "extern C" for C++, which works fine on
3466 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3467 SCOPE with something explicit in config.h. In addition:
3468
3469 (a) pcreposix.h still had just "extern" instead of either of these macros;
3470 I have replaced it with PCRE_DATA_SCOPE.
3471
3472 (b) Functions such as _pcre_xclass(), which are internal to the library,
3473 but external in the C sense, all had PCRE_EXPORT in their definitions.
3474 This is apparently wrong for the Windows case, so I have removed it.
3475 (It makes no difference on Unix-like systems.)
3476
3477 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3478 of recursive calls to match(). This is different to MATCH_LIMIT because
3479 that limits the total number of calls to match(), not all of which increase
3480 the depth of recursion. Limiting the recursion depth limits the amount of
3481 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3482 when PCRE is compiled, and changed at run time. A patch from Google adds
3483 this functionality to the C++ interface.
3484
3485 18. Changes to the handling of Unicode character properties:
3486
3487 (a) Updated the table to Unicode 4.1.0.
3488
3489 (b) Recognize characters that are not in the table as "Cn" (undefined).
3490
3491 (c) I revised the way the table is implemented to a much improved format
3492 which includes recognition of ranges. It now supports the ranges that
3493 are defined in UnicodeData.txt, and it also amalgamates other
3494 characters into ranges. This has reduced the number of entries in the
3495 table from around 16,000 to around 3,000, thus reducing its size
3496 considerably. I realized I did not need to use a tree structure after
3497 all - a binary chop search is just as efficient. Having reduced the
3498 number of entries, I extended their size from 6 bytes to 8 bytes to
3499 allow for more data.
3500
3501 (d) Added support for Unicode script names via properties such as \p{Han}.
3502
3503 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3504 matching that character.
3505
3506 20. When matching a repeated Unicode property with a minimum greater than zero,
3507 (for example \pL{2,}), PCRE could look past the end of the subject if it
3508 reached it while seeking the minimum number of characters. This could
3509 happen only if some of the characters were more than one byte long, because
3510 there is a check for at least the minimum number of bytes.
3511
3512 21. Refactored the implementation of \p and \P so as to be more general, to
3513 allow for more different types of property in future. This has changed the
3514 compiled form incompatibly. Anybody with saved compiled patterns that use
3515 \p or \P will have to recompile them.
3516
3517 22. Added "Any" and "L&" to the supported property types.
3518
3519 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3520 but give a compile time error if the value is greater than 0xff.
3521
3522 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3523 accidentally not being installed or uninstalled.
3524
3525 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3526 made were to insert the current release number. This seemed silly, because
3527 it made things harder for people building PCRE on systems that don't run
3528 "configure". I have turned pcre.h into a distributed file, no longer built
3529 by "configure", with the version identification directly included. There is
3530 no longer a pcre.h.in file.
3531
3532 However, this change necessitated a change to the pcre-config script as
3533 well. It is built from pcre-config.in, and one of the substitutions was the
3534 release number. I have updated configure.ac so that ./configure now finds
3535 the release number by grepping pcre.h.
3536
3537 26. Added the ability to run the tests under valgrind.
3538
3539
3540 Version 6.4 05-Sep-05
3541 ---------------------
3542
3543 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3544 "--" to be printed when multiple files were scanned, even when none of the
3545 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3546 consider it to be a bug, and have restored the previous behaviour.
3547
3548 2. A couple of code tidies to get rid of compiler warnings.
3549
3550 3. The pcretest program used to cheat by referring to symbols in the library
3551 whose names begin with _pcre_. These are internal symbols that are not
3552 really supposed to be visible externally, and in some environments it is
3553 possible to suppress them. The cheating is now confined to including
3554 certain files from the library's source, which is a bit cleaner.
3555
3556 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3557 file's purpose clearer.
3558
3559 5. Reorganized pcre_ucp_findchar().
3560
3561
3562 Version 6.3 15-Aug-05
3563 ---------------------
3564
3565 1. The file libpcre.pc.in did not have general read permission in the tarball.
3566
3567 2. There were some problems when building without C++ support:
3568
3569 (a) If C++ support was not built, "make install" and "make test" still
3570 tried to test it.
3571
3572 (b) There were problems when the value of CXX was explicitly set. Some
3573 changes have been made to try to fix these, and ...
3574
3575 (c) --disable-cpp can now be used to explicitly disable C++ support.
3576
3577 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3578 backslash in a target when C++ was disabled. This confuses some
3579 versions of "make", apparently. Using an intermediate variable solves
3580 this. (Same for CPP_LOBJ.)
3581
3582 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3583 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3584 necessary on certain architectures.
3585
3586 4. Added a setting of -export-symbols-regex to the link command to remove
3587 those symbols that are exported in the C sense, but actually are local
3588 within the library, and not documented. Their names all begin with
3589 "_pcre_". This is not a perfect job, because (a) we have to except some
3590 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3591 available (and never for static libraries). I have made a note to try to
3592 find a way round (a) in the future.
3593
3594
3595 Version 6.2 01-Aug-05
3596 ---------------------
3597
3598 1. There was no test for integer overflow of quantifier values. A construction
3599 such as {1111111111111111} would give undefined results. What is worse, if
3600 a minimum quantifier for a parenthesized subpattern overflowed and became
3601 negative, the calculation of the memory size went wrong. This could have
3602 led to memory overwriting.
3603
3604 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3605
3606 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3607 operating environments where this matters.
3608
3609 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3610 PCRE options from within the C++ wrapper.
3611
3612 5. Named capturing subpatterns were not being correctly counted when a pattern
3613 was compiled. This caused two problems: (a) If there were more than 100
3614 such subpatterns, the calculation of the memory needed for the whole
3615 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3616 back references of the form \12, where the number was greater than 9, were
3617 not recognized as back references, even though there were sufficient
3618 previous subpatterns.
3619
3620 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3621 versions of gcc, e.g. 2.95.4.
3622
3623
3624 Version 6.1 21-Jun-05
3625 ---------------------
3626
3627 1. There was one reference to the variable "posix" in pcretest.c that was not
3628 surrounded by "#if !defined NOPOSIX".
3629
3630 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3631 the cross-check on the old pcre_info() function, for the benefit of the
3632 cut-down version of PCRE that is currently imported into Exim.
3633
3634 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3635 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3636 patterns that start (?i) but I don't think that matters. The use of (?i) is
3637 just an example; this all applies to the other options as well.
3638
3639 4. Since libtool seems to echo the compile commands it is issuing, the output
3640 from "make" can be reduced a bit by putting "@" in front of each libtool
3641 compile command.
3642
3643 5. Patch from the folks at Google for configure.in to be a bit more thorough
3644 in checking for a suitable C++ installation before trying to compile the
3645 C++ stuff. This should fix a reported problem when a compiler was present,
3646 but no suitable headers.
3647
3648 6. The man pages all had just "PCRE" as their title. I have changed them to
3649 be the relevant file name. I have also arranged that these names are
3650 retained in the file doc/pcre.txt, which is a concatenation in text format
3651 of all the man pages except the little individual ones for each function.
3652
3653 7. The NON-UNIX-USE file had not been updated for the different set of source
3654 files that come with release 6. I also added a few comments about the C++
3655 wrapper.
3656
3657
3658 Version 6.0 07-Jun-05
3659 ---------------------
3660
3661 1. Some minor internal re-organization to help with my DFA experiments.
3662
3663 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3664 didn't matter for the library itself when fully configured, but did matter
3665 when compiling without UCP support, or within Exim, where the ucp files are
3666 not imported.
3667
3668 3. Refactoring of the library code to split up the various functions into
3669 different source modules. The addition of the new DFA matching code (see
3670 below) to a single monolithic source would have made it really too
3671 unwieldy, quite apart from causing all the code to be include in a
3672 statically linked application, when only some functions are used. This is
3673 relevant even without the DFA addition now that patterns can be compiled in
3674 one application and matched in another.
3675
3676 The downside of splitting up is that there have to be some external
3677 functions and data tables that are used internally in different modules of
3678 the library but which are not part of the API. These have all had their
3679 names changed to start with "_pcre_" so that they are unlikely to clash
3680 with other external names.
3681
3682 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3683 a different (DFA) algorithm. Although it is slower than the original
3684 function, it does have some advantages for certain types of matching
3685 problem.
