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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1488 - (show annotations)
Wed Jun 18 18:38:00 2014 UTC (5 years, 3 months ago) by ph10
File size: 267534 byte(s)
Error occurred while calculating annotation data.
Recognize characters with multiple other cases when creating starting bit map.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.36 xx-xxx-2014
5 ------------------------
6
7 1. Got rid of some compiler warnings in the C++ modules that were shown up by
8 -Wmissing-field-initializers and -Wunused-parameter.
9
10 2. The tests for quantifiers being too big (greater than 65535) were being
11 applied after reading the number, and stupidly assuming that integer
12 overflow would give a negative number. The tests are now applied as the
13 numbers are read.
14
15 3. Tidy code in pcre_exec.c where two branches that used to be different are
16 now the same.
17
18 4. The JIT compiler did not generate match limit checks for certain
19 bracketed expressions with quantifiers. This may lead to exponential
20 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
21 issue should be resolved now.
22
23 5. Fixed an issue, which occures when nested alternatives are optimized
24 with table jumps.
25
26 6. Inserted two casts and changed some ints to size_t in the light of some
27 reported 64-bit compiler warnings (Bugzilla 1477).
28
29 7. Fixed a bug concerned with zero-minimum possessive groups that could match
30 an empty string, which sometimes were behaving incorrectly in the
31 interpreter (though correctly in the JIT matcher). This pcretest input is
32 an example:
33
34 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
35 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
36
37 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
38 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
39 for an empty string was breaking the inner loop and carrying on at a lower
40 level, when possessive repeated groups should always return to a higher
41 level as they have no backtrack points in them. The empty string test now
42 occurs at the outer level.
43
44 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
45 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
46
47 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
48 doesn't).
49
50 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
51 to hold a pattern to be incorrectly computed (too small) when there were
52 named back references to duplicated names. This could cause "internal
53 error: code overflow" or "double free or corruption" or other memory
54 handling errors.
55
56 11. When named subpatterns had the same prefixes, back references could be
57 confused. For example, in this pattern:
58
59 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
60
61 the reference to 'Name' was incorrectly treated as a reference to a
62 duplicate name.
63
64 12. A pattern such as /^s?c/mi8 where the optional character has more than
65 one "other case" was incorrectly compiled such that it would only try to
66 match starting at "c".
67
68 13. When a pattern starting with \s was studied, VT was not included in the
69 list of possible starting characters; this should have been part of the
70 8.34/18 patch.
71
72 14. If a character class started [\Qx]... where x is any character, the class
73 was incorrectly terminated at the ].
74
75 15. If a pattern that started with a caseless match for a character with more
76 than one "other case" was studied, PCRE did not set up the starting code
77 unit bit map for the list of possible characters. Now it does. This is an
78 optimization improvement, not a bug fix.
79
80
81 Version 8.35 04-April-2014
82 --------------------------
83
84 1. A new flag is set, when property checks are present in an XCLASS.
85 When this flag is not set, PCRE can perform certain optimizations
86 such as studying these XCLASS-es.
87
88 2. The auto-possessification of character sets were improved: a normal
89 and an extended character set can be compared now. Furthermore
90 the JIT compiler optimizes more character set checks.
91
92 3. Got rid of some compiler warnings for potentially uninitialized variables
93 that show up only when compiled with -O2.
94
95 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
96 of a match later then the end of the match. The pcretest program was not
97 handling the case sensibly - it was outputting from the start to the next
98 binary zero. It now reports this situation in a message, and outputs the
99 text from the end to the start.
100
101 5. Fast forward search is improved in JIT. Instead of the first three
102 characters, any three characters with fixed position can be searched.
103 Search order: first, last, middle.
104
105 6. Improve character range checks in JIT. Characters are read by an inprecise
106 function now, which returns with an unknown value if the character code is
107 above a certain treshold (e.g: 256). The only limitation is that the value
108 must be bigger than the treshold as well. This function is useful, when
109 the characters above the treshold are handled in the same way.
110
111 7. The macros whose names start with RAWUCHAR are placeholders for a future
112 mode in which only the bottom 21 bits of 32-bit data items are used. To
113 make this more memorable for those maintaining the code, the names have
114 been changed to start with UCHAR21, and an extensive comment has been added
115 to their definition.
116
117 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
118 to the export list in Makefile.am (they were accidentally omitted from the
119 8.34 tarball).
120
121 9. The informational output from pcretest used the phrase "starting byte set"
122 which is inappropriate for the 16-bit and 32-bit libraries. As the output
123 for "first char" and "need char" really means "non-UTF-char", I've changed
124 "byte" to "char", and slightly reworded the output. The documentation about
125 these values has also been (I hope) clarified.
126
127 10. Another JIT related optimization: use table jumps for selecting the correct
128 backtracking path, when more than four alternatives are present inside a
129 bracket.
130
131 11. Empty match is not possible, when the minimum length is greater than zero,
132 and there is no \K in the pattern. JIT should avoid empty match checks in
133 such cases.
134
135 12. In a caseless character class with UCP support, when a character with more
136 than one alternative case was not the first character of a range, not all
137 the alternative cases were added to the class. For example, s and \x{17f}
138 are both alternative cases for S: the class [RST] was handled correctly,
139 but [R-T] was not.
140
141 13. The configure.ac file always checked for pthread support when JIT was
142 enabled. This is not used in Windows, so I have put this test inside a
143 check for the presence of windows.h (which was already tested for).
144
145 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
146 The algorithm provides a way to skip certain starting offsets, and usually
147 faster than linear prefix searches.
148
149 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
150 as for 'fr_FR' and 'french'. For some reason, however, it then used the
151 Windows-specific input and output files, which have 'french' screwed in.
152 So this could never have worked. One of the problems with locales is that
153 they aren't always the same. I have now updated RunTest so that it checks
154 the output of the locale test (test 3) against three different output
155 files, and it allows the test to pass if any one of them matches. With luck
156 this should make the test pass on some versions of Solaris where it was
157 failing. Because of the uncertainty, the script did not used to stop if
158 test 3 failed; it now does. If further versions of a French locale ever
159 come to light, they can now easily be added.
160
161 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
162 there was a message during ./configure, but it did not stop. This now
163 provokes an error. The invalid example in README has been corrected.
164 If a value less than the minimum is given, the minimum value has always
165 been used, but now a warning is given.
166
167 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
168 was a bug in the test system, which is now fixed. Also, the list of various
169 configurations that are tested for each release did not have one with both
170 16/32 bits and --enable-bar-anycrlf. It now does.
171
172 18. pcretest was missing "-C bsr" for displaying the \R default setting.
173
174 19. Little endian PowerPC systems are supported now by the JIT compiler.
175
176 20. The fast forward newline mechanism could enter to an infinite loop on
177 certain invalid UTF-8 input. Although we don't support these cases
178 this issue can be fixed by a performance optimization.
179
180 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
181 not take account if existing stack usage. There is now a new global
182 variable called pcre_stack_guard that can be set to point to an external
183 function to check stack availability. It is called at the start of
184 processing every parenthesized group.
185
186 22. A typo in the code meant that in ungreedy mode the max/min qualifier
187 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
188 not match "ab".
189
190 23. When UTF was disabled, the JIT program reported some incorrect compile
191 errors. These messages are silenced now.
192
193 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
194 compiler.
195
196 25. Change all the temporary files used in RunGrepTest to be different to those
197 used by RunTest so that the tests can be run simultaneously, for example by
198 "make -j check".
199
200
201 Version 8.34 15-December-2013
202 -----------------------------
203
204 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
205 executable memory. Patch inspired by Carsten Klein.
206
207 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
208 this macro is never tested and has no effect, because the work to support
209 coverage involves only compiling and linking options and special targets in
210 the Makefile. The comment in config.h implied that defining the macro would
211 enable coverage support, which is totally false. There was also support for
212 setting this macro in the CMake files (my fault, I just copied it from
213 configure). SUPPORT_GCOV has now been removed.
214
215 3. Make a small performance improvement in strlen16() and strlen32() in
216 pcretest.
217
218 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
219 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
220 these cases). There was also one in pcretest.c.
221
222 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
223
224 6. In UTF mode, the code for checking whether a group could match an empty
225 string (which is used for indefinitely repeated groups to allow for
226 breaking an infinite loop) was broken when the group contained a repeated
227 negated single-character class with a character that occupied more than one
228 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
229 UTF-8 mode). The effect was undefined: the group might or might not be
230 deemed as matching an empty string, or the program might have crashed.
231
232 7. The code for checking whether a group could match an empty string was not
233 recognizing that \h, \H, \v, \V, and \R must match a character.
234
235 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
236 an empty string. If it can, pcretest shows this in its information output.
237
238 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
239 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
240 matched by pcre_exec() without using JIT:
241
242 (a) If the rest of the pattern did not match after a maximal run of
243 grapheme clusters, the code for backing up to try with fewer of them
244 did not always back up over a full grapheme when characters that do not
245 have the modifier quality were involved, e.g. Hangul syllables.
246
247 (b) If the match point in a subject started with modifier character, and
248 there was no match, the code could incorrectly back up beyond the match
249 point, and potentially beyond the first character in the subject,
250 leading to a segfault or an incorrect match result.
251
252 10. A conditional group with an assertion condition could lead to PCRE
253 recording an incorrect first data item for a match if no other first data
254 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
255 first data item, and therefore matched "ca" after "c" instead of at the
256 start.
257
258 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
259 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
260
261 12. The source of pcregrep now includes z/OS-specific code so that it can be
262 compiled for z/OS as part of the special z/OS distribution.
263
264 13. Added the -T and -TM options to pcretest.
265
266 14. The code in pcre_compile.c for creating the table of named capturing groups
267 has been refactored. Instead of creating the table dynamically during the
268 actual compiling pass, the information is remembered during the pre-compile
269 pass (on the stack unless there are more than 20 named groups, in which
270 case malloc() is used) and the whole table is created before the actual
271 compile happens. This has simplified the code (it is now nearly 150 lines
272 shorter) and prepared the way for better handling of references to groups
273 with duplicate names.
274
275 15. A back reference to a named subpattern when there is more than one of the
276 same name now checks them in the order in which they appear in the pattern.
277 The first one that is set is used for the reference. Previously only the
278 first one was inspected. This change makes PCRE more compatible with Perl.
279
280 16. Unicode character properties were updated from Unicode 6.3.0.
281
282 17. The compile-time code for auto-possessification has been refactored, based
283 on a patch by Zoltan Herczeg. It now happens after instead of during
284 compilation. The code is cleaner, and more cases are handled. The option
285 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
286 options in pcretest are provided to set it. It can also be set by
287 (*NO_AUTO_POSSESS) at the start of a pattern.
288
289 18. The character VT has been added to the default ("C" locale) set of
290 characters that match \s and are generally treated as white space,
291 following this same change in Perl 5.18. There is now no difference between
292 "Perl space" and "POSIX space". Whether VT is treated as white space in
293 other locales depends on the locale.
294
295 19. The code for checking named groups as conditions, either for being set or
296 for being recursed, has been refactored (this is related to 14 and 15
297 above). Processing unduplicated named groups should now be as fast at
298 numerical groups, and processing duplicated groups should be faster than
299 before.
300
301 20. Two patches to the CMake build system, by Alexander Barkov:
302
303 (1) Replace the "source" command by "." in CMakeLists.txt because
304 "source" is a bash-ism.
305
306 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
307 without these the CMake build does not work on Solaris.
308
309 21. Perl has changed its handling of \8 and \9. If there is no previously
310 encountered capturing group of those numbers, they are treated as the
311 literal characters 8 and 9 instead of a binary zero followed by the
312 literals. PCRE now does the same.
313
314 22. Following Perl, added \o{} to specify codepoints in octal, making it
315 possible to specify values greater than 0777 and also making them
316 unambiguous.
317
318 23. Perl now gives an error for missing closing braces after \x{... instead of
319 treating the string as literal. PCRE now does the same.