3686
3687 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3688 including restarting after a partial match.
3689
3690 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3691 defined when compiling for Windows was sent to me. I have put it into the
3692 code, though I have no means of testing or verifying it.
3693
3694 7. Added the pcre_refcount() auxiliary function.
3695
3696 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3697 match before or at the first newline in the subject string. In pcretest,
3698 the /f option on a pattern can be used to set this.
3699
3700 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3701 would behave wrongly. This has been present in PCRE since release 4.0.
3702
3703 10. A number of changes to the pcregrep command:
3704
3705 (a) Refactored how -x works; insert ^(...)$ instead of setting
3706 PCRE_ANCHORED and checking the length, in preparation for adding
3707 something similar for -w.
3708
3709 (b) Added the -w (match as a word) option.
3710
3711 (c) Refactored the way lines are read and buffered so as to have more
3712 than one at a time available.
3713
3714 (d) Implemented a pcregrep test script.
3715
3716 (e) Added the -M (multiline match) option. This allows patterns to match
3717 over several lines of the subject. The buffering ensures that at least
3718 8K, or the rest of the document (whichever is the shorter) is available
3719 for matching (and similarly the previous 8K for lookbehind assertions).
3720
3721 (f) Changed the --help output so that it now says
3722
3723 -w, --word-regex(p)
3724
3725 instead of two lines, one with "regex" and the other with "regexp"
3726 because that confused at least one person since the short forms are the
3727 same. (This required a bit of code, as the output is generated
3728 automatically from a table. It wasn't just a text change.)
3729
3730 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3731 option but starts with a hyphen. Could be a pattern or a path name
3732 starting with a hyphen, for instance.
3733
3734 (h) "-" can be given as a file name to represent stdin.
3735
3736 (i) When file names are being printed, "(standard input)" is used for
3737 the standard input, for compatibility with GNU grep. Previously
3738 "<stdin>" was used.
3739
3740 (j) The option --label=xxx can be used to supply a name to be used for
3741 stdin when file names are being printed. There is no short form.
3742
3743 (k) Re-factored the options decoding logic because we are going to add
3744 two more options that take data. Such options can now be given in four
3745 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3746
3747 (l) Added the -A, -B, and -C options for requesting that lines of context
3748 around matches be printed.
3749
3750 (m) Added the -L option to print the names of files that do not contain
3751 any matching lines, that is, the complement of -l.
3752
3753 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3754 continue to scan other files.
3755
3756 (o) The -s option was incorrectly implemented. For compatibility with other
3757 greps, it now suppresses the error message for a non-existent or non-
3758 accessible file (but not the return code). There is a new option called
3759 -q that suppresses the output of matching lines, which was what -s was
3760 previously doing.
3761
3762 (p) Added --include and --exclude options to specify files for inclusion
3763 and exclusion when recursing.
3764
3765 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3766 Hopefully, it now does.
3767
3768 12. Missing cast in pcre_study().
3769
3770 13. Added an "uninstall" target to the makefile.
3771
3772 14. Replaced "extern" in the function prototypes in Makefile.in with
3773 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3774 world, but is set differently for Windows.
3775
3776 15. Added a second compiling function called pcre_compile2(). The only
3777 difference is that it has an extra argument, which is a pointer to an
3778 integer error code. When there is a compile-time failure, this is set
3779 non-zero, in addition to the error test pointer being set to point to an
3780 error message. The new argument may be NULL if no error number is required
3781 (but then you may as well call pcre_compile(), which is now just a
3782 wrapper). This facility is provided because some applications need a
3783 numeric error indication, but it has also enabled me to tidy up the way
3784 compile-time errors are handled in the POSIX wrapper.
3785
3786 16. Added VPATH=.libs to the makefile; this should help when building with one
3787 prefix path and installing with another. (Or so I'm told by someone who
3788 knows more about this stuff than I do.)
3789
3790 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3791 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3792 match everything, including newlines. This is not POSIX-compatible, but
3793 somebody wanted the feature. From pcretest it can be activated by using
3794 both the P and the s flags.
3795
3796 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3797
3798 19. libpcre.pc was being incorrectly installed as executable.
3799
3800 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3801 it now also looks for '\r' so that it will work unmodified on Windows.
3802
3803 21. Added Google's contributed C++ wrapper to the distribution.
3804
3805 22. Added some untidy missing memory free() calls in pcretest, to keep
3806 Electric Fence happy when testing.
3807
3808
3809
3810 Version 5.0 13-Sep-04
3811 ---------------------
3812
3813 1. Internal change: literal characters are no longer packed up into items
3814 containing multiple characters in a single byte-string. Each character
3815 is now matched using a separate opcode. However, there may be more than one
3816 byte in the character in UTF-8 mode.
3817
3818 2. The pcre_callout_block structure has two new fields: pattern_position and
3819 next_item_length. These contain the offset in the pattern to the next match
3820 item, and its length, respectively.
3821
3822 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3823 insertion of callouts before each pattern item. Added the /C option to
3824 pcretest to make use of this.
3825
3826 4. On the advice of a Windows user, the lines
3827
3828 #if defined(_WIN32) || defined(WIN32)
3829 _setmode( _fileno( stdout ), 0x8000 );
3830 #endif /* defined(_WIN32) || defined(WIN32) */
3831
3832 have been added to the source of pcretest. This apparently does useful
3833 magic in relation to line terminators.
3834
3835 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3836 for the benefit of those environments where the "b" makes a difference.
3837
3838 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3839 to know about it. I have put a hack into configure.in that adds in code
3840 to set GCC=yes if CC=icc. This seems to end up at a point in the
3841 generated configure script that is early enough to affect the setting of
3842 compiler options, which is what is needed, but I have no means of testing
3843 whether it really works. (The user who reported this had patched the
3844 generated configure script, which of course I cannot do.)
3845
3846 LATER: After change 22 below (new libtool files), the configure script
3847 seems to know about icc (and also ecc). Therefore, I have commented out
3848 this hack in configure.in.
3849
3850 7. Added support for pkg-config (2 patches were sent in).
3851
3852 8. Negated POSIX character classes that used a combination of internal tables
3853 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3854 [[:^ascii]]. Typically, they would match almost any characters. The other
3855 POSIX classes were not broken in this way.
3856
3857 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3858 to find the match, as PCRE was deluded into thinking that the match had to
3859 start at the start point or following a newline. The same bug applied to
3860 patterns with negative forward assertions or any backward assertions
3861 preceding ".*" at the start, unless the pattern required a fixed first
3862 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3863
3864 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3865 starting at the last subject character, bytes beyond the end of the subject
3866 string were read.
3867
3868 11. Renamed the variable "class" as "classbits" to make life easier for C++
3869 users. (Previously there was a macro definition, but it apparently wasn't
3870 enough.)
3871
3872 12. Added the new field "tables" to the extra data so that tables can be passed
3873 in at exec time, or the internal tables can be re-selected. This allows
3874 a compiled regex to be saved and re-used at a later time by a different
3875 program that might have everything at different addresses.
3876
3877 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3878 -R library as well as a -L library.
3879
3880 14. The debugging options of pcretest (-d on the command line or D on a
3881 pattern) showed incorrect output for anything following an extended class
3882 that contained multibyte characters and which was followed by a quantifier.
3883
3884 15. Added optional support for general category Unicode character properties
3885 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3886 support. It adds about 90K to the size of the library. The meanings of the
3887 inbuilt class escapes such as \d and \s have NOT been changed.
3888
3889 16. Updated pcredemo.c to include calls to free() to release the memory for the
3890 compiled pattern.
3891
3892 17. The generated file chartables.c was being created in the source directory
3893 instead of in the building directory. This caused the build to fail if the
3894 source directory was different from the building directory, and was
3895 read-only.
3896
3897 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3898 file. No doubt somebody will tell me if they don't make sense... Also added
3899 Dan Mooney's comments about building on OpenVMS.
3900
3901 19. Added support for partial matching via the PCRE_PARTIAL option for
3902 pcre_exec() and the \P data escape in pcretest.
3903
3904 20. Extended pcretest with 3 new pattern features:
3905
3906 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3907 write the compiled pattern to the file whose name is "rest-of-line".
3908 This is a straight binary dump of the data, with the saved pointer to
3909 the character tables forced to be NULL. The study data, if any, is
3910 written too. After writing, pcretest reads a new pattern.
3911
3912 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3913 compiled pattern from the given file. There must not be any
3914 occurrences of "<" in the file name (pretty unlikely); if there are,
3915 pcretest will instead treat the initial "<" as a pattern delimiter.
3916 After reading in the pattern, pcretest goes on to read data lines as
3917 usual.
3918
3919 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3920 and 16-bit fields in a compiled pattern, to simulate a pattern that
3921 was compiled on a host of opposite endianness.
3922
3923 21. The pcre-exec() function can now cope with patterns that were compiled on
3924 hosts of opposite endianness, with this restriction:
3925
3926 As for any compiled expression that is saved and used later, the tables
3927 pointer field cannot be preserved; the extra_data field in the arguments
3928 to pcre_exec() should be used to pass in a tables address if a value
3929 other than the default internal tables were used at compile time.
3930
3931 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3932 now diagnosed as an error. Previously, most of the time, a negative number
3933 would have been treated as zero, but if in addition "ovector" was passed as
3934 NULL, a crash could occur.
3935
3936 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3937 new versions from the libtool 1.5 distribution (the last one is a copy of
3938 a file called libtool.m4). This seems to have fixed the need to patch
3939 "configure" to support Darwin 1.3 (which I used to do). However, I still
3940 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3941 workstation).
3942
3943 24. Changed the PCRE licence to be the more standard "BSD" licence.
3944
3945
3946 Version 4.5 01-Dec-03
3947 ---------------------
3948
3949 1. There has been some re-arrangement of the code for the match() function so
3950 that it can be compiled in a version that does not call itself recursively.
3951 Instead, it keeps those local variables that need separate instances for
3952 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3953 needs to "recurse". Keeping track of where control must go is done by means
3954 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3955 hide most of the details from the main code, and operates only if
3956 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3957 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3958 operating.
3959
3960 To make it easier for callers to provide specially tailored get/free
3961 functions for this usage, two new functions, pcre_stack_malloc, and
3962 pcre_stack_free, are used. They are always called in strict stacking order,
3963 and the size of block requested is always the same.
3964
3965 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3966 PCRE has been compiled to use the stack or the heap for recursion. The
3967 -C option of pcretest uses this to show which version is compiled.
3968
3969 A new data escape \S, is added to pcretest; it causes the amounts of store
3970 obtained and freed by both kinds of malloc/free at match time to be added
3971 to the output.
3972
3973 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3974 what's available on my current Linux desktop machine.
3975
3976 3. When matching a UTF-8 string, the test for a valid string at the start has
3977 been extended. If start_offset is not zero, PCRE now checks that it points
3978 to a byte that is the start of a UTF-8 character. If not, it returns
3979 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3980 this is necessary because there may be backward assertions in the pattern.
3981 When matching the same subject several times, it may save resources to use
3982 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3983
3984 4. The code for checking the validity of UTF-8 strings has been tightened so
3985 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3986 containing "overlong sequences".
3987
3988 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3989 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3990 should have been "&", but it just so happened that all the cases this let
3991 through by mistake were picked up later in the function.
3992
3993 6. I had used a variable called "isblank" - this is a C99 function, causing
3994 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3995
3996 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3997 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3998 so the version is shown; (c) stop "make test" echoing "./RunTest".
3999
4000 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4001
4002 9. The prototype for memmove() for systems that don't have it was using
4003 size_t, but the inclusion of the header that defines size_t was later. I've
4004 moved the #includes for the C headers earlier to avoid this.
4005
4006 10. Added some adjustments to the code to make it easier to compiler on certain
4007 special systems:
4008
4009 (a) Some "const" qualifiers were missing.
4010 (b) Added the macro EXPORT before all exported functions; by default this
4011 is defined to be empty.
4012 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4013 that it reads its output file name as an argument instead of writing
4014 to the standard output and assuming this can be redirected.
4015
4016 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4017 class containing characters with values greater than 255, PCRE compilation
4018 went into a loop.
4019
4020 12. A recursive reference to a subpattern that was within another subpattern
4021 that had a minimum quantifier of zero caused PCRE to crash. For example,
4022 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4023 recursion. If the recursively-called subpattern itself had a zero repeat,
4024 that was OK.
4025
4026 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4027 buffer into which it was copied (for escape processing) was still set at
4028 1024, so long lines caused crashes.
4029
4030 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4031 "internal error: code overflow...". This applied to any character class
4032 that was followed by a possessive quantifier.
4033
4034 15. Modified the Makefile to add libpcre.la as a prerequisite for
4035 libpcreposix.la because I was told this is needed for a parallel build to
4036 work.
4037
4038 16. If a pattern that contained .* following optional items at the start was
4039 studied, the wrong optimizing data was generated, leading to matching
4040 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4041 matching string must start with a or b or c. The correct conclusion for
4042 this pattern is that a match can start with any character.
4043
4044
4045 Version 4.4 13-Aug-03
4046 ---------------------
4047
4048 1. In UTF-8 mode, a character class containing characters with values between
4049 127 and 255 was not handled correctly if the compiled pattern was studied.
4050 In fixing this, I have also improved the studying algorithm for such
4051 classes (slightly).
4052
4053 2. Three internal functions had redundant arguments passed to them. Removal
4054 might give a very teeny performance improvement.
4055
4056 3. Documentation bug: the value of the capture_top field in a callout is *one
4057 more than* the number of the hightest numbered captured substring.
4058
4059 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4060 in incorrectly linking with a previously installed version. They now link
4061 explicitly with libpcre.la.
4062
4063 5. configure.in no longer needs to recognize Cygwin specially.
4064
4065 6. A problem in pcre.in for Windows platforms is fixed.
4066
4067 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4068 pcretest, it used to include the size of the study block as part of its
4069 output. Unfortunately, the structure contains a field that has a different
4070 size on different hardware architectures. This meant that the tests that
4071 showed this size failed. As the block is currently always of a fixed size,
4072 this information isn't actually particularly useful in pcretest output, so
4073 I have just removed it.
4074
4075 8. Three pre-processor statements accidentally did not start in column 1.
4076 Sadly, there are *still* compilers around that complain, even though
4077 standard C has not required this for well over a decade. Sigh.
4078
4079 9. In pcretest, the code for checking callouts passed small integers in the
4080 callout_data field, which is a void * field. However, some picky compilers
4081 complained about the casts involved for this on 64-bit systems. Now
4082 pcretest passes the address of the small integer instead, which should get
4083 rid of the warnings.
4084
4085 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4086 both compile and run time, and gives an error if an invalid UTF-8 sequence
4087 is found. There is a option for disabling this check in cases where the
4088 string is known to be correct and/or the maximum performance is wanted.
4089
4090 11. In response to a bug report, I changed one line in Makefile.in from
4091
4092 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4093 to
4094 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4095
4096 to look similar to other lines, but I have no way of telling whether this
4097 is the right thing to do, as I do not use Windows. No doubt I'll get told
4098 if it's wrong...
4099
4100
4101 Version 4.3 21-May-03
4102 ---------------------
4103
4104 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4105 Makefile.
4106
4107 2. Some refactoring to improve the quality of the code:
4108
4109 (i) The utf8_table... variables are now declared "const".
4110
4111 (ii) The code for \cx, which used the "case flipping" table to upper case
4112 lower case letters, now just substracts 32. This is ASCII-specific,
4113 but the whole concept of \cx is ASCII-specific, so it seems
4114 reasonable.
4115
4116 (iii) PCRE was using its character types table to recognize decimal and
4117 hexadecimal digits in the pattern. This is silly, because it handles
4118 only 0-9, a-f, and A-F, but the character types table is locale-
4119 specific, which means strange things might happen. A private
4120 table is now used for this - though it costs 256 bytes, a table is
4121 much faster than multiple explicit tests. Of course, the standard
4122 character types table is still used for matching digits in subject
4123 strings against \d.
4124
4125 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4126 ending in _t are). So I've renamed it as ESC_tee.
4127
4128 3. The first argument for regexec() in the POSIX wrapper should have been
4129 defined as "const".