320
321 24. RunTest used to grumble if an inappropriate test was selected explicitly,
322 but just skip it when running all tests. This make it awkward to run ranges
323 of tests when one of them was inappropriate. Now it just skips any
324 inappropriate tests, as it always did when running all tests.
325
326 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
327 character types such as \d or \w, too many callouts were inserted, and the
328 data that they returned was rubbish.
329
330 26. In UCP mode, \s was not matching two of the characters that Perl matches,
331 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
332 were matched by \h. The code has now been refactored so that the lists of
333 the horizontal and vertical whitespace characters used for \h and \v (which
334 are defined only in one place) are now also used for \s.
335
336 27. Add JIT support for the 64 bit TileGX architecture.
337 Patch by Jiong Wang (Tilera Corporation).
338
339 28. Possessive quantifiers for classes (both explicit and automatically
340 generated) now use special opcodes instead of wrapping in ONCE brackets.
341
342 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
343 (because it's meaningless), this was not happening when PCRE_CASELESS was
344 set. Not wrong, but inefficient.
345
346 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
347 properties for \w, \d, etc) is present in a test regex. Otherwise if the
348 test contains no characters greater than 255, Perl doesn't realise it
349 should be using Unicode semantics.
350
351 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
352 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
353 does in Unicode mode.
354
355 32. Added the "forbid" facility to pcretest so that putting tests into the
356 wrong test files can sometimes be quickly detected.
357
358 33. There is now a limit (default 250) on the depth of nesting of parentheses.
359 This limit is imposed to control the amount of system stack used at compile
360 time. It can be changed at build time by --with-parens-nest-limit=xxx or
361 the equivalent in CMake.
362
363 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
364 errors. Perl warns for these when in warning mode, but PCRE has no facility
365 for giving warnings.
366
367 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
368 However, this was not working for (?!) because it is optimized to (*FAIL),
369 for which PCRE does not allow quantifiers. The optimization is now disabled
370 when a quantifier follows (?!). I can't see any use for this, but it makes
371 things uniform.
372
373 36. Perl no longer allows group names to start with digits, so I have made this
374 change also in PCRE. It simplifies the code a bit.
375
376 37. In extended mode, Perl ignores spaces before a + that indicates a
377 possessive quantifier. PCRE allowed a space before the quantifier, but not
378 before the possessive +. It now does.
379
380 38. The use of \K (reset reported match start) within a repeated possessive
381 group such as (a\Kb)*+ was not working.
382
383 40. Document that the same character tables must be used at compile time and
384 run time, and that the facility to pass tables to pcre_exec() and
385 pcre_dfa_exec() is for use only with saved/restored patterns.
386
387 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
388 features for Builds with MSVC:
389
390 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
391 builds). The .rc files can be used to set FileDescription and many other
392 attributes.
393
394 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
395 This allows higher-level build scripts which want .pdb files to avoid
396 hard-coding the exact files needed."
397
398 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
399 mean "start of word" and "end of word", respectively, as a transition aid.
400
401 43. A minimizing repeat of a class containing codepoints greater than 255 in
402 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
403 compiled to use the heap for recursion.
404
405 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
406 is configured.
407
408
409 Version 8.33 28-May-2013
410 ------------------------
411
412 1. Added 'U' to some constants that are compared to unsigned integers, to
413 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
414 variables that are added to signed variables, to ensure the result is
415 signed and can be negated.
416
417 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
418 CMake config files.
419
420 3. Revise the creation of config.h.generic so that all boolean macros are
421 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
422 overriding via -D on the command line possible.
423
424 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
425 to unsigned int is reported to make a quite noticeable speed difference in
426 a specific Windows environment. Testing on Linux did also appear to show
427 some benefit (and it is clearly not harmful). Also fixed the definition of
428 Xop which should be unsigned.
429
430 5. Related to (4), changing the definition of the intermediate variable cc
431 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
432 improvements.
433
434 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
435 unnecessary spaces.
436
437 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
438 and later.
439
440 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
441
442 9. Optimizing fast_forward_start_bits in JIT.
443
444 10. Adding support for callouts in JIT, and fixing some issues revealed
445 during this work. Namely:
446
447 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
448
449 (b) Minimum length was not checked before the matching is started.
450
451 11. The value of capture_last that is passed to callouts was incorrect in some
452 cases when there was a capture on one path that was subsequently abandoned
453 after a backtrack. Also, the capture_last value is now reset after a
454 recursion, since all captures are also reset in this case.
455
456 12. The interpreter no longer returns the "too many substrings" error in the
457 case when an overflowing capture is in a branch that is subsequently
458 abandoned after a backtrack.
459
460 13. In the pathological case when an offset vector of size 2 is used, pcretest
461 now prints out the matched string after a yield of 0 or 1.
462
463 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
464 Only supported by the JIT compiler at the moment.
465
466 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
467
468 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
469 offset of the starting point of the matching process, provided the offsets
470 vector is large enough.
471
472 17. The \A escape now records a lookbehind value of 1, though its execution
473 does not actually inspect the previous character. This is to ensure that,
474 in partial multi-segment matching, at least one character from the old
475 segment is retained when a new segment is processed. Otherwise, if there
476 are no lookbehinds in the pattern, \A might match incorrectly at the start
477 of a new segment.
478
479 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
480
481 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
482 gives some modest performance improvement in 8-bit mode.
483
484 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
485 be expressed in certain programming languages using Universal Character
486 Names.
487
488 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
489 which points out that "non characters" are not "characters that may not
490 appear in Unicode strings" but rather "characters that are reserved for
491 internal use and have only local meaning".
492
493 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
494 there was a conditional group that depended on an assertion, if the
495 assertion was false, the callout that immediately followed the alternation
496 in the condition was skipped when pcre_exec() was used for matching.
497
498 23. Allow an explicit callout to be inserted before an assertion that is the
499 condition for a conditional group, for compatibility with automatic
500 callouts, which always insert a callout at this point.
501
502 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
503 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
504
505 25. (*PRUNE) is now supported by the JIT compiler.
506
507 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
508
509 27. Fix the case where there are two or more SKIPs with arguments that may be
510 ignored.
511
512 28. (*SKIP) is now supported by the JIT compiler.
513
514 29. (*THEN) is now supported by the JIT compiler.
515
516 30. Update RunTest with additional test selector options.
517
518 31. The way PCRE handles backtracking verbs has been changed in two ways.
519
520 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
521 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
522 backtracking. In some cases this makes it more Perl-compatible, but Perl's
523 rather obscure rules do not always do the same thing.
524
525 (2) Previously, backtracking verbs were confined within assertions. This is
526 no longer the case for positive assertions, except for (*ACCEPT). Again,
527 this sometimes improves Perl compatibility, and sometimes does not.
528
529 32. A number of tests that were in test 2 because Perl did things differently
530 have been moved to test 1, because either Perl or PCRE has changed, and
531 these tests are now compatible.
532
533 32. Backtracking control verbs are now handled in the same way in JIT and
534 interpreter.
535
536 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
537 contained a forward subroutine reference caused a compile error.
538
539 34. Auto-detect and optimize limited repetitions in JIT.
540
541 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
542 blocking (*UTF) etc.
543
544 36. In the interpreter, maximizing pattern repetitions for characters and
545 character types now use tail recursion, which reduces stack usage.
546
547 37. The value of the max lookbehind was not correctly preserved if a compiled
548 and saved regex was reloaded on a host of different endianness.
549
550 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
551 of the compiled pattern block, expand the flags field from 16 to 32 bits
552 because it was almost full.
553
554 39. Try madvise first before posix_madvise.
555
556 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
557 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
558 original change locked out all matches of empty strings. This has been
559 changed so that one match of an empty string per line is recognized.
560 Subsequent searches on the same line (for colouring or for --only-matching,
561 for example) do not recognize empty strings.
562
563 41. Applied a user patch to fix a number of spelling mistakes in comments.
564
565 42. Data lines longer than 65536 caused pcretest to crash.
566
567 43. Clarified the data type for length and startoffset arguments for pcre_exec
568 and pcre_dfa_exec in the function-specific man pages, where they were
569 explicitly stated to be in bytes, never having been updated. I also added
570 some clarification to the pcreapi man page.
571
572 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
573 a segmentation fault.
574
575
576 Version 8.32 30-November-2012
577 -----------------------------
578
579 1. Improved JIT compiler optimizations for first character search and single
580 character iterators.
581
582 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
583 Patch by Daniel Richard G.
584
585 3. Single character iterator optimizations in the JIT compiler.
586
587 4. Improved JIT compiler optimizations for character ranges.
588
589 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
590 Reported by Giuseppe D'Angelo.
591
592 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
593 of a line, was being set incorrectly in cases where .* appeared inside
594 atomic brackets at the start of a pattern, or where there was a subsequent
595 *PRUNE or *SKIP.
596
597 7. Improved instruction cache flush for POWER/PowerPC.
598 Patch by Daniel Richard G.
599
600 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
601 grep:
602
603 (a) There is now no limit to the number of patterns to be matched.
604
605 (b) An error is given if a pattern is too long.
606
607 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
608 are now supported.
609
610 (d) --exclude-from and --include-from (multiple use) have been added.
611
612 (e) Exclusions and inclusions now apply to all files and directories, not
613 just to those obtained from scanning a directory recursively.
614
615 (f) Multiple uses of -f and --file-list are now supported.
616
617 (g) In a Windows environment, the default for -d has been changed from
618 "read" (the GNU grep default) to "skip", because otherwise the presence
619 of a directory in the file list provokes an error.
620
621 (h) The documentation has been revised and clarified in places.
622
623 9. Improve the matching speed of capturing brackets.
624
625 10. Changed the meaning of \X so that it now matches a Unicode extended
626 grapheme cluster.
627
628 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
629 out POSIX threads when JIT support is configured.
630
631 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
632
633 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
634 structure could go wrong in environments where size_t is not the same size
635 as int.
636
637 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
638 set.
639
640 15. The EBCDIC support had decayed; later updates to the code had included
641 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
642 general tidy up of EBCDIC-related issues, and the documentation was also
643 not quite right. There is now a test that can be run on ASCII systems to
644 check some of the EBCDIC-related things (but is it not a full test).
645
646 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
647 in a small tidy to the code.
648
649 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
650
651 18. If the --only-matching (-o) option in pcregrep is specified multiple
652 times, each one causes appropriate output. For example, -o1 -o2 outputs the
653 substrings matched by the 1st and 2nd capturing parentheses. A separating
654 string can be specified by --om-separator (default empty).
655
656 19. Improving the first n character searches.
657
658 20. Turn case lists for horizontal and vertical white space into macros so that
659 they are defined only once.
660
661 21. This set of changes together give more compatible Unicode case-folding
662 behaviour for characters that have more than one other case when UCP
663 support is available.
664
665 (a) The Unicode property table now has offsets into a new table of sets of
666 three or more characters that are case-equivalent. The MultiStage2.py
667 script that generates these tables (the pcre_ucd.c file) now scans
668 CaseFolding.txt instead of UnicodeData.txt for character case
669 information.
670
671 (b) The code for adding characters or ranges of characters to a character
672 class has been abstracted into a generalized function that also handles
673 case-independence. In UTF-mode with UCP support, this uses the new data
674 to handle characters with more than one other case.
675
676 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
677 whose other case is greater than 256 are now correctly matched
678 caselessly. Previously, the high codepoint matched the low one, but not
679 vice versa.
680
681 (d) The processing of \h, \H, \v, and \ in character classes now makes use
682 of the new class addition function, using character lists defined as
683 macros alongside the case definitions of 20 above.
684
685 (e) Caseless back references now work with characters that have more than
686 one other case.
687
688 (f) General caseless matching of characters with more than one other case
689 is supported.
690
691 22. Unicode character properties were updated from Unicode 6.2.0
692
693 23. Improved CMake support under Windows. Patch by Daniel Richard G.
694
695 24. Add support for 32-bit character strings, and UTF-32
696
697 25. Major JIT compiler update (code refactoring and bugfixing).