4130
4131 4. Changed pcretest to use malloc() for its buffers so that they can be
4132 Electric Fenced for debugging.
4133
4134 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4135 to read one or more bytes before the start of the subject string. Often this
4136 had no effect on PCRE's behaviour, but in some circumstances it could
4137 provoke a segmentation fault.
4138
4139 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4140 to try to read one or more bytes before the start of the subject string.
4141
4142 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4143 UTF-8 support could misbehave in various ways if the subject string
4144 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4145 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4146 back over UTF-8 characters.)
4147
4148
4149 Version 4.2 14-Apr-03
4150 ---------------------
4151
4152 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4153
4154 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4155 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4156 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4157 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4158 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4159 and BUILD_EXEEXT
4160 Note: automatic setting of the BUILD variables is not yet working
4161 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4162 compile-time but not at link-time
4163 [LINK]: use for linking executables only
4164 make different versions for Windows and non-Windows
4165 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4166 libraries
4167 [LINK_FOR_BUILD]: new variable
4168 [OBJEXT]: use throughout
4169 [EXEEXT]: use throughout
4170 <winshared>: new target
4171 <wininstall>: new target
4172 <dftables.o>: use native compiler
4173 <dftables>: use native linker
4174 <install>: handle Windows platform correctly
4175 <clean>: ditto
4176 <check>: ditto
4177 copy DLL to top builddir before testing
4178
4179 As part of these changes, -no-undefined was removed again. This was reported
4180 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4181 in any case.
4182
4183 3. Some tidies to get rid of compiler warnings:
4184
4185 . In the match_data structure, match_limit was an unsigned long int, whereas
4186 match_call_count was an int. I've made them both unsigned long ints.
4187
4188 . In pcretest the fact that a const uschar * doesn't automatically cast to
4189 a void * provoked a warning.
4190
4191 . Turning on some more compiler warnings threw up some "shadow" variables
4192 and a few more missing casts.
4193
4194 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4195 option, a class that contained a single character with a value between 128
4196 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4197
4198 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4199 option, a class that contained several characters, but with at least one
4200 whose value was between 128 and 255 caused PCRE to crash.
4201
4202
4203 Version 4.1 12-Mar-03
4204 ---------------------
4205
4206 1. Compiling with gcc -pedantic found a couple of places where casts were
4207 needed, and a string in dftables.c that was longer than standard compilers are
4208 required to support.
4209
4210 2. Compiling with Sun's compiler found a few more places where the code could
4211 be tidied up in order to avoid warnings.
4212
4213 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4214 first of these names is deprecated in the latest Autoconf in favour of the name
4215 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4216 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4217 analogy I have changed it to CFLAGS_FOR_BUILD.
4218
4219 4. Added -no-undefined to the linking command in the Makefile, because this is
4220 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4221 linking step for the pcreposix library.
4222
4223 5. PCRE was failing to diagnose the case of two named groups with the same
4224 name.
4225
4226 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4227 literal character that is needed in the subject for a match, and scans along to
4228 ensure that it is present before embarking on the full matching process. This
4229 saves time in cases of nested unlimited repeats that are never going to match.
4230 Problem: the scan can take a lot of time if the subject is very long (e.g.
4231 megabytes), thus penalizing straightforward matches. It is now done only if the
4232 amount of subject to be scanned is less than 1000 bytes.
4233
4234 7. A lesser problem with the same optimization is that it was recording the
4235 first character of an anchored pattern as "needed", thus provoking a search
4236 right along the subject, even when the first match of the pattern was going to
4237 fail. The "needed" character is now not set for anchored patterns, unless it
4238 follows something in the pattern that is of non-fixed length. Thus, it still
4239 fulfils its original purpose of finding quick non-matches in cases of nested
4240 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4241
4242
4243 Version 4.0 17-Feb-03
4244 ---------------------
4245
4246 1. If a comment in an extended regex that started immediately after a meta-item
4247 extended to the end of string, PCRE compiled incorrect data. This could lead to
4248 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4249
4250 2. Moved to autoconf 2.53 and libtool 1.4.2.
4251
4252 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4253 the special perltest8 script is no longer needed - all the tests can be run
4254 from a single perltest script.
4255
4256 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4257 by \s. It has now been removed in PCRE. This means it isn't recognized as
4258 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4259 class [:space:] *does* include VT, thereby creating a mess.
4260
4261 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4262 space and tab.
4263
4264 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4265 its new features into the main test script, reducing the number of scripts.
4266
4267 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4268 were backward compatible, and made the (?i) apply to the whole pattern, as if
4269 /i were given. Now it behaves more logically, and applies the option setting
4270 only to what follows. PCRE has been changed to follow suit. However, if it
4271 finds options settings right at the start of the pattern, it extracts them into
4272 the global options, as before. Thus, they show up in the info data.
4273
4274 8. Added support for the \Q...\E escape sequence. Characters in between are
4275 treated as literals. This is slightly different from Perl in that $ and @ are
4276 also handled as literals inside the quotes. In Perl, they will cause variable
4277 interpolation. Note the following examples:
4278
4279 Pattern PCRE matches Perl matches
4280
4281 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4282 \Qabc\$xyz\E abc\$xyz abc\$xyz
4283 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4284
4285 For compatibility with Perl, \Q...\E sequences are recognized inside character
4286 classes as well as outside them.
4287
4288 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4289 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4290 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4291 signed/unsigned warnings.
4292
4293 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4294 option for pcretest, so I've replaced it by a simple function that does just
4295 that job.
4296
4297 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4298 "pcregrep -".
4299
4300 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4301 Java package. This provides some syntactic sugar for simple cases of what my
4302 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4303 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4304 item, you can use this simplified notation. Note that only makes sense with
4305 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4306 greediness, whatever the setting of the PCRE_UNGREEDY option.
4307
4308 13. A change of greediness default within a pattern was not taking effect at
4309 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4310 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4311 was abstracted outside.
4312
4313 14. PCRE now supports the \G assertion. It is true when the current matching
4314 position is at the start point of the match. This differs from \A when the
4315 starting offset is non-zero. Used with the /g option of pcretest (or similar
4316 code), it works in the same way as it does for Perl's /g option. If all
4317 alternatives of a regex begin with \G, the expression is anchored to the start
4318 match position, and the "anchored" flag is set in the compiled expression.
4319
4320 15. Some bugs concerning the handling of certain option changes within patterns
4321 have been fixed. These applied to options other than (?ims). For example,
4322 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4323 been the other way round. Some of this was related to change 7 above.
4324
4325 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4326 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4327 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4328 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4329
4330 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4331 mode. Unlike ".", it always matches newline, whatever the setting of
4332 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4333 assertions. Perl allows it, but it doesn't (in general) work because it can't
4334 calculate the length of the lookbehind. At least, that's the case for Perl
4335 5.8.0 - I've been told they are going to document that it doesn't work in
4336 future.
4337
4338 18. Added an error diagnosis for escapes that PCRE does not support: these are
4339 \L, \l, \N, \P, \p, \U, \u, and \X.
4340
4341 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4342 reading past the end of the pattern in cases such as /[abcd/.
4343
4344 20. PCRE was getting more memory than necessary for patterns with classes that
4345 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4346
4347 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4348 compiling PCRE for use with Virtual Pascal.
4349
4350 22. Small fix to the Makefile to make it work properly if the build is done
4351 outside the source tree.
4352
4353 23. Added a new extension: a condition to go with recursion. If a conditional
4354 subpattern starts with (?(R) the "true" branch is used if recursion has
4355 happened, whereas the "false" branch is used only at the top level.
4356
4357 24. When there was a very long string of literal characters (over 255 bytes
4358 without UTF support, over 250 bytes with UTF support), the computation of how
4359 much memory was required could be incorrect, leading to segfaults or other
4360 strange effects.
4361
4362 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4363 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4364 there was a subsequent back reference to those brackets. This meant that, for
4365 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4366 possible to check for precisely this case. All we can do is abandon the
4367 optimization if .* occurs inside capturing brackets when there are any back
4368 references whatsoever. (See below for a better fix that came later.)
4369
4370 26. The handling of the optimization for finding the first character of a
4371 non-anchored pattern, and for finding a character that is required later in the
4372 match were failing in some cases. This didn't break the matching; it just
4373 failed to optimize when it could. The way this is done has been re-implemented.