698 Experimental Sparc 32 support is added.
699
700 26. Applied a modified version of Daniel Richard G's patch to create
701 pcre.h.generic and config.h.generic by "make" instead of in the
702 PrepareRelease script.
703
704 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
705 pcre_compile.c when checking for a zero character.
706
707 28. Introducing a native interface for JIT. Through this interface, the compiled
708 machine code can be directly executed. The purpose of this interface is to
709 provide fast pattern matching, so several sanity checks are not performed.
710 However, feature tests are still performed. The new interface provides
711 1.4x speedup compared to the old one.
712
713 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
714 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
715 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
716
717 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
718 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
719 the "old" RFC 2279). Instead, it ended up passing a negative length to
720 pcre_exec().
721
722 31. Add support for GCC's visibility feature to hide internal functions.
723
724 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
725 "unknown -C option" after outputting 0 or 1.
726
727 33. There is now support for generating a code coverage report for the test
728 suite in environments where gcc is the compiler and lcov is installed. This
729 is mainly for the benefit of the developers.
730
731 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
732 unaddressable using valgrind annotations, allowing valgrind to detect
733 invalid memory accesses. This is mainly for the benefit of the developers.
734
735 25. (*UTF) can now be used to start a pattern in any of the three libraries.
736
737 26. Give configure error if --enable-cpp but no C++ compiler found.
738
739
740 Version 8.31 06-July-2012
741 -------------------------
742
743 1. Fixing a wrong JIT test case and some compiler warnings.
744
745 2. Removed a bashism from the RunTest script.
746
747 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
748 to unsigned type, result still unsigned" that was given by an MS compiler
749 on encountering the code "-sizeof(xxx)".
750
751 4. Partial matching support is added to the JIT compiler.
752
753 5. Fixed several bugs concerned with partial matching of items that consist
754 of more than one character:
755
756 (a) /^(..)\1/ did not partially match "aba" because checking references was
757 done on an "all or nothing" basis. This also applied to repeated
758 references.
759
760 (b) \R did not give a hard partial match if \r was found at the end of the
761 subject.
762
763 (c) \X did not give a hard partial match after matching one or more
764 characters at the end of the subject.
765
766 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
767 a partial match for the string "\r".
768
769 (e) When newline was set to CRLF, the metacharacter "." did not recognize
770 a partial match for a CR character at the end of the subject string.
771
772 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
773 running pcretest, the text "(JIT)" added to the output whenever JIT is
774 actually used to run the match.
775
776 7. Individual JIT compile options can be set in pcretest by following -s+[+]
777 or /S+[+] with a digit between 1 and 7.
778
779 8. OP_NOT now supports any UTF character not just single-byte ones.
780
781 9. (*MARK) control verb is now supported by the JIT compiler.
782
783 10. The command "./RunTest list" lists the available tests without actually
784 running any of them. (Because I keep forgetting what they all are.)
785
786 11. Add PCRE_INFO_MAXLOOKBEHIND.
787
788 12. Applied a (slightly modified) user-supplied patch that improves performance
789 when the heap is used for recursion (compiled with --disable-stack-for-
790 recursion). Instead of malloc and free for each heap frame each time a
791 logical recursion happens, frames are retained on a chain and re-used where
792 possible. This sometimes gives as much as 30% improvement.
793
794 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
795 call.
796
797 14. As documented, (*COMMIT) is now confined to within a positive assertion.
798
799 15. It is now possible to link pcretest with libedit as an alternative to
800 libreadline.
801
802 16. (*COMMIT) control verb is now supported by the JIT compiler.
803
804 17. The Unicode data tables have been updated to Unicode 6.1.0.
805
806 18. Added --file-list option to pcregrep.
807
808 19. Added binary file support to pcregrep, including the -a, --binary-files,
809 -I, and --text options.
810
811 20. The madvise function is renamed for posix_madvise for QNX compatibility
812 reasons. Fixed by Giuseppe D'Angelo.
813
814 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
815
816 22. Changed the option for creating symbolic links for 16-bit man pages from
817 -s to -sf so that re-installing does not cause issues.
818
819 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
820
821 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
822 matches in certain environments (the workspace was not being correctly
823 retained). Also added to pcre_dfa_exec() a simple plausibility check on
824 some of the workspace data at the beginning of a restart.
825
826 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
827 was not doing so when it should - probably a typo introduced by SVN 528
828 (change 8.10/14).
829
830 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
831 \w+ when the character tables indicated that \x{c4} was a word character.
832 There were several related cases, all because the tests for doing a table
833 lookup were testing for characters less than 127 instead of 255.
834
835 27. If a pattern contains capturing parentheses that are not used in a match,
836 their slots in the ovector are set to -1. For those that are higher than
837 any matched groups, this happens at the end of processing. In the case when
838 there were back references that the ovector was too small to contain
839 (causing temporary malloc'd memory to be used during matching), and the
840 highest capturing number was not used, memory off the end of the ovector
841 was incorrectly being set to -1. (It was using the size of the temporary
842 memory instead of the true size.)
843
844 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
845 ovector size, it uses memory at the end of the block that it has got.
846
847 29. Check for an overlong MARK name and give an error at compile time. The
848 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
849
850 30. JIT compiler update.
851
852 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
853 Rill for the patch.
854
855 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
856
857 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
858
859 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
860 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
861 was enabled.
862
863 35. Improve JIT code generation for greedy plus quantifier.
864
865 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
866 1 to "aa" instead of to an empty string. The bug affected repeated groups
867 that could potentially match an empty string.
868
869 37. Optimizing single character iterators in JIT.
870
871 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
872 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
873 codepoints that are too big for the mode are faulted, and in a UTF mode,
874 disallowed codepoints are also faulted.
875
876 39. If PCRE was compiled with UTF support, in three places in the DFA
877 matcher there was code that should only have been obeyed in UTF mode, but
878 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
879 processing when bytes with values greater than 127 were present. In 16-bit
880 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
881 both cases the values are those that cannot be the first data item in a UTF
882 character. The three items that might have provoked this were recursions,
883 possessively repeated groups, and atomic groups.
884
885 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
886 and pcregrep, because some OS require shared objects to be explicitly
887 passed to ld, causing the link step to fail if they are not.
888
889 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
890 mode, patterns that started with \h* or \R* might be incorrectly matched.
891
892
893 Version 8.30 04-February-2012
894 -----------------------------
895
896 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
897 name is defined in ctype.h.
898
899 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
900 only in quite long subpatterns.
901
902 3. Removed the function pcre_info(), which has been obsolete and deprecated
903 since it was replaced by pcre_fullinfo() in February 2000.
904
905 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
906 match a (*MARK), and the match failed at the start of the subject, a
907 reference to memory before the start of the subject could occur. This bug
908 was introduced by fix 17 of release 8.21.
909
910 5. A reference to an unset group with zero minimum repetition was giving
911 totally wrong answers (in non-JavaScript-compatibility mode). For example,
912 /(another)?(\1?)test/ matched against "hello world test". This bug was
913 introduced in release 8.13.
914
915 6. Add support for 16-bit character strings (a large amount of work involving
916 many changes and refactorings).
917
918 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
919 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
920 from a file.
921
922 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
923 rounding is not applied in this particular case).
924
925 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
926 if they appear, or are escaped, in patterns.
927
928 10. Get rid of a number of -Wunused-but-set-variable warnings.
929
930 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
931 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
932 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
933 also returns the mark "x". This bug applied to capturing parentheses,
934 non-capturing parentheses, and atomic parentheses. It also applied to some
935 assertions.
936
937 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
938 information out of configure.ac instead of relying on pcre.h.generic, which
939 is not stored in the repository.
940
941 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
942 -lreadline.
943
944 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
945
946 15. Applied Graycode's patch to put the top-level frame on the stack rather
947 than the heap when not using the stack for recursion. This gives a
948 performance improvement in many cases when recursion is not deep.
949
950 16. Experimental code added to "pcretest -C" to output the stack frame size.
951
952
953 Version 8.21 12-Dec-2011
954 ------------------------
955
956 1. Updating the JIT compiler.
957
958 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
959 are added as well.
960
961 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
962 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
963 calling _pcre_jit_exec. Some extra comments are added.
964
965 4. (*MARK) settings inside atomic groups that do not contain any capturing
966 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
967 was introduced by change 18 for 8.20.
968
969 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
970 ECMA-262 standard.
971
972 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
973 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
974 This bug was probably introduced by change 9 of 8.13.
975
976 7. While fixing 6 above, I noticed that a number of other items were being
977 incorrectly rejected as "not fixed length". This arose partly because newer
978 opcodes had not been added to the fixed-length checking code. I have (a)
979 corrected the bug and added tests for these items, and (b) arranged for an
980 error to occur if an unknown opcode is encountered while checking for fixed
981 length instead of just assuming "not fixed length". The items that were
982 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
983 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
984 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
985
986 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
987 being incorrectly compiled and would have given unpredicatble results.
988
989 9. A possessively repeated subpattern with minimum repeat count greater than
990 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
991 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
992 the first (A) could occur when it should not.
993
994 10. Add a cast and remove a redundant test from the code.
995
996 11. JIT should use pcre_malloc/pcre_free for allocation.
997
998 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
999 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1000 is anything other than /usr, -L is still shown).
1001
1002 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1003
1004 14. Perl does not support \N without a following name in a [] class; PCRE now
1005 also gives an error.
1006
1007 15. If a forward reference was repeated with an upper limit of around 2000,
1008 it caused the error "internal error: overran compiling workspace". The
1009 maximum number of forward references (including repeats) was limited by the
1010 internal workspace, and dependent on the LINK_SIZE. The code has been
1011 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1012 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1013 of around 200,000 forward references. While doing this, I also speeded up
1014 the filling in of repeated forward references.
1015
1016 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1017 incorrectly expecting the subject to contain another "a" after the start.
1018
1019 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1020 in the match, the SKIP should be ignored. This was not happening; instead
1021 the SKIP was being treated as NOMATCH. For patterns such as
1022 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1023 tested.
1024
1025 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1026 now much more compatible with Perl, in particular in cases where the result
1027 is a non-match for a non-anchored pattern. For example, if
1028 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1029 "m", where previously it did not return a name. A side effect of this
1030 change is that for partial matches, the last encountered mark name is
1031 returned, as for non matches. A number of tests that were previously not
1032 Perl-compatible have been moved into the Perl-compatible test files. The
1033 refactoring has had the pleasing side effect of removing one argument from
1034 the match() function, thus reducing its stack requirements.
1035
1036 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1037 subsequent uses of /S (without +) incorrectly behaved like /S+.
1038
1039 21. Retrieve executable code size support for the JIT compiler and fixing
1040 some warnings.
1041
1042 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1043 not work when the shorter character appeared right at the end of the
1044 subject string.
1045
1046 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1047 systems.
1048
1049 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1050 output it when the /M option is used in pcretest.
1051
1052 25. The CheckMan script was not being included in the distribution. Also, added
1053 an explicit "perl" to run Perl scripts from the PrepareRelease script
1054 because this is reportedly needed in Windows.
1055
1056 26. If study data was being save in a file and studying had not found a set of
1057 "starts with" bytes for the pattern, the data written to the file (though
1058 never used) was taken from uninitialized memory and so caused valgrind to
1059 complain.
1060
1061 27. Updated RunTest.bat as provided by Sheri Pierce.
1062
1063 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1064
1065 29. Computation of memory usage for the table of capturing group names was
1066 giving an unnecessarily large value.
1067
1068
1069 Version 8.20 21-Oct-2011
1070 ------------------------
1071
1072 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1073 a POSIX class. After further experiments with Perl, which convinced me that
1074 Perl has bugs and confusions, a closing square bracket is no longer allowed
1075 in a POSIX name. This bug also affected patterns with classes that started
1076 with full stops.
1077
1078 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1079 captured substring, but while checking the failing first alternative,
1080 substring 1 is temporarily captured. If the output vector supplied to
1081 pcre_exec() was not big enough for this capture, the yield of the function
1082 was still zero ("insufficient space for captured substrings"). This cannot
1083 be totally fixed without adding another stack variable, which seems a lot
1084 of expense for a edge case. However, I have improved the situation in cases
1085 such as /(a)(b)x|abc/ matched against "abc", where the return code
1086 indicates that fewer than the maximum number of slots in the ovector have
1087 been set.
1088
1089 3. Related to (2) above: when there are more back references in a pattern than
1090 slots in the output vector, pcre_exec() uses temporary memory during
1091 matching, and copies in the captures as far as possible afterwards. It was
1092 using the entire output vector, but this conflicts with the specification
1093 that only 2/3 is used for passing back captured substrings. Now it uses
1094 only the first 2/3, for compatibility. This is, of course, another edge
1095 case.
1096
1097 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1098 main code base, and can be used by building with --enable-jit. When this is
1099 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1100 runtime --no-jit option is given.
1101
1102 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1103 ovector, the return from the function was zero, implying that there were
1104 other matches that did not fit. The correct "exactly full" value is now
1105 returned.
1106
1107 6. If a subpattern that was called recursively or as a subroutine contained
1108 (*PRUNE) or any other control that caused it to give a non-standard return,
1109 invalid errors such as "Error -26 (nested recursion at the same subject
1110 position)" or even infinite loops could occur.
1111
1112 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1113 computing the minimum length on reaching *ACCEPT, and so ended up with the
1114 wrong value of 1 rather than 0. Further investigation indicates that
1115 computing a minimum subject length in the presence of *ACCEPT is difficult
1116 (think back references, subroutine calls), and so I have changed the code
1117 so that no minimum is registered for a pattern that contains *ACCEPT.
1118
1119 8. If (*THEN) was present in the first (true) branch of a conditional group,
1120 it was not handled as intended. [But see 16 below.]
1121
1122 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1123 Sheri Pierce.
1124
1125 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1126 the first byte in a match must be "a".
1127
1128 11. Change 17 for 8.13 increased the recursion depth for patterns like
1129 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1130 pattern contains any instances of (*THEN). If it does not, the old
1131 optimizations are restored. It would be nice to do this on a per-group
1132 basis, but at the moment that is not feasible.
1133
1134 12. In some environments, the output of pcretest -C is CRLF terminated. This
1135 broke RunTest's code that checks for the link size. A single white space
1136 character after the value is now allowed for.
1137
1138 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1139 For "fr", it uses the Windows-specific input and output files.
1140
1141 14. If (*THEN) appeared in a group that was called recursively or as a
1142 subroutine, it did not work as intended. [But see next item.]
1143
1144 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1145 pattern fragments (but not containing any | characters). If A and B are
1146 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1147 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1148 D. In other words, Perl considers parentheses that do not contain any |
1149 characters to be part of a surrounding alternative, whereas PCRE was
1150 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1151 differently. PCRE now behaves in the same way as Perl, except in the case
1152 of subroutine/recursion calls such as (?1) which have in any case always
1153 been different (but PCRE had them first :-).
1154
1155 16. Related to 15 above: Perl does not treat the | in a conditional group as
1156 creating alternatives. Such a group is treated in the same way as an
1157 ordinary group without any | characters when processing (*THEN). PCRE has
1158 been changed to match Perl's behaviour.
1159
1160 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1161 RunGrepTest script failed.
1162
1163 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1164 inevitable for groups that contain captures, but it can lead to a lot of
1165 stack use in large patterns. The old behaviour has been restored for atomic
1166 groups that do not contain any capturing parentheses.
1167
1168 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1169 suppress the check for a minimum subject length at run time. (If it was
1170 given to pcre_exec() or pcre_dfa_exec() it did work.)
1171
1172 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1173 fail to work when decoding hex characters in data strings in EBCDIC
1174 environments.
1175
1176 21. It appears that in at least one Mac OS environment, the isxdigit() function
1177 is implemented as a macro that evaluates to its argument more than once,
1178 contravening the C 90 Standard (I haven't checked a later standard). There
1179 was an instance in pcretest which caused it to go wrong when processing
1180 \x{...} escapes in subject strings. The has been rewritten to avoid using
1181 things like p++ in the argument of isxdigit().
1182
1183
1184 Version 8.13 16-Aug-2011
1185 ------------------------
1186
1187 1. The Unicode data tables have been updated to Unicode 6.0.0.
1188
1189 2. Two minor typos in pcre_internal.h have been fixed.
1190
1191 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1192 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1193 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1194
1195 4. There were a number of related bugs in the code for matching backrefences
1196 caselessly in UTF-8 mode when codes for the characters concerned were
1197 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1198 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1199 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1200 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1201 2-byte code at the end of the subject (it thought there wasn't enough data
1202 left).
1203
1204 5. Comprehensive information about what went wrong is now returned by
1205 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1206 as the output vector has at least 2 elements. The offset of the start of
1207 the failing character and a reason code are placed in the vector.
1208
1209 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1210 now returned is for the first byte of the failing character, instead of the
1211 last byte inspected. This is an incompatible change, but I hope it is small
1212 enough not to be a problem. It makes the returned offset consistent with
1213 pcre_exec() and pcre_dfa_exec().
1214
1215 7. pcretest now gives a text phrase as well as the error number when
1216 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1217 failure, the offset and reason code are output.
1218
1219 8. When \R was used with a maximizing quantifier it failed to skip backwards
1220 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1221 back over a single character (\n). This seems wrong (because it treated the
1222 two characters as a single entity when going forwards), conflicts with the
1223 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1224 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1225 has been changed.
1226
1227 9. Some internal refactoring has changed the processing so that the handling
1228 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1229 time (the PCRE_DOTALL option was changed this way some time ago: version
1230 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1231 which was always a bit of a fudge. It also means that there is one less
1232 argument for the match() function, which reduces its stack requirements
1233 slightly. This change also fixes an incompatibility with Perl: the pattern
1234 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1235
1236 10. More internal refactoring has drastically reduced the number of recursive
1237 calls to match() for possessively repeated groups such as (abc)++ when
1238 using pcre_exec().
1239
1240 11. While implementing 10, a number of bugs in the handling of groups were
1241 discovered and fixed:
1242
1243 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1244 (a|)*(?1) gave a compile-time internal error.
1245 ((a|)+)+ did not notice that the outer group could match an empty string.
1246 (^a|^)+ was not marked as anchored.
1247 (.*a|.*)+ was not marked as matching at start or after a newline.
1248
1249 12. Yet more internal refactoring has removed another argument from the match()
1250 function. Special calls to this function are now indicated by setting a
1251 value in a variable in the "match data" data block.
1252
1253 13. Be more explicit in pcre_study() instead of relying on "default" for
1254 opcodes that mean there is no starting character; this means that when new
1255 ones are added and accidentally left out of pcre_study(), testing should
1256 pick them up.
1257
1258 14. The -s option of pcretest has been documented for ages as being an old
1259 synonym of -m (show memory usage). I have changed it to mean "force study
1260 for every regex", that is, assume /S for every regex. This is similar to -i
1261 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1262 using it. It makes it easier to run collections of tests with and without
1263 study enabled, and thereby test pcre_study() more easily. All the standard
1264 tests are now run with and without -s (but some patterns can be marked as
1265 "never study" - see 20 below).
1266
1267 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1268 restoration of the capturing data to the outer values was not happening
1269 correctly.
1270
1271 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1272 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1273 pattern had matched an empty string, and so incorrectly returned a no
1274 match.
1275
1276 17. There was optimizing code for the last branch of non-capturing parentheses,
1277 and also for the obeyed branch of a conditional subexpression, which used
1278 tail recursion to cut down on stack usage. Unfortunately, now that there is
1279 the possibility of (*THEN) occurring in these branches, tail recursion is
1280 no longer possible because the return has to be checked for (*THEN). These
1281 two optimizations have therefore been removed. [But see 8.20/11 above.]
1282
1283 18. If a pattern containing \R was studied, it was assumed that \R always
1284 matched two bytes, thus causing the minimum subject length to be
1285 incorrectly computed because \R can also match just one byte.
1286
1287 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1288 was incorrectly computed.
1289
1290 20. If /S is present twice on a test pattern in pcretest input, it now
1291 *disables* studying, thereby overriding the use of -s on the command line
1292 (see 14 above). This is necessary for one or two tests to keep the output
1293 identical in both cases.
1294
1295 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1296 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1297
1298 22. When an atomic group that contained a capturing parenthesis was
1299 successfully matched, but the branch in which it appeared failed, the
1300 capturing was not being forgotten if a higher numbered group was later
1301 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1302 group 1 to "a", when in fact it should be unset. This applied to multi-
1303 branched capturing and non-capturing groups, repeated or not, and also to
1304 positive assertions (capturing in negative assertions does not happen
1305 in PCRE) and also to nested atomic groups.
1306
1307 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1308 subject after a captured substring, to make it easier to tell which of a
1309 number of identical substrings has been captured.
1310
1311 24. The way atomic groups are processed by pcre_exec() has been changed so that
1312 if they are repeated, backtracking one repetition now resets captured
1313 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1314 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1315 "aaa". Previously, it would have been "a". As part of this code
1316 refactoring, the way recursive calls are handled has also been changed.
1317
1318 25. If an assertion condition captured any substrings, they were not passed
1319 back unless some other capturing happened later. For example, if
1320 (?(?=(a))a) was matched against "a", no capturing was returned.
1321
1322 26. When studying a pattern that contained subroutine calls or assertions,
1323 the code for finding the minimum length of a possible match was handling
1324 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1325 group 1 called group 2 while simultaneously a separate group 2 called group
1326 1). A stack overflow occurred in this case. I have fixed this by limiting
1327 the recursion depth to 10.
1328
1329 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1330 Fortmann. This supports explicit test numbers on the command line, and has
1331 argument validation and error reporting.
1332
1333 28. An instance of \X with an unlimited repeat could fail if at any point the
1334 first character it looked at was a mark character.
1335
1336 29. Some minor code refactoring concerning Unicode properties and scripts
1337 should reduce the stack requirement of match() slightly.
1338
1339 30. Added the '=' option to pcretest to check the setting of unused capturing
1340 slots at the end of the pattern, which are documented as being -1, but are
1341 not included in the return count.
1342
1343 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1344 compiled something random. Now it gives a compile-time error (as does
1345 Perl).
1346
1347 32. A *MARK encountered during the processing of a positive assertion is now
1348 recorded and passed back (compatible with Perl).
1349
1350 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1351 had alternative anchored branches, the search for a second match in a line
1352 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1353 matched the line "0102" twice. The same bug affected patterns that started
1354 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1355 twice.
1356
1357 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1358 does, and because of capturing effects, quantifying parenthesized
1359 assertions may at times be useful. Quantifiers are now allowed for
1360 parenthesized assertions.
1361
1362 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1363
1364 36. \g was being checked for fancy things in a character class, when it should
1365 just be a literal "g".
1366
1367 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1368 appearance of a nested POSIX class supersedes an apparent external class.
1369 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1370 unescaped square brackets may also appear as part of class names. For
1371 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1372 more like Perl. (But see 8.20/1 above.)
1373
1374 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1375 was because it thought it was \N{name}, which is not supported).
1376
1377 39. Add minix to OS list not supporting the -S option in pcretest.
1378
1379 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1380 cannot analyze patterns in sufficient detail to catch mutual recursions
1381 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1382 subgroup is called recursively as a subpattern for a second time at the
1383 same position in the subject string. In previous releases this might have
1384 been caught by the recursion limit, or it might have run out of stack.
1385
1386 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1387 happen only once. PCRE was, however incorrectly giving a compile time error
1388 "recursive call could loop indefinitely" because it cannot analyze the
1389 pattern in sufficient detail. The compile time test no longer happens when
1390 PCRE is compiling a conditional subpattern, but actual runaway loops are
1391 now caught at runtime (see 40 above).
1392
1393 42. It seems that Perl allows any characters other than a closing parenthesis
1394 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1395 has been changed to be the same.