4374
4375 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4376
4377 28. Added a new feature that provides some of the functionality that Perl
4378 provides with (?{...}). The facility is termed a "callout". The way it is done
4379 in PCRE is for the caller to provide an optional function, by setting
4380 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4381 global variable. By default it is unset, which disables all calling out. To get
4382 the function called, the regex must include (?C) at appropriate points. This
4383 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4384 This provides a means of identifying different callout points. When PCRE
4385 reaches such a point in the regex, if pcre_callout has been set, the external
4386 function is called. It is provided with data in a structure called
4387 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4388 matching continues; if it returns a non-zero value, the match at the current
4389 point fails. However, backtracking will occur if possible. [This was changed
4390 later and other features added - see item 49 below.]
4391
4392 29. pcretest is upgraded to test the callout functionality. It provides a
4393 callout function that displays information. By default, it shows the start of
4394 the match and the current position in the text. There are some new data escapes
4395 to vary what happens:
4396
4397 \C+ in addition, show current contents of captured substrings
4398 \C- do not supply a callout function
4399 \C!n return 1 when callout number n is reached
4400 \C!n!m return 1 when callout number n is reached for the mth time
4401
4402 30. If pcregrep was called with the -l option and just a single file name, it
4403 output "<stdin>" if a match was found, instead of the file name.
4404
4405 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4406 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4407 pcre_exec(). This saves a malloc/free per call. The default value of
4408 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4409 when configuring.
4410
4411 32. The default maximum size of a compiled pattern is 64K. There have been a
4412 few cases of people hitting this limit. The code now uses macros to handle the
4413 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4414 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4415 configuring. Tests 2 and 5 work only with 2-byte links because they output
4416 debugging information about compiled patterns.
4417
4418 33. Internal code re-arrangements:
4419
4420 (a) Moved the debugging function for printing out a compiled regex into
4421 its own source file (printint.c) and used #include to pull it into
4422 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4423 separate copies.
4424
4425 (b) Defined the list of op-code names for debugging as a macro in
4426 internal.h so that it is next to the definition of the opcodes.
4427
4428 (c) Defined a table of op-code lengths for simpler skipping along compiled
4429 code. This is again a macro in internal.h so that it is next to the
4430 definition of the opcodes.
4431
4432 34. Added support for recursive calls to individual subpatterns, along the
4433 lines of Robin Houston's patch (but implemented somewhat differently).
4434
4435 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4436 allow it to read and process whole directories in Win32. This code was
4437 contributed by Lionel Fourquaux; it has not been tested by me.
4438
4439 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4440 used to name a group. Names consist of alphanumerics and underscores, and must
4441 be unique. Back references use the syntax (?P=name) and recursive calls use
4442 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4443 numbers. The function pcre_fullinfo() can be used after compilation to extract
4444 a name/number map. There are three relevant calls:
4445
4446 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4447 PCRE_INFO_NAMECOUNT yields the number of entries
4448 PCRE_INFO_NAMETABLE yields a pointer to the map.
4449
4450 The map is a vector of fixed-size entries. The size of each entry depends on
4451 the length of the longest name used. The first two bytes of each entry are the
4452 group number, most significant byte first. There follows the corresponding
4453 name, zero terminated. The names are in alphabetical order.
4454
4455 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4456 case instead of 255. Making it the same both with and without UTF-8 support
4457 means that the same test output works with both.
4458
4459 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4460 calling malloc() with a zero argument.
4461
4462 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4463 optimization. I've improved things by keeping a bitmap of backreferences with
4464 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4465 fact referenced, the optimization can be applied. It is unlikely that a
4466 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4467 the match to follow \n) will appear inside brackets with a number greater than
4468 31, but if it does, any back reference > 31 suppresses the optimization.
4469
4470 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4471 of disabling numbered capturing parentheses. Any opening parenthesis that is
4472 not followed by ? behaves as if it were followed by ?: but named parentheses
4473 can still be used for capturing (and they will acquire numbers in the usual
4474 way).
4475
4476 41. Redesigned the return codes from the match() function into yes/no/error so
4477 that errors can be passed back from deep inside the nested calls. A malloc
4478 failure while inside a recursive subpattern call now causes the
4479 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4480
4481 42. It is now possible to set a limit on the number of times the match()
4482 function is called in a call to pcre_exec(). This facility makes it possible to
4483 limit the amount of recursion and backtracking, though not in a directly
4484 obvious way, because the match() function is used in a number of different
4485 circumstances. The count starts from zero for each position in the subject
4486 string (for non-anchored patterns). The default limit is, for compatibility, a
4487 large number, namely 10 000 000. You can change this in two ways:
4488
4489 (a) When configuring PCRE before making, you can use --with-match-limit=n
4490 to set a default value for the compiled library.
4491
4492 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4493 a different value is set. See 45 below.
4494
4495 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4496
4497 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4498 of things that can be changed at compile time. The first argument specifies
4499 what is wanted and the second points to where the information is to be placed.
4500 The current list of available information is:
4501
4502 PCRE_CONFIG_UTF8
4503
4504 The output is an integer that is set to one if UTF-8 support is available;
4505 otherwise it is set to zero.
4506
4507 PCRE_CONFIG_NEWLINE
4508
4509 The output is an integer that it set to the value of the code that is used for
4510 newline. It is either LF (10) or CR (13).
4511
4512 PCRE_CONFIG_LINK_SIZE
4513
4514 The output is an integer that contains the number of bytes used for internal
4515 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4516
4517 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4518
4519 The output is an integer that contains the threshold above which the POSIX
4520 interface uses malloc() for output vectors. See item 31 above.
4521
4522 PCRE_CONFIG_MATCH_LIMIT
4523
4524 The output is an unsigned integer that contains the default limit of the number
4525 of match() calls in a pcre_exec() execution. See 42 above.
4526
4527 44. pcretest has been upgraded by the addition of the -C option. This causes it
4528 to extract all the available output from the new pcre_config() function, and to
4529 output it. The program then exits immediately.
4530
4531 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4532 order to support additional features. One way would have been to define
4533 pcre_exec2() (for example) with extra arguments, but this would not have been
4534 extensible, and would also have required all calls to the original function to
4535 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4536 is used for passing in "extra" data from pcre_study().
4537
4538 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4539 contains the following fields:
4540
4541 flags a bitmap indicating which of the following fields are set
4542 study_data opaque data from pcre_study()
4543 match_limit a way of specifying a limit on match() calls for a specific
4544 call to pcre_exec()
4545 callout_data data for callouts (see 49 below)
4546
4547 The flag bits are also defined in pcre.h, and are
4548
4549 PCRE_EXTRA_STUDY_DATA
4550 PCRE_EXTRA_MATCH_LIMIT
4551 PCRE_EXTRA_CALLOUT_DATA
4552
4553 The pcre_study() function now returns one of these new pcre_extra blocks, with
4554 the actual study data pointed to by the study_data field, and the
4555 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4556 before. That is, this change is entirely upwards-compatible and requires no
4557 change to existing code.
4558
4559 If you want to pass in additional data to pcre_exec(), you can either place it
4560 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4561 block.
4562
4563 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4564 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4565 times with different match limits, until it finds the minimum value needed for
4566 pcre_exec() to complete. The value is then output. This can be instructive; for
4567 most simple matches the number is quite small, but for pathological cases it
4568 gets very large very quickly.
4569
4570 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4571 returns the size of the data block pointed to by the study_data field in a
4572 pcre_extra block, that is, the value that was passed as the argument to
4573 pcre_malloc() when PCRE was getting memory in which to place the information
4574 created by pcre_study(). The fourth argument should point to a size_t variable.
4575 pcretest has been extended so that this information is shown after a successful
4576 pcre_study() call when information about the compiled regex is being displayed.
4577
4578 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4579 because what follows is always an absolute path. (Later: it turns out that this
4580 is more than cosmetic for MinGW, because it doesn't like empty path
4581 components.)
4582
4583 49. Some changes have been made to the callout feature (see 28 above):
4584
4585 (i) A callout function now has three choices for what it returns:
4586
4587 0 => success, carry on matching
4588 > 0 => failure at this point, but backtrack if possible
4589 < 0 => serious error, return this value from pcre_exec()
4590
4591 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4592 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4593 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4594 use by callout functions. It will never be used by PCRE itself.