1396
1397 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1398 as not to get warnings when autogen.sh is called. Also changed
1399 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1400
1401 44. To help people who use pcregrep to scan files containing exceedingly long
1402 lines, the following changes have been made:
1403
1404 (a) The default value of the buffer size parameter has been increased from
1405 8K to 20K. (The actual buffer used is three times this size.)
1406
1407 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1408 PCRE is built.
1409
1410 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1411 to be set at run time.
1412
1413 (d) Numerical values in pcregrep options can be followed by K or M, for
1414 example --buffer-size=50K.
1415
1416 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1417 given and the return code is set to 2.
1418
1419 45. Add a pointer to the latest mark to the callout data block.
1420
1421 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1422 partial match of an empty string instead of no match. This was specific to
1423 the use of ".".
1424
1425 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1426 complete match instead of a partial match. This bug was dependent on both
1427 the PCRE_UTF8 and PCRE_DOTALL options being set.
1428
1429 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1430 starting byte set, because \b was not being ignored.
1431
1432
1433 Version 8.12 15-Jan-2011
1434 ------------------------
1435
1436 1. Fixed some typos in the markup of the man pages, and wrote a script that
1437 checks for such things as part of the documentation building process.
1438
1439 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1440 --match-limit and --recursion-limit options (added for 8.11). In
1441 particular, this made one of the standard tests fail. (The integer value
1442 went into the wrong half of a long int.)
1443
1444 3. If the --colour option was given to pcregrep with -v (invert match), it
1445 did strange things, either producing crazy output, or crashing. It should,
1446 of course, ignore a request for colour when reporting lines that do not
1447 match.
1448
1449 4. Another pcregrep bug caused similar problems if --colour was specified with
1450 -M (multiline) and the pattern match finished with a line ending.
1451
1452 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1453 matched in multiline mode, the following line was shown as part of the
1454 match. This seems wrong, so I have changed it.
1455
1456 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1457 the check for further matches in the same line (so they could be coloured)
1458 to overrun the end of the current line. If another match was found, it was
1459 incorrectly shown (and then shown again when found in the next line).
1460
1461 7. If pcregrep was compiled under Windows, there was a reference to the
1462 function pcregrep_exit() before it was defined. I am assuming this was
1463 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1464 reported by a user. I've moved the definition above the reference.
1465
1466
1467 Version 8.11 10-Dec-2010
1468 ------------------------
1469
1470 1. (*THEN) was not working properly if there were untried alternatives prior
1471 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1472 backtracked to try for "b" instead of moving to the next alternative branch
1473 at the same level (in this case, to look for "c"). The Perl documentation
1474 is clear that when (*THEN) is backtracked onto, it goes to the "next
1475 alternative in the innermost enclosing group".
1476
1477 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1478 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1479 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1480 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1481 (*THEN).
1482
1483 3. If \s appeared in a character class, it removed the VT character from
1484 the class, even if it had been included by some previous item, for example
1485 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1486 of \s, but is part of the POSIX "space" class.)
1487
1488 4. A partial match never returns an empty string (because you can always
1489 match an empty string at the end of the subject); however the checking for
1490 an empty string was starting at the "start of match" point. This has been
1491 changed to the "earliest inspected character" point, because the returned
1492 data for a partial match starts at this character. This means that, for
1493 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1494 (previously it gave "no match").
1495
1496 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1497 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1498 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1499 has an implication that the given string is incomplete (because a partial
1500 match is preferred over a full match). For this reason, these items now
1501 give a partial match in this situation. [Aside: previously, the one case
1502 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1503 match rather than a full match, which was wrong by the old rules, but is
1504 now correct.]
1505
1506 6. There was a bug in the handling of #-introduced comments, recognized when
1507 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1508 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1509 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1510 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1511 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1512 places in pcre_compile().
1513
1514 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1515 comments when looking ahead for named forward references to subpatterns,
1516 the only newline sequence it recognized was NL. It now handles newlines
1517 according to the set newline convention.
1518
1519 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1520 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1521 cater for a lack of strerror(). These oversights have been fixed.
1522
1523 9. Added --match-limit and --recursion-limit to pcregrep.
1524
1525 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1526
1527 11. When the -o option was used, pcregrep was setting a return code of 1, even
1528 when matches were found, and --line-buffered was not being honoured.
1529
1530 12. Added an optional parentheses number to the -o and --only-matching options
1531 of pcregrep.
1532
1533 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1534 can match an empty string. The code to do it in pcretest and pcredemo
1535 needed fixing:
1536
1537 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1538 only one byte after an empty string match just before CRLF (this case
1539 just got forgotten; "any" and "anycrlf" were OK).
1540
1541 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1542 mode when an empty string match preceded an ASCII character followed by
1543 a non-ASCII character. (The code for advancing by one character rather
1544 than one byte was nonsense.)
1545
1546 (c) The pcredemo.c sample program did not have any code at all to handle
1547 the cases when CRLF is a valid newline sequence.
1548
1549 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1550 as a starting offset was within the subject string. There is now a new
1551 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1552 negative or greater than the length of the string. In order to test this,
1553 pcretest is extended to allow the setting of negative starting offsets.
1554
1555 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1556 starting offset points to the beginning of a UTF-8 character was
1557 unnecessarily clumsy. I tidied it up.
1558
1559 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1560 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1561
1562 17. Nobody had reported that the --include_dir option, which was added in
1563 release 7.7 should have been called --include-dir (hyphen, not underscore)
1564 for compatibility with GNU grep. I have changed it to --include-dir, but
1565 left --include_dir as an undocumented synonym, and the same for
1566 --exclude-dir, though that is not available in GNU grep, at least as of
1567 release 2.5.4.
1568
1569 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1570 characters from a string of bytes) have been redefined so as not to use
1571 loops, in order to improve performance in some environments. At the same
1572 time, I abstracted some of the common code into auxiliary macros to save
1573 repetition (this should not affect the compiled code).
1574
1575 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1576 compile-time error is now given if \c is not followed by an ASCII
1577 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1578 different, and any byte value is allowed.)
1579
1580 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1581 START_OPTIMIZE option, which is now allowed at compile time - but just
1582 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1583 to pcregrep and other applications that have no direct access to PCRE
1584 options. The new /Y option in pcretest sets this option when calling
1585 pcre_compile().
1586
1587 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1588 back references. Groups containing recursive back references were forced to
1589 be atomic by that change, but in the case of named groups, the amount of
1590 memory required was incorrectly computed, leading to "Failed: internal
1591 error: code overflow". This has been fixed.
1592
1593 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1594 pcretest.c, to avoid build problems in some Borland environments.
1595
1596
1597 Version 8.10 25-Jun-2010
1598 ------------------------
1599
1600 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1601 THEN.
1602
1603 2. (*ACCEPT) was not working when inside an atomic group.
1604
1605 3. Inside a character class, \B is treated as a literal by default, but
1606 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1607 causes the error). The code is unchanged, but I tidied the documentation.
1608
1609 4. Inside a character class, PCRE always treated \R and \X as literals,
1610 whereas Perl faults them if its -w option is set. I have changed PCRE so
1611 that it faults them when PCRE_EXTRA is set.
1612
1613 5. Added support for \N, which always matches any character other than
1614 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1615
1616 6. When compiling pcregrep with newer versions of gcc which may have
1617 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1618 declared with attribute warn_unused_result" were given. Just casting the
1619 result to (void) does not stop the warnings; a more elaborate fudge is
1620 needed. I've used a macro to implement this.
1621
1622 7. Minor change to pcretest.c to avoid a compiler warning.
1623
1624 8. Added four artifical Unicode properties to help with an option to make
1625 \s etc use properties (see next item). The new properties are: Xan
1626 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1627
1628 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1629 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1630 this option. Modified pcretest to add /W to test this facility. Added
1631 REG_UCP to make it available via the POSIX interface.
1632
1633 10. Added --line-buffered to pcregrep.
1634
1635 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1636 studied, and the match started with a letter with a code point greater than
1637 127 whose first byte was different to the first byte of the other case of
1638 the letter, the other case of this starting letter was not recognized
1639 (#976).
1640
1641 12. If a pattern that was studied started with a repeated Unicode property
1642 test, for example, \p{Nd}+, there was the theoretical possibility of
1643 setting up an incorrect bitmap of starting bytes, but fortunately it could
1644 not have actually happened in practice until change 8 above was made (it
1645 added property types that matched character-matching opcodes).
1646
1647 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1648 possible starting bytes for non-anchored patterns.
1649
1650 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1651 \R, and also a number of cases that involve Unicode properties, both
1652 explicit and implicit when PCRE_UCP is set.
1653
1654 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1655 input, it could crash or give wrong results if characters with values
1656 greater than 0xc0 were present in the subject string. (Detail: it assumed
1657 UTF-8 input when processing these items.)
1658
1659 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1660 size_t is 64-bit (#991).
1661
1662 17. Added a check for running out of memory when PCRE is compiled with
1663 --disable-stack-for-recursion (#990).
1664
1665 18. If the last data line in a file for pcretest does not have a newline on
1666 the end, a newline was missing in the output.
1667
1668 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1669 less than 128) in its various bitmaps. However, there is a facility for
1670 generating tables according to the current locale when PCRE is compiled. It
1671 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1672 characters, are recognized by isspace() and therefore were getting set in
1673 these tables, and indeed these tables seem to approximate to ISO 8859. This
1674 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1675 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1676 which of course cannot start UTF-8 characters. I have changed the code so
1677 that only real ASCII characters (less than 128) and the correct starting
1678 bytes for UTF-8 encodings are set for characters greater than 127 when in
1679 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1680 altogether.)
1681
1682 20. Added the /T option to pcretest so as to be able to run tests with non-
1683 standard character tables, thus making it possible to include the tests
1684 used for 19 above in the standard set of tests.
1685
1686 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1687 reference to a subpattern the other side of a comment that contains an
1688 opening parenthesis caused either an internal compiling error, or a
1689 reference to the wrong subpattern.
1690
1691
1692 Version 8.02 19-Mar-2010
1693 ------------------------
1694
1695 1. The Unicode data tables have been updated to Unicode 5.2.0.
1696
1697 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1698 configured.
1699
1700 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1701 original author of that file, following a query about its status.
1702
1703 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1704 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1705
1706 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1707 quantifier applied to a forward-referencing subroutine call, could compile
1708 incorrect code or give the error "internal error: previously-checked
1709 referenced subpattern not found".
1710
1711 6. Both MS Visual Studio and Symbian OS have problems with initializing
1712 variables to point to external functions. For these systems, therefore,
1713 pcre_malloc etc. are now initialized to local functions that call the
1714 relevant global functions.
1715
1716 7. There were two entries missing in the vectors called coptable and poptable
1717 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1718 I've fixed the data, and added a kludgy way of testing at compile time that
1719 the lengths are correct (equal to the number of opcodes).
1720
1721 8. Following on from 7, I added a similar kludge to check the length of the
1722 eint vector in pcreposix.c.
1723
1724 9. Error texts for pcre_compile() are held as one long string to avoid too
1725 much relocation at load time. To find a text, the string is searched,
1726 counting zeros. There was no check for running off the end of the string,
1727 which could happen if a new error number was added without updating the
1728 string.
1729
1730 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1731
1732 11. \K was not working if it appeared in an atomic group or in a group that
1733 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1734 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1735 the assertion is positive, but not if it is negative.
1736
1737 12. Change 11 fortuitously reduced the size of the stack frame used in the
1738 "match()" function of pcre_exec.c by one pointer. Forthcoming
1739 implementation of support for (*MARK) will need an extra pointer on the
1740 stack; I have reserved it now, so that the stack frame size does not
1741 decrease.
1742
1743 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1744 item in branch that calls a recursion is a subroutine call - as in the
1745 second branch in the above example - was incorrectly given the compile-
1746 time error "recursive call could loop indefinitely" because pcre_compile()
1747 was not correctly checking the subroutine for matching a non-empty string.
1748
1749 14. The checks for overrunning compiling workspace could trigger after an
1750 overrun had occurred. This is a "should never occur" error, but it can be
1751 triggered by pathological patterns such as hundreds of nested parentheses.
1752 The checks now trigger 100 bytes before the end of the workspace.
1753
1754 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1755
1756
1757 Version 8.01 19-Jan-2010
1758 ------------------------
1759
1760 1. If a pattern contained a conditional subpattern with only one branch (in
1761 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1762 computed the wrong minimum data length (which is of course zero for such
1763 subpatterns). This could cause incorrect "no match" results.
1764
1765 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1766 the pattern is reset in the first branch, pcre_compile() failed with
1767 "internal error: code overflow at offset...". This happened only when
1768 the reset was to the original external option setting. (An optimization
1769 abstracts leading options settings into an external setting, which was the
1770 cause of this.)
1771
1772 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1773 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1774 assertion pattern did not match (meaning that the assertion was true), it
1775 was incorrectly treated as false if the SKIP had been reached during the
1776 matching. This also applied to assertions used as conditions.
1777
1778 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1779 assertion subpattern, including such a pattern used as a condition,
1780 unpredictable results occurred, instead of the error return
1781 PCRE_ERROR_DFA_UITEM.
1782
1783 5. The C++ GlobalReplace function was not working like Perl for the special
1784 situation when an empty string is matched. It now does the fancy magic
1785 stuff that is necessary.
1786
1787 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1788 removed. (These were left over from very, very early versions of PCRE.)
1789
1790 7. Some cosmetic changes to the code to make life easier when compiling it
1791 as part of something else:
1792
1793 (a) Change DEBUG to PCRE_DEBUG.
1794
1795 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1796 called "current" as "current_branch", to prevent a collision with the
1797 Linux macro when compiled as a kernel module.
1798
1799 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1800 prevent a collision with the Linux macro when compiled as a kernel
1801 module.
1802
1803 8. In pcre_compile() there are some checks for integer overflows that used to
1804 cast potentially large values to (double). This has been changed to that
1805 when building, a check for int64_t is made, and if it is found, it is used
1806 instead, thus avoiding the use of floating point arithmetic. (There is no
1807 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1808 double.
1809
1810 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1811 2005 (difference between two addresses compared to an unsigned value).
1812
1813 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1814 custom one, because of the following reported problem in Windows:
1815
1816 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1817 under Win32.
1818 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1819 therefore missing the function definition.
1820 - The compiler thus generates a "C" signature for the test function.
1821 - The linker fails to find the "C" function.
1822 - PCRE fails to configure if asked to do so against libbz2.
1823
1824 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1825 messages were output:
1826
1827 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1828 rerunning libtoolize, to keep the correct libtool macros in-tree.
1829 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1830
1831 I have done both of these things.
1832
1833 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1834 most of the time, it *can* run out if it is given a pattern that contains a
1835 runaway infinite recursion. I updated the discussion in the pcrestack man
1836 page.
1837
1838 13. Now that we have gone to the x.xx style of version numbers, the minor
1839 version may start with zero. Using 08 or 09 is a bad idea because users
1840 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1841 interpreted as invalid octal numbers. I've updated the previous comment in
1842 configure.ac, and also added a check that gives an error if 08 or 09 are
1843 used.
1844
1845 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1846 causing partial matching to fail when the end of the subject matched \W
1847 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1848
1849 15. There were some discrepancies between the declarations in pcre_internal.h
1850 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1851 their definitions. The declarations used "const uschar *" and the
1852 definitions used USPTR. Even though USPTR is normally defined as "const
1853 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1854 reported that: "This difference in casting confuses some C++ compilers, for
1855 example, SunCC recognizes above declarations as different functions and
1856 generates broken code for hbpcre." I have changed the declarations to use
1857 USPTR.
1858
1859 16. GNU libtool is named differently on some systems. The autogen.sh script now
1860 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1861 (FreeBSD).
1862
1863 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1864 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1865 comment: "Figure out how to create a longlong from a string: strtoll and
1866 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1867 instance, but it only takes 2 args instead of 3!"
1868
1869 18. A subtle bug concerned with back references has been fixed by a change of
1870 specification, with a corresponding code fix. A pattern such as
1871 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1872 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1873 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1874 same bug. Such groups have to be quantified to be useful, or contained
1875 inside another quantified group. (If there's no repetition, the reference
1876 can never match.) The problem arises because, having left the group and
1877 moved on to the rest of the pattern, a later failure that backtracks into
1878 the group uses the captured value from the final iteration of the group
1879 rather than the correct earlier one. I have fixed this in PCRE by forcing
1880 any group that contains a reference to itself to be an atomic group; that
1881 is, there cannot be any backtracking into it once it has completed. This is
1882 similar to recursive and subroutine calls.
1883
1884
1885 Version 8.00 19-Oct-09
1886 ----------------------
1887
1888 1. The table for translating pcre_compile() error codes into POSIX error codes
1889 was out-of-date, and there was no check on the pcre_compile() error code
1890 being within the table. This could lead to an OK return being given in
1891 error.
1892
1893 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1894 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1895 in a Windows environment.
1896
1897 3. The pcregrep --count option prints the count for each file even when it is
1898 zero, as does GNU grep. However, pcregrep was also printing all files when
1899 --files-with-matches was added. Now, when both options are given, it prints
1900 counts only for those files that have at least one match. (GNU grep just
1901 prints the file name in this circumstance, but including the count seems
1902 more useful - otherwise, why use --count?) Also ensured that the
1903 combination -clh just lists non-zero counts, with no names.
1904
1905 4. The long form of the pcregrep -F option was incorrectly implemented as
1906 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1907 but it seems right to fix it, and I didn't think it was worth preserving
1908 the old behaviour.
1909
1910 5. The command line items --regex=pattern and --regexp=pattern were not
1911 recognized by pcregrep, which required --regex pattern or --regexp pattern
1912 (with a space rather than an '='). The man page documented the '=' forms,
1913 which are compatible with GNU grep; these now work.
1914
1915 6. No libpcreposix.pc file was created for pkg-config; there was just
1916 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1917
1918 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1919 when UCP support is not needed, by modifying the Python script that
1920 generates it from Unicode data files. This should not matter if the module
1921 is correctly used as a library, but I received one complaint about 50K of
1922 unwanted data. My guess is that the person linked everything into his
1923 program rather than using a library. Anyway, it does no harm.
1924
1925 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1926 was a minimum greater than 1 for a wide character in a possessive
1927 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1928 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1929 character. Chaos in the form of incorrect output or a compiling loop could
1930 result.
1931
1932 9. The restrictions on what a pattern can contain when partial matching is
1933 requested for pcre_exec() have been removed. All patterns can now be
1934 partially matched by this function. In addition, if there are at least two
1935 slots in the offset vector, the offset of the earliest inspected character
1936 for the match and the offset of the end of the subject are set in them when
1937 PCRE_ERROR_PARTIAL is returned.
1938
1939 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1940 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1941 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1942 and may be more useful for multi-segment matching.
1943
1944 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1945 used to be given if ever the end of the subject was reached; now it is
1946 given only if matching could not proceed because another character was
1947 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1948 string "Z", which now yields "no match" instead of "partial match". In the
1949 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1950 final character ended with (*FAIL).
1951
1952 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1953 if the pattern had a "must contain" character that was already found in the
1954 earlier partial match, unless partial matching was again requested. For
1955 example, with the pattern /dog.(body)?/, the "must contain" character is
1956 "g". If the first part-match was for the string "dog", restarting with
1957 "sbody" failed. This bug has been fixed.
1958
1959 13. The string returned by pcre_dfa_exec() after a partial match has been
1960 changed so that it starts at the first inspected character rather than the
1961 first character of the match. This makes a difference only if the pattern
1962 starts with a lookbehind assertion or \b or \B (\K is not supported by
1963 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1964 matching functions compatible, and I think it's the right thing to do.
1965
1966 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1967 so that the demonstration program is easily available in environments where
1968 PCRE has not been installed from source.
1969
1970 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1971 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1972 library.
1973
1974 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1975 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1976 is not the first non-POSIX option to be added. Clearly some people find
1977 these options useful.
1978
1979 17. If a caller to the POSIX matching function regexec() passes a non-zero
1980 value for nmatch with a NULL value for pmatch, the value of
1981 nmatch is forced to zero.
1982
1983 18. RunGrepTest did not have a test for the availability of the -u option of
1984 the diff command, as RunTest does. It now checks in the same way as
1985 RunTest, and also checks for the -b option.
1986
1987 19. If an odd number of negated classes containing just a single character
1988 interposed, within parentheses, between a forward reference to a named
1989 subpattern and the definition of the subpattern, compilation crashed with
1990 an internal error, complaining that it could not find the referenced
1991 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1992 [The bug was that it was starting one character too far in when skipping
1993 over the character class, thus treating the ] as data rather than
1994 terminating the class. This meant it could skip too much.]
1995
1996 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1997 /g option in pcretest when the pattern contains \K, which makes it possible
1998 to have an empty string match not at the start, even when the pattern is
1999 anchored. Updated pcretest and pcredemo to use this option.
2000
2001 21. If the maximum number of capturing subpatterns in a recursion was greater
2002 than the maximum at the outer level, the higher number was returned, but
2003 with unset values at the outer level. The correct (outer level) value is
2004 now given.
2005
2006 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2007 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2008 make it do so. The string so far is captured, making this feature
2009 compatible with Perl.
2010
2011 23. The tests have been re-organized, adding tests 11 and 12, to make it
2012 possible to check the Perl 5.10 features against Perl 5.10.
2013
2014 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2015 pattern matches a fixed length string. PCRE did not allow this; now it
2016 does. Neither allows recursion.
2017
2018 25. I finally figured out how to implement a request to provide the minimum
2019 length of subject string that was needed in order to match a given pattern.
2020 (It was back references and recursion that I had previously got hung up
2021 on.) This code has now been added to pcre_study(); it finds a lower bound
2022 to the length of subject needed. It is not necessarily the greatest lower
2023 bound, but using it to avoid searching strings that are too short does give
2024 some useful speed-ups. The value is available to calling programs via
2025 pcre_fullinfo().
2026
2027 26. While implementing 25, I discovered to my embarrassment that pcretest had
2028 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2029 study optimizations had never been tested with that matching function.
2030 Oops. What is worse, even when it was passed study data, there was a bug in
2031 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2032 were also very few tests of studied patterns with pcre_dfa_exec().
2033
2034 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2035 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2036 on the other side of the coin, they are no longer allowed to have different
2037 names, because these cannot be distinguished in PCRE, and this has caused
2038 confusion. (This is a difference from Perl.)
2039
2040 28. When duplicate subpattern names are present (necessarily with different
2041 numbers, as required by 27 above), and a test is made by name in a
2042 conditional pattern, either for a subpattern having been matched, or for
2043 recursion in such a pattern, all the associated numbered subpatterns are
2044 tested, and the overall condition is true if the condition is true for any
2045 one of them. This is the way Perl works, and is also more like the way
2046 testing by number works.
2047
2048
2049 Version 7.9 11-Apr-09
2050 ---------------------
2051
2052 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2053 (pcretest), all targets were linked against these libraries. This included
2054 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2055 libraries. This caused unwanted dependencies to be created. This problem
2056 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2057 pcretest is linked with readline.
2058
2059 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2060 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2061 moved outside it again, because FALSE and TRUE are already defined in AIX,
2062 but BOOL is not.
2063
2064 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2065 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2066
2067 4. The pcregrep documentation said spaces were inserted as well as colons (or
2068 hyphens) following file names and line numbers when outputting matching
2069 lines. This is not true; no spaces are inserted. I have also clarified the
2070 wording for the --colour (or --color) option.
2071
2072 5. In pcregrep, when --colour was used with -o, the list of matching strings
2073 was not coloured; this is different to GNU grep, so I have changed it to be
2074 the same.
2075
2076 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2077 each matching line was coloured. Now it goes on to look for further matches
2078 of any of the test patterns, which is the same behaviour as GNU grep.