4595
4596 (ii) The pcre_extra structure (see 45 above) has a void * field called
4597 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4598 pcre_callout_block structure has a field of the same name. The contents of
4599 the field passed in the pcre_extra structure are passed to the callout
4600 function in the corresponding field in the callout block. This makes it
4601 easier to use the same callout-containing regex from multiple threads. For
4602 testing, the pcretest program has a new data escape
4603
4604 \C*n pass the number n (may be negative) as callout_data
4605
4606 If the callout function in pcretest receives a non-zero value as
4607 callout_data, it returns that value.
4608
4609 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4610 there were some redundant $(CFLAGS) in commands that are now specified as
4611 $(LINK), which already includes $(CFLAGS).
4612
4613 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4614 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4615 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4616 one-byte characters throughout. Note that case-insensitive matching applies
4617 only to characters whose values are less than 256. PCRE doesn't support the
4618 notion of cases for higher-valued characters.
4619
4620 (i) A character class whose characters are all within 0-255 is handled as
4621 a bit map, and the map is inverted for negative classes. Previously, a
4622 character > 255 always failed to match such a class; however it should
4623 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4624
4625 (ii) A negated character class with a single character < 255 is coded as
4626 "not this character" (OP_NOT). This wasn't working properly when the test
4627 character was multibyte, either singly or repeated.
4628
4629 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4630 mode, for example: \x{100}{2,3}.
4631
4632 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4633 singly or repeated) now correctly test multibyte characters. However,
4634 PCRE doesn't recognize any characters with values greater than 255 as
4635 digits, spaces, or word characters. Such characters always match \D, \S,
4636 and \W, and never match \d, \s, or \w.
4637
4638 (v) Classes may now contain characters and character ranges with values
4639 greater than 255. For example: [ab\x{100}-\x{400}].
4640
4641 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4642 PCRE in UTF-8 mode.
4643
4644 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4645 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4646 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4647 value.)
4648
4649 53. The single man page has become too large. I have therefore split it up into
4650 a number of separate man pages. These also give rise to individual HTML pages;
4651 these are now put in a separate directory, and there is an index.html page that
4652 lists them all. Some hyperlinking between the pages has been installed.
4653
4654 54. Added convenience functions for handling named capturing parentheses.
4655
4656 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4657 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4658 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4659 are faulted.
4660
4661 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4662 calling configure. These values are used when compiling the dftables.c program
4663 which is run to generate the source of the default character tables. They
4664 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4665 you will need to set these values.
4666
4667 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4668
4669
4670 Version 3.9 02-Jan-02
4671 ---------------------
4672
4673 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4674
4675 2. If --disable-static was given, the building process failed when trying to
4676 build pcretest and pcregrep. (For some reason it was using libtool to compile
4677 them, which is not right, as they aren't part of the library.)
4678
4679
4680 Version 3.8 18-Dec-01
4681 ---------------------
4682
4683 1. The experimental UTF-8 code was completely screwed up. It was packing the
4684 bytes in the wrong order. How dumb can you get?
4685
4686
4687 Version 3.7 29-Oct-01
4688 ---------------------
4689
4690 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4691 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4692 this didn't happen under Solaris 8, where I normally test things.
4693
4694 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4695 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4696 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4697 this fix an hour or so after the initial 3.7 release.)
4698
4699
4700 Version 3.6 23-Oct-01
4701 ---------------------
4702
4703 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4704 offsets passed as NULL with zero offset count.
4705
4706 2. The config.guess and config.sub files had not been updated when I moved to
4707 the latest autoconf.
4708
4709
4710 Version 3.5 15-Aug-01
4711 ---------------------
4712
4713 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4714 had been forgotten.
4715
4716 2. By using declared but undefined structures, we can avoid using "void"
4717 definitions in pcre.h while keeping the internal definitions of the structures
4718 private.
4719
4720 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4721 user point of view, this means that both static and shared libraries are built
4722 by default, but this can be individually controlled. More of the work of
4723 handling this static/shared cases is now inside libtool instead of PCRE's make
4724 file.
4725
4726 4. The pcretest utility is now installed along with pcregrep because it is
4727 useful for users (to test regexs) and by doing this, it automatically gets
4728 relinked by libtool. The documentation has been turned into a man page, so
4729 there are now .1, .txt, and .html versions in /doc.
4730
4731 5. Upgrades to pcregrep:
4732 (i) Added long-form option names like gnu grep.
4733 (ii) Added --help to list all options with an explanatory phrase.
4734 (iii) Added -r, --recursive to recurse into sub-directories.
4735 (iv) Added -f, --file to read patterns from a file.
4736
4737 6. pcre_exec() was referring to its "code" argument before testing that
4738 argument for NULL (and giving an error if it was NULL).
4739
4740 7. Upgraded Makefile.in to allow for compiling in a different directory from
4741 the source directory.
4742
4743 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4744 options bits, the pointer it was passed was to an int instead of to an unsigned
4745 long int. This mattered only on 64-bit systems.
4746
4747 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4748 generated) instead of pcre.in, which it its source. Also made the same change
4749 in several of the .c files.
4750
4751 10. A new release of gcc defines printf() as a macro, which broke pcretest
4752 because it had an ifdef in the middle of a string argument for printf(). Fixed
4753 by using separate calls to printf().
4754
4755 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4756 script, to force use of CR or LF instead of \n in the source. On non-Unix
4757 systems, the value can be set in config.h.
4758
4759 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4760 absolute limit. Changed the text of the error message to make this clear, and
4761 likewise updated the man page.
4762
4763 13. The limit of 99 on the number of capturing subpatterns has been removed.
4764 The new limit is 65535, which I hope will not be a "real" limit.
4765
4766
4767 Version 3.4 22-Aug-00
4768 ---------------------
4769
4770 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4771
4772 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4773
4774
4775 Version 3.3 01-Aug-00
4776 ---------------------
4777
4778 1. If an octal character was given, but the value was greater than \377, it
4779 was not getting masked to the least significant bits, as documented. This could
4780 lead to crashes in some systems.
4781
4782 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4783 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4784
4785 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4786 These just pass their arguments on to (pcre_free)(), but they are provided
4787 because some uses of PCRE bind it to non-C systems that can call its functions,
4788 but cannot call free() or pcre_free() directly.
4789
4790 4. Add "make test" as a synonym for "make check". Corrected some comments in
4791 the Makefile.
4792
4793 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4794 Makefile.
4795
4796 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4797 command called pgrep for grepping around the active processes.
4798
4799 7. Added the beginnings of support for UTF-8 character strings.
4800
4801 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4802 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4803 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4804 out for the ar command.)
4805
4806
4807 Version 3.2 12-May-00
4808 ---------------------
4809
4810 This is purely a bug fixing release.
4811
4812 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4813 of ZA. This was just one example of several cases that could provoke this bug,
4814 which was introduced by change 9 of version 2.00. The code for breaking
4815 infinite loops after an iteration that matches an empty string was't working
4816 correctly.
4817
4818 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4819 when matched against abbab (for example). After matching an empty string, it
4820 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4821 caused it to match further down the string than it should.
4822
4823 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4824 was there because it doesn't seem to be needed, and it causes trouble on some
4825 systems, as it is not a Standard C header. It has been removed.
4826
4827 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4828 were reported on the Macintosh. The changes were from
4829
4830 while ((c = *(++ptr)) != 0 && c != '\n');
4831 to
4832 while ((c = *(++ptr)) != 0 && c != '\n') ;
4833
4834 Totally extraordinary, but if that's what it takes...
4835
4836 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4837 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4838 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4839 assumes the way PCRE uses memmove() (always moving upwards).
4840
4841 6. PCRE is being used in one environment where strchr() is not available. There
4842 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4843 faster code anyway.
4844
4845
4846 Version 3.1 09-Feb-00
4847 ---------------------
4848
4849 The only change in this release is the fixing of some bugs in Makefile.in for
4850 the "install" target:
4851
4852 (1) It was failing to install pcreposix.h.
4853
4854 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4855
4856
4857 Version 3.0 01-Feb-00
4858 ---------------------
4859
4860 1. Add support for the /+ modifier to perltest (to output $` like it does in
4861 pcretest).
4862
4863 2. Add support for the /g modifier to perltest.
4864
4865 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4866 matches null strings.