2079
2080 7. A pattern that could match an empty string could cause pcregrep to loop; it
2081 doesn't make sense to accept an empty string match in pcregrep, so I have
2082 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2083 seems to be how GNU grep behaves. [But see later change 40 for release
2084 8.33.]
2085
2086 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2087 start or after a newline", because the conditional assertion was not being
2088 correctly handled. The rule now is that both the assertion and what follows
2089 in the first alternative must satisfy the test.
2090
2091 9. If auto-callout was enabled in a pattern with a conditional group whose
2092 condition was an assertion, PCRE could crash during matching, both with
2093 pcre_exec() and pcre_dfa_exec().
2094
2095 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2096 used for matching.
2097
2098 11. Unicode property support in character classes was not working for
2099 characters (bytes) greater than 127 when not in UTF-8 mode.
2100
2101 12. Added the -M command line option to pcretest.
2102
2103 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2104
2105 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2106
2107 16. Added comments and documentation about mis-use of no_arg in the C++
2108 wrapper.
2109
2110 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2111 from Martin Jerabek that uses macro names for all relevant character and
2112 string constants.
2113
2114 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2115 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2116 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2117 these, but not everybody uses configure.
2118
2119 19. A conditional group that had only one branch was not being correctly
2120 recognized as an item that could match an empty string. This meant that an
2121 enclosing group might also not be so recognized, causing infinite looping
2122 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2123 with the subject "ab", where knowledge that the repeated group can match
2124 nothing is needed in order to break the loop.
2125
2126 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2127 exec(), but without supplying a callout function, matching went wrong.
2128
2129 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2130 leak if the size of the offset vector was greater than 30. When the vector
2131 is smaller, the saved offsets during recursion go onto a local stack
2132 vector, but for larger vectors malloc() is used. It was failing to free
2133 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2134 error, in fact).
2135
2136 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2137 heapframe that is used only when UTF-8 support is enabled. This caused no
2138 problem, but was untidy.
2139
2140 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2141 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2142 included within another project.
2143
2144 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2145 slightly modified by me:
2146
2147 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2148 not building pcregrep.
2149
2150 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2151 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2152
2153 25. Forward references, both numeric and by name, in patterns that made use of
2154 duplicate group numbers, could behave incorrectly or give incorrect errors,
2155 because when scanning forward to find the reference group, PCRE was not
2156 taking into account the duplicate group numbers. A pattern such as
2157 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2158
2159 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2160 the feature of a custom pointer more persuasive (as requested by a user).
2161
2162 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2163 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2164 pre-defined.
2165
2166 28. Added support for (*UTF8) at the start of a pattern.
2167
2168 29. Arrange for flags added by the "release type" setting in CMake to be shown
2169 in the configuration summary.
2170
2171
2172 Version 7.8 05-Sep-08
2173 ---------------------
2174
2175 1. Replaced UCP searching code with optimized version as implemented for Ad
2176 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2177 stage table and inline lookup instead of a function, giving speed ups of 2
2178 to 5 times on some simple patterns that I tested. Permission was given to
2179 distribute the MultiStage2.py script that generates the tables (it's not in
2180 the tarball, but is in the Subversion repository).
2181
2182 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2183 scripts.
2184
2185 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2186 a group with a zero qualifier. The result of the study could be incorrect,
2187 or the function might crash, depending on the pattern.
2188
2189 4. Caseless matching was not working for non-ASCII characters in back
2190 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2191 It now works when Unicode Property Support is available.
2192
2193 5. In pcretest, an escape such as \x{de} in the data was always generating
2194 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2195 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2196 truncation.
2197
2198 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2199
2200 7. Added two (int) casts to pcregrep when printing the difference of two
2201 pointers, in case they are 64-bit values.
2202
2203 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2204 test 2 if it fails.
2205
2206 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2207 and a #define of that name to empty if it is not externally set. This is to
2208 allow users of MSVC to set it if necessary.
2209
2210 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2211 the convenience functions in the pcre_get.c source file.
2212
2213 11. An option change at the start of a pattern that had top-level alternatives
2214 could cause overwriting and/or a crash. This command provoked a crash in
2215 some environments:
2216
2217 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2218
2219 This potential security problem was recorded as CVE-2008-2371.
2220
2221 12. For a pattern where the match had to start at the beginning or immediately
2222 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2223 pcre_dfa_exec() could read past the end of the passed subject if there was
2224 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2225 pcretest so that it places the subject at the end of its malloc-ed buffer.
2226
2227 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2228 exec() might read past the end of the data buffer in UTF-8 mode.
2229
2230 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2231 the data contained the byte 0x85 as part of a UTF-8 character within its
2232 first line. This applied both to normal and DFA matching.
2233
2234 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2235 /^[^d]*?$/8 failed to match "abc".
2236
2237 16. Added a missing copyright notice to pcrecpp_internal.h.
2238
2239 17. Make it more clear in the documentation that values returned from
2240 pcre_exec() in ovector are byte offsets, not character counts.
2241
2242 18. Tidied a few places to stop certain compilers from issuing warnings.
2243
2244 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2245 supplied by Stefan Weber. I made a further small update for 7.8 because
2246 there is a change of source arrangements: the pcre_searchfuncs.c module is
2247 replaced by pcre_ucd.c.
2248
2249
2250 Version 7.7 07-May-08
2251 ---------------------
2252
2253 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2254 a string to a long long, pretend we don't even have a long long." This is
2255 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2256
2257 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2258 pre-7.6 versions, which defined a global no_arg variable instead of putting
2259 it in the RE class. (See also #8 below.)
2260
2261 3. Remove a line of dead code, identified by coverity and reported by Nuno
2262 Lopes.
2263
2264 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2265
2266 (1) The include/exclude patterns were being applied to the whole pathnames
2267 of files, instead of just to the final components.
2268
2269 (2) If there was more than one level of directory, the subdirectories were
2270 skipped unless they satisfied the include/exclude conditions. This is
2271 inconsistent with GNU grep (and could even be seen as contrary to the
2272 pcregrep specification - which I improved to make it absolutely clear).
2273 The action now is always to scan all levels of directory, and just
2274 apply the include/exclude patterns to regular files.
2275
2276 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2277 --exclude_dir in the tests to avoid scanning .svn directories.
2278
2279 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2280 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2281 doesn't support NULs in patterns.
2282
2283 7. Added some missing "const"s to declarations of static tables in
2284 pcre_compile.c and pcre_dfa_exec.c.
2285
2286 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2287 caused by fix #2 above. (Subsequently also a second patch to fix the
2288 first patch. And a third patch - this was a messy problem.)
2289
2290 9. Applied Craig's patch to remove the use of push_back().
2291
2292 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2293 matching function regexec().
2294
2295 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2296 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2297 references. PCRE supports relative numbers with this syntax (I don't think
2298 Oniguruma does).
2299
2300 12. Previously, a group with a zero repeat such as (...){0} was completely
2301 omitted from the compiled regex. However, this means that if the group
2302 was called as a subroutine from elsewhere in the pattern, things went wrong
2303 (an internal error was given). Such groups are now left in the compiled
2304 pattern, with a new opcode that causes them to be skipped at execution
2305 time.
2306
2307 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2308 to the way PCRE behaves:
2309
2310 (a) A lone ] character is dis-allowed (Perl treats it as data).
2311
2312 (b) A back reference to an unmatched subpattern matches an empty string
2313 (Perl fails the current match path).
2314
2315 (c) A data ] in a character class must be notated as \] because if the
2316 first data character in a class is ], it defines an empty class. (In
2317 Perl it is not possible to have an empty class.) The empty class []
2318 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2319 The negative empty class [^] matches any one character, independently
2320 of the DOTALL setting.
2321
2322 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2323 non-existent subpattern following a character class starting with ']' and
2324 containing () gave an internal compiling error instead of "reference to
2325 non-existent subpattern". Fortunately, when the pattern did exist, the
2326 compiled code was correct. (When scanning forwards to check for the
2327 existence of the subpattern, it was treating the data ']' as terminating
2328 the class, so got the count wrong. When actually compiling, the reference
2329 was subsequently set up correctly.)
2330
2331 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2332 it was being rejected as not supported by pcre_dfa_exec(), even though
2333 other assertions are supported. I have made pcre_dfa_exec() support
2334 (*FAIL).
2335
2336 16. The implementation of 13c above involved the invention of a new opcode,
2337 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2338 cannot be changed at match time, I realized I could make a small
2339 improvement to matching performance by compiling OP_ALLANY instead of
2340 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2341 on the OP_ANY path.
2342
2343 17. Compiling pcretest on Windows with readline support failed without the
2344 following two fixes: (1) Make the unistd.h include conditional on
2345 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2346
2347 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2348 ncurses library to be included for pcretest when ReadLine support is
2349 requested, but also to allow for it to be overridden. This patch came from
2350 Daniel Bergström.
2351
2352 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2353 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2354 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2355 spotting this.
2356
2357
2358 Version 7.6 28-Jan-08
2359 ---------------------
2360
2361 1. A character class containing a very large number of characters with
2362 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2363 overflow.
2364
2365 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2366 HAVE_LONG_LONG is not defined.
2367
2368 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2369 bring them up to date and include new features. This patch includes:
2370
2371 - Fixed PH's badly added libz and libbz2 support.
2372 - Fixed a problem with static linking.
2373 - Added pcredemo. [But later removed - see 7 below.]
2374 - Fixed dftables problem and added an option.
2375 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2376 HAVE_LONG_LONG.
2377 - Added readline support for pcretest.
2378 - Added an listing of the option settings after cmake has run.
2379
2380 4. A user submitted a patch to Makefile that makes it easy to create
2381 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2382 Makefile.am that cause it to include this special target, without
2383 affecting anything else. Note that the same mingw target plus all
2384 the other distribution libraries and programs are now supported
2385 when configuring with CMake (see 6 below) instead of with
2386 Configure/Make.
2387
2388 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2389 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2390 exported in the Windows port". It has not yet been confirmed that the patch
2391 solves the problem, but it does no harm.
2392
2393 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2394 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2395 with CMake, and also correct the comment about stack recursion.
2396
2397 7. Remove the automatic building of pcredemo from the ./configure system and
2398 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2399 of a program that users should build themselves after PCRE is installed, so
2400 building it automatically is not really right. What is more, it gave
2401 trouble in some build environments.
2402
2403 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2404
2405
2406 Version 7.5 10-Jan-08
2407 ---------------------
2408
2409 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2410 values in parens when parsing an RE using the C++ wrapper."
2411
2412 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2413 Characters greater than 255 were excluded from the class instead of being
2414 included.
2415
2416 3. The same bug as (2) above applied to negated POSIX classes such as
2417 [:^space:].
2418
2419 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2420 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2421 I have changed it.
2422
2423 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2424 first named subpattern) and a construct such as (?&a) would reference the
2425 first named subpattern whose name started with "a" (in other words, the
2426 length check was missing). Both these problems are fixed. "Subpattern name
2427 expected" is now given for (?&) (a zero-length name), and this patch also
2428 makes it give the same error for \k'' (previously it complained that that
2429 was a reference to a non-existent subpattern).
2430
2431 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2432 this is right because (?- can be followed by option settings as well as by
2433 digits. I have, however, made the messages clearer.
2434
2435 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2436 than the number used in the conditional) now cause a compile-time error.
2437 This is actually not compatible with Perl, which accepts such patterns, but
2438 treats the conditional as always being FALSE (as PCRE used to), but it
2439 seems to me that giving a diagnostic is better.
2440
2441 8. Change "alphameric" to the more common word "alphanumeric" in comments
2442 and messages.
2443
2444 9. Fix two occurrences of "backslash" in comments that should have been
2445 "backspace".
2446
2447 10. Remove two redundant lines of code that can never be obeyed (their function
2448 was moved elsewhere).
2449
2450 11. The program that makes PCRE's Unicode character property table had a bug
2451 which caused it to generate incorrect table entries for sequences of
2452 characters that have the same character type, but are in different scripts.