4867
4868 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4869 pattern. Perl treats empty patterns specially - it reuses the most recent
4870 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4871 effect.
4872
4873 5. The POSIX interface was broken in that it was just handing over the POSIX
4874 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4875 required a bigger vector, with some working space on the end. This means that
4876 the POSIX wrapper now has to get and free some memory, and copy the results.
4877
4878 6. Added some simple autoconf support, placing the test data and the
4879 documentation in separate directories, re-organizing some of the
4880 information files, and making it build pcre-config (a GNU standard). Also added
4881 libtool support for building PCRE as a shared library, which is now the
4882 default.
4883
4884 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4885 09 are not valid octal constants. Single digits will be used for minor values
4886 less than 10.
4887
4888 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4889 existing programs that set these in the POSIX interface can use PCRE without
4890 modification.
4891
4892 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4893 return all that pcre_info() returns, plus additional data. The pcre_info()
4894 function is retained for compatibility, but is considered to be obsolete.
4895
4896 10. Added experimental recursion feature (?R) to handle one common case that
4897 Perl 5.6 will be able to do with (?p{...}).
4898
4899 11. Added support for POSIX character classes like [:alpha:], which Perl is
4900 adopting.
4901
4902
4903 Version 2.08 31-Aug-99
4904 ----------------------
4905
4906 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4907 trying to match at the startoffset position, but instead was moving forward to
4908 the next newline as if a previous match had failed.
4909
4910 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4911 and could get into a loop if a null string was matched other than at the start
4912 of the subject.
4913
4914 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4915 be distinguished at compile time, and for completeness also added PCRE_DATE.
4916
4917 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4918 in GnuWin32 environments.
4919
4920
4921 Version 2.07 29-Jul-99
4922 ----------------------
4923
4924 1. The documentation is now supplied in plain text form and HTML as well as in
4925 the form of man page sources.
4926
4927 2. C++ compilers don't like assigning (void *) values to other pointer types.
4928 In particular this affects malloc(). Although there is no problem in Standard
4929 C, I've put in casts to keep C++ compilers happy.
4930
4931 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4932 should be (const char *).
4933
4934 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4935 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4936 However, I haven't made this a standard facility. The documentation doesn't
4937 mention it, and the Makefile doesn't support it.
4938
4939 5. The Makefile now contains an "install" target, with editable destinations at
4940 the top of the file. The pcretest program is not installed.
4941
4942 6. pgrep -V now gives the PCRE version number and date.
4943
4944 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4945 causing the entire string to be ignored, instead of just the last character.
4946
4947 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4948 non-matching string, it can take a very, very long time, even for strings of
4949 quite modest length, because of the nested recursion. PCRE now does better in
4950 some of these cases. It does this by remembering the last required literal
4951 character in the pattern, and pre-searching the subject to ensure it is present
4952 before running the real match. In other words, it applies a heuristic to detect
4953 some types of certain failure quickly, and in the above example, if presented
4954 with a string that has no trailing " it gives "no match" very quickly.
4955
4956 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4957 other alternatives are tried instead.
4958
4959
4960 Version 2.06 09-Jun-99
4961 ----------------------
4962
4963 1. Change pcretest's output for amount of store used to show just the code
4964 space, because the remainder (the data block) varies in size between 32-bit and
4965 64-bit systems.
4966
4967 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4968 start matching at. This allows lookbehinds to work when searching for multiple
4969 occurrences in a string.
4970
4971 3. Added additional options to pcretest for testing multiple occurrences:
4972
4973 /+ outputs the rest of the string that follows a match
4974 /g loops for multiple occurrences, using the new startoffset argument
4975 /G loops for multiple occurrences by passing an incremented pointer
4976
4977 4. PCRE wasn't doing the "first character" optimization for patterns starting
4978 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4979 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4980 the letter 'x'. On long subject strings, this gives a significant speed-up.
4981
4982
4983 Version 2.05 21-Apr-99
4984 ----------------------
4985
4986 1. Changed the type of magic_number from int to long int so that it works
4987 properly on 16-bit systems.
4988
4989 2. Fixed a bug which caused patterns starting with .* not to work correctly
4990 when the subject string contained newline characters. PCRE was assuming
4991 anchoring for such patterns in all cases, which is not correct because .* will
4992 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4993 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4994 must be retried after every newline in the subject.
4995
4996
4997 Version 2.04 18-Feb-99
4998 ----------------------
4999
5000 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5001 computation of the store needed to hold the pattern was incorrect (too large).
5002 If such patterns were nested a few deep, this could multiply and become a real
5003 problem.
5004
5005 2. Added /M option to pcretest to show the memory requirement of a specific
5006 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5007
5008 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5009 compiled in such a way that the backtracking after subsequent failure was
5010 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5011 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5012
5013
5014 Version 2.03 02-Feb-99
5015 ----------------------
5016
5017 1. Fixed typo and small mistake in man page.
5018
5019 2. Added 4th condition (GPL supersedes if conflict) and created separate
5020 LICENCE file containing the conditions.
5021
5022 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5023 Perl, that is the internal \ allows the delimiter to be included in the
5024 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5025 the final delimiter, add \ to the end of the pattern (to test the error).
5026
5027 4. Added the convenience functions for extracting substrings after a successful
5028 match. Updated pcretest to make it able to test these functions.
5029
5030
5031 Version 2.02 14-Jan-99
5032 ----------------------
5033
5034 1. Initialized the working variables associated with each extraction so that
5035 their saving and restoring doesn't refer to uninitialized store.
5036
5037 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5038 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5039 fix the problem.
5040
5041 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5042 calls, and wasn't printing the correct value for compiling calls. Increased the
5043 default value of LOOPREPEAT, and the number of significant figures in the
5044 times.
5045
5046 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5047
5048 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5049 a building problem on Windows NT with a FAT file system.
5050
5051
5052 Version 2.01 21-Oct-98
5053 ----------------------
5054
5055 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5056 to character tables built by pcre_maketables() in the current locale. If NULL
5057 is passed, the default tables are used.
5058
5059
5060 Version 2.00 24-Sep-98
5061 ----------------------
5062
5063 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5064 it any more.
5065
5066 2. Allow quantification of (?>) groups, and make it work correctly.
5067
5068 3. The first character computation wasn't working for (?>) groups.
5069
5070 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5071 end of the subject) and add 5.005's \z, which really does match only at the
5072 very end of the subject.
5073
5074 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5075
5076 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5077 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5078 localized options. All options to pcre_study() were also removed.
5079
5080 7. Add other new features from 5.005:
5081
5082 $(?<= positive lookbehind
5083 $(?<! negative lookbehind
5084 (?imsx-imsx) added the unsetting capability
5085 such a setting is global if at outer level; local otherwise
5086 (?imsx-imsx:) non-capturing groups with option setting
5087 (?(cond)re|re) conditional pattern matching
5088
5089 A backreference to itself in a repeated group matches the previous
5090 captured string.
5091
5092 8. General tidying up of studying (both automatic and via "study")
5093 consequential on the addition of new assertions.
5094
5095 9. As in 5.005, unlimited repeated groups that could match an empty substring
5096 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5097 runtime if any iteration does actually match an empty substring.
5098
5099 10. Include the RunTest script in the distribution.
5100
5101 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5102 discrepancies, some of which were old and were also with respect to 5.004. They
5103 have now been fixed.
5104
5105
5106 Version 1.09 28-Apr-98
5107 ----------------------
5108
5109 1. A negated single character class followed by a quantifier with a minimum
5110 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5111 program crashes, or just wrong answers. This did not apply to negated classes
5112 containing more than one character, or to minima other than one.
5113
5114
5115 Version 1.08 27-Mar-98
5116 ----------------------
5117
5118 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5119
5120 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5121 latter must appear before anything that relies on it in the pattern.
5122
5123
5124 Version 1.07 16-Feb-98
5125 ----------------------
5126
5127 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5128 repeat of a potentially empty string).
5129
5130
5131 Version 1.06 23-Jan-98
5132 ----------------------
5133
5134 1. Added Markus Oberhumer's little patches for C++.
5135
5136 2. Literal strings longer than 255 characters were broken.
5137
5138
5139 Version 1.05 23-Dec-97
5140 ----------------------
5141
5142 1. Negated character classes containing more than one character were failing if
5143 PCRE_CASELESS was set at run time.
5144
5145
5146 Version 1.04 19-Dec-97
5147 ----------------------
5148
5149 1. Corrected the man page, where some "const" qualifiers had been omitted.
5150
5151 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5152 input syntax.
5153
5154 3. Fixed memory leak which occurred when a regex with back references was
5155 matched with an offsets vector that wasn't big enough. The temporary memory
5156 that is used in this case wasn't being freed if the match failed.
5157
5158 4. Tidied pcretest to ensure it frees memory that it gets.
5159
5160 5. Temporary memory was being obtained in the case where the passed offsets
5161 vector was exactly big enough.
5162
5163 6. Corrected definition of offsetof() from change 5 below.
5164
5165 7. I had screwed up change 6 below and broken the rules for the use of
5166 setjmp(). Now fixed.
5167
5168
5169 Version 1.03 18-Dec-97
5170 ----------------------
5171
5172 1. A erroneous regex with a missing opening parenthesis was correctly
5173 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5174 on some systems.
5175
5176 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5177 it was reported that one broken compiler failed on the former because "code" is
5178 also an independent variable.
5179
5180 3. The erroneous regex a[]b caused an array overrun reference.
5181
5182 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5183 fail on data ending with that character. (It was going on too far, and checking
5184 the next character, typically a binary zero.) This was specific to the
5185 optimized code for single-character negative classes.
5186
5187 5. Added a contributed patch from the TIN world which does the following:
5188
5189 + Add an undef for memmove, in case the the system defines a macro for it.
5190
5191 + Add a definition of offsetof(), in case there isn't one. (I don't know
5192 the reason behind this - offsetof() is part of the ANSI standard - but
5193 it does no harm).
5194
5195 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5196 most of the places where whitespace preceded '#'. I have given up and
5197 allowed the remaining 2 cases to be at the margin.
5198
5199 + Rename some variables in pcre to eliminate shadowing. This seems very
5200 pedantic, but does no harm, of course.
5201
5202 6. Moved the call to setjmp() into its own function, to get rid of warnings
5203 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5204
5205 7. Constructs such as \d{8,} were compiling into the equivalent of
5206 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5207 outcome, but in this particular case used more store than had been allocated,
5208 which caused the bug to be discovered because it threw up an internal error.
5209
5210 8. The debugging code in both pcre and pcretest for outputting the compiled
5211 form of a regex was going wrong in the case of back references followed by
5212 curly-bracketed repeats.
5213
5214
5215 Version 1.02 12-Dec-97
5216 ----------------------
5217
5218 1. Typos in pcre.3 and comments in the source fixed.
5219
5220 2. Applied a contributed patch to get rid of places where it used to remove
5221 'const' from variables, and fixed some signed/unsigned and uninitialized
5222 variable warnings.
5223
5224 3. Added the "runtest" target to Makefile.
5225
5226 4. Set default compiler flag to -O2 rather than just -O.
5227
5228
5229 Version 1.01 19-Nov-97
5230 ----------------------
5231
5232 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5233 like /([ab]*)*/, that is, for classes with more than one character in them.
5234
5235 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5236 as /((?>a*))*/ (a PCRE_EXTRA facility).
5237
5238
5239 Version 1.00 18-Nov-97
5240 ----------------------
5241
5242 1. Added compile-time macros to support systems such as SunOS4 which don't have
5243 memmove() or strerror() but have other things that can be used instead.
5244
5245 2. Arranged that "make clean" removes the executables.
5246
5247
5248 Version 0.99 27-Oct-97
5249 ----------------------
5250
5251 1. Fixed bug in code for optimizing classes with only one character. It was
5252 initializing a 32-byte map regardless, which could cause it to run off the end
5253 of the memory it had got.
5254
5255 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5256
5257
5258 Version 0.98 22-Oct-97
5259 ----------------------
5260
5261 1. Fixed bug in code for handling temporary memory usage when there are more
5262 back references than supplied space in the ovector. This could cause segfaults.
5263
5264
5265 Version 0.97 21-Oct-97
5266 ----------------------
5267
5268 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5269
5270 2. Optimized negated single characters not to use a bit map.
5271
5272 3. Brought error texts together as macro definitions; clarified some of them;
5273 fixed one that was wrong - it said "range out of order" when it meant "invalid
5274 escape sequence".
5275
5276 4. Changed some char * arguments to const char *.
5277
5278 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5279
5280 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5281 pcretest.
5282
5283
5284 Version 0.96 16-Oct-97
5285 ----------------------
5286
5287 1. Added a simple "pgrep" utility to the distribution.
5288
5289 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5290 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5291 where "ddd" means "one or more decimal digits".
5292
5293 3. Fixed serious bug. If a pattern had a back reference, but the call to
5294 pcre_exec() didn't supply a large enough ovector to record the related
5295 identifying subpattern, the match always failed. PCRE now remembers the number
5296 of the largest back reference, and gets some temporary memory in which to save
5297 the offsets during matching if necessary, in order to ensure that
5298 backreferences always work.
5299
5300 4. Increased the compatibility with Perl in a number of ways:
5301
5302 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5303 to request this handling. The option can be set at compile or exec time.
5304
5305 (b) $ matches before a terminating newline by default; an option
5306 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5307 mode). The option can be set at compile or exec time.
5308
5309 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5310 the same as Perl's. If the decimal number it represents is less than 10
5311 or there aren't that many previous left capturing parentheses, an octal
5312 escape is read. Inside a character class, it's always an octal escape,
5313 even if it is a single digit.
5314
5315 (d) An escaped but undefined alphabetic character is taken as a literal,
5316 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5317 escapes.
5318
5319 (e) {0} is now permitted. (The previous item is removed from the compiled
5320 pattern).
5321
5322 5. Changed all the names of code files so that the basic parts are no longer
5323 than 10 characters, and abolished the teeny "globals.c" file.
5324
5325 6. Changed the handling of character classes; they are now done with a 32-byte
5326 bit map always.
5327
5328 7. Added the -d and /D options to pcretest to make it possible to look at the
5329 internals of compilation without having to recompile pcre.
5330
5331
5332 Version 0.95 23-Sep-97
5333 ----------------------
5334
5335 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5336 \x20 at the start of a run of normal characters. These were being treated as
5337 real characters, instead of the source characters being re-checked.
5338
5339
5340 Version 0.94 18-Sep-97
5341 ----------------------
5342
5343 1. The functions are now thread-safe, with the caveat that the global variables
5344 containing pointers to malloc() and free() or alternative functions are the
5345 same for all threads.
5346
5347 2. Get pcre_study() to generate a bitmap of initial characters for non-
5348 anchored patterns when this is possible, and use it if passed to pcre_exec().
5349
5350
5351 Version 0.93 15-Sep-97
5352 ----------------------
5353
5354 1. /(b)|(:+)/ was computing an incorrect first character.
5355
5356 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5357 but not actually doing anything yet.
5358
5359 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5360 as Perl does (e.g. [-az] or [az-]).
5361
5362 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5363 all possible positions.
5364
5365 5. Split up into different modules to avoid including unneeded functions in a
5366 compiled binary. However, compile and exec are still in one module. The "study"
5367 function is split off.
5368
5369 6. The character tables are now in a separate module whose source is generated
5370 by an auxiliary program - but can then be edited by hand if required. There are
5371 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5372 toupper() in the code.
5373
5374 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5375 make them global. Abolish the function for setting them, as the caller can now
5376 set them directly.
5377
5378
5379 Version 0.92 11-Sep-97
5380 ----------------------
5381
5382 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5383 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5384
5385 2. Caseless matching was not working in character classes if the characters in
5386 the pattern were in upper case.
5387
5388 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5389
5390 4. Make PCRE_ANCHORED public and accept as a compile option.
5391
5392 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5393 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5394 pass them.
5395
5396 6. Give an error if bad option bits passed at compile or run time.
5397
5398 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5399 pcretest to cause it to pass that flag.
5400
5401 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5402 options, and the first character, if set.
5403
5404 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5405
5406
5407 Version 0.91 10-Sep-97
5408 ----------------------
5409
5410 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5411 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5412
5413 2. PCRE was looping on encountering an indefinitely repeated back reference to
5414 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5415 Perl does - treats the match as successful.
5416
5417 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5