2453 It amalgamated them into a single range, with the script of the first of
2454 them. In other words, some characters were in the wrong script. There were
2455 thirteen such cases, affecting characters in the following ranges:
2456
2457 U+002b0 - U+002c1
2458 U+0060c - U+0060d
2459 U+0061e - U+00612
2460 U+0064b - U+0065e
2461 U+0074d - U+0076d
2462 U+01800 - U+01805
2463 U+01d00 - U+01d77
2464 U+01d9b - U+01dbf
2465 U+0200b - U+0200f
2466 U+030fc - U+030fe
2467 U+03260 - U+0327f
2468 U+0fb46 - U+0fbb1
2469 U+10450 - U+1049d
2470
2471 12. The -o option (show only the matching part of a line) for pcregrep was not
2472 compatible with GNU grep in that, if there was more than one match in a
2473 line, it showed only the first of them. It now behaves in the same way as
2474 GNU grep.
2475
2476 13. If the -o and -v options were combined for pcregrep, it printed a blank
2477 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2478 does the same. The return code can be used to tell if there were any
2479 non-matching lines.
2480
2481 14. Added --file-offsets and --line-offsets to pcregrep.
2482
2483 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2484 infinitely looping recursion. The bug was that positive lookaheads were not
2485 being skipped when checking for a possible empty match (negative lookaheads
2486 and both kinds of lookbehind were skipped).
2487
2488 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2489 inclusion of <windows.h> to before rather than after the definition of
2490 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2491
2492 17. Specifying a possessive quantifier with a specific limit for a Unicode
2493 character property caused pcre_compile() to compile bad code, which led at
2494 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2495 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2496 caused the error; without that there was no problem.
2497
2498 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2499
2500 19. Added --enable-pcretest-libreadline.
2501
2502 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2503 RE::GlobalReplace(). As a result, the number of replacements returned was
2504 double what it should be. I removed one of the increments, but Craig sent a
2505 later patch that removed the other one (the right fix) and added unit tests
2506 that check the return values (which was not done before).
2507
2508 21. Several CMake things:
2509
2510 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2511 the names libpcre and libpcreposix, not just pcre and pcreposix.
2512
2513 (2) The above change means that pcretest and pcregrep are now correctly
2514 linked with the newly-built libraries, not previously installed ones.
2515
2516 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2517
2518 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2519 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2520 UTF-8 newline character). The key issue is that the pattern starts .*;
2521 this means that the match must be either at the beginning, or after a
2522 newline. The bug was in the code for advancing after a failed match and
2523 checking that the new position followed a newline. It was not taking
2524 account of UTF-8 characters correctly.
2525
2526 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2527 character classes. PCRE was not treating the sequence [:...:] as a
2528 character class unless the ... were all letters. Perl, however, seems to
2529 allow any characters between [: and :], though of course it rejects as
2530 unknown any "names" that contain non-letters, because all the known class
2531 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2532 for example, whereas PCRE did not - it did not recognize a POSIX character
2533 class. This seemed a bit dangerous, so the code has been changed to be
2534 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2535 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2536 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2537 Perl does, and where it didn't before.
2538
2539 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2540 Windows environments %n is disabled by default.
2541
2542
2543 Version 7.4 21-Sep-07
2544 ---------------------
2545
2546 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2547 means that a class such as [\s] counted as "explicit reference to CR or
2548 LF". That isn't really right - the whole point of the change was to try to
2549 help when there was an actual mention of one of the two characters. So now
2550 the change happens only if \r or \n (or a literal CR or LF) character is
2551 encountered.
2552
2553 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2554 of both had grown to the point where there were only 3 bits left.
2555 Fortunately, there was spare space in the data structure, and so I have
2556 moved the internal flags into a new 16-bit field to free up more option
2557 bits.
2558
2559 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2560 but did not set the internal JCHANGED flag - either of these is enough to
2561 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2562 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2563 start sets both bits.
2564
2565 4. Added options (at build time, compile time, exec time) to change \R from
2566 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2567
2568 5. doc/pcresyntax.html was missing from the distribution.
2569
2570 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2571 compatibility, even though it is no longer used.
2572
2573 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2574 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2575 windows.h file is present (where different names are used). [This was
2576 reversed later after testing - see 16 below.]
2577
2578 8. Changed all #include <config.h> to #include "config.h". There were also
2579 some further <pcre.h> cases that I changed to "pcre.h".
2580
2581 9. When pcregrep was used with the --colour option, it missed the line ending
2582 sequence off the lines that it output.
2583
2584 10. It was pointed out to me that arrays of string pointers cause lots of
2585 relocations when a shared library is dynamically loaded. A technique of
2586 using a single long string with a table of offsets can drastically reduce
2587 these. I have refactored PCRE in four places to do this. The result is
2588 dramatic:
2589
2590 Originally: 290
2591 After changing UCP table: 187
2592 After changing error message table: 43
2593 After changing table of "verbs" 36
2594 After changing table of Posix names 22
2595
2596 Thanks to the folks working on Gregex for glib for this insight.
2597
2598 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2599 unicode-properties was also set.
2600
2601 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2602
2603 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2604 checked only for CRLF.
2605
2606 14. Added casts to pcretest.c to avoid compiler warnings.
2607
2608 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2609
2610 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2611 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2612 entirely. This removes changes made in 7 above.
2613
2614 17. The CMake files have been updated, and there is now more information about
2615 building with CMake in the NON-UNIX-USE document.
2616
2617
2618 Version 7.3 28-Aug-07
2619 ---------------------
2620
2621 1. In the rejigging of the build system that eventually resulted in 7.1, the
2622 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2623 brackets there is not right, since it causes compilers to look for an
2624 installed pcre.h, not the version that is in the source that is being
2625 compiled (which of course may be different). I have changed it back to:
2626
2627 #include "pcre.h"
2628
2629 I have a vague recollection that the change was concerned with compiling in
2630 different directories, but in the new build system, that is taken care of
2631 by the VPATH setting the Makefile.
2632
2633 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2634 when the subject happened to end in the byte 0x85 (e.g. if the last
2635 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2636 characters but of course it shouldn't be taken as a newline when it is part
2637 of another character. The bug was that, for an unlimited repeat of . in
2638 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2639 characters when looking for a newline.
2640
2641 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2642
2643 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2644 in debug output.
2645
2646 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2647 long printing in the pcrecpp unittest when running under MinGW.
2648
2649 6. ESC_K was left out of the EBCDIC table.
2650
2651 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2652 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2653 limit also applies to "virtual nesting" when a pattern is recursive, and in
2654 this case 1000 isn't so big. I have been able to remove this limit at the
2655 expense of backing off one optimization in certain circumstances. Normally,
2656 when pcre_exec() would call its internal match() function recursively and
2657 immediately return the result unconditionally, it uses a "tail recursion"
2658 feature to save stack. However, when a subpattern that can match an empty
2659 string has an unlimited repetition quantifier, it no longer makes this
2660 optimization. That gives it a stack frame in which to save the data for
2661 checking that an empty string has been matched. Previously this was taken
2662 from the 1000-entry workspace that had been reserved. So now there is no
2663 explicit limit, but more stack is used.
2664
2665 8. Applied Daniel's patches to solve problems with the import/export magic
2666 syntax that is required for Windows, and which was going wrong for the
2667 pcreposix and pcrecpp parts of the library. These were overlooked when this
2668 problem was solved for the main library.
2669
2670 9. There were some crude static tests to avoid integer overflow when computing
2671 the size of patterns that contain repeated groups with explicit upper
2672 limits. As the maximum quantifier is 65535, the maximum group length was
2673 set at 30,000 so that the product of these two numbers did not overflow a
2674 32-bit integer. However, it turns out that people want to use groups that
2675 are longer than 30,000 bytes (though not repeat them that many times).
2676 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2677 made it possible to implement the integer overflow checks in a much more
2678 dynamic way, which I have now done. The artificial limitation on group
2679 length has been removed - we now have only the limit on the total length of
2680 the compiled pattern, which depends on the LINK_SIZE setting.
2681
2682 10. Fixed a bug in the documentation for get/copy named substring when
2683 duplicate names are permitted. If none of the named substrings are set, the
2684 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2685 empty string.
2686
2687 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2688 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2689 because the ] is interpreted as the first data character and the
2690 terminating ] is not found. PCRE has been made compatible with Perl in this
2691 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2692 cause memory overwriting.
2693
2694 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2695 string has been matched (to stop an infinite loop). It was not recognizing
2696 a conditional subpattern that could match an empty string if that
2697 subpattern was within another subpattern. For example, it looped when
2698 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2699 condition was not nested. This bug has been fixed.
2700
2701 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2702 past the start of the subject in the presence of bytes with the top bit
2703 set, for example "\x8aBCD".
2704
2705 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2706 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2707
2708 14. Optimized (?!) to (*FAIL).
2709
2710 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2711 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2712 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2713 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2714 does: it's just the validity check that is more restrictive.
2715
2716 16. Inserted checks for integer overflows during escape sequence (backslash)
2717 processing, and also fixed erroneous offset values for syntax errors during
2718 backslash processing.
2719
2720 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2721 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2722
2723 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2724 caused an overrun.
2725
2726 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2727 something other than just ASCII characters) inside a group that had an
2728 unlimited repeat caused a loop at compile time (while checking to see
2729 whether the group could match an empty string).
2730
2731 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2732 [\P{Any}] did so. (Error in the code for printing property names.)
2733
2734 21. An orphan \E inside a character class could cause a crash.
2735
2736 22. A repeated capturing bracket such as (A)? could cause a wild memory
2737 reference during compilation.
2738
2739 23. There are several functions in pcre_compile() that scan along a compiled
2740 expression for various reasons (e.g. to see if it's fixed length for look
2741 behind). There were bugs in these functions when a repeated \p or \P was
2742 present in the pattern. These operators have additional parameters compared
2743 with \d, etc, and these were not being taken into account when moving along
2744 the compiled data. Specifically:
2745
2746 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2747 length.
2748
2749 (b) An item such as \pL+ within a repeated group could cause crashes or
2750 loops.
2751
2752 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2753 "reference to non-existent subpattern" error.
2754
2755 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2756
2757 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2758 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2759
2760 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2761
2762 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2763 character were causing crashes (broken optimization).
2764
2765 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2766 \p or \P) caused a compile-time loop.
2767
2768 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2769 break. For example, the unstudied pattern [\r\n]A does not match the string
2770 "\r\nA" because change 7.0/46 below moves the current point on by two
2771 characters after failing to match at the start. However, the pattern \nA
2772 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2773 the same is true. There doesn't seem any very clean way out of this, but
2774 what I have chosen to do makes the common cases work: PCRE now takes note
2775 of whether there can be an explicit match for \r or \n anywhere in the
2776 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2777 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2778 pattern has explicit CR or LF references.
2779
2780 29. Added (*CR) etc for changing newline setting at start of pattern.
2781
2782
2783 Version 7.2 19-Jun-07
2784 ---------------------
2785
2786 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2787 which is apparently normally available under Windows.
2788
2789 2. Re-jig the pcregrep tests with different newline settings in an attempt
2790 to make them independent of the local environment's newline setting.
2791
2792 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2793
2794 4. Some of the "internals" tests were previously cut out when the link size
2795 was not 2, because the output contained actual offsets. The recent new
2796 "Z" feature of pcretest means that these can be cut out, making the tests
2797 usable with all link sizes.
2798
2799 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2800 stack recursion. This gives a massive performance boost under BSD, but just
2801 a small improvement under Linux. However, it saves one field in the frame
2802 in all cases.
2803
2804 6. Added more features from the forthcoming Perl 5.10:
2805
2806 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2807 recursion call. It refers to the nth most recently opened parentheses.
2808
2809 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2810 to be opened parentheses.
2811
2812 (c) Conditions that refer to capturing parentheses can be specified
2813 relatively, for example, (?(-2)... or (?(+3)...
2814
2815 (d) \K resets the start of the current match so that everything before
2816 is not part of it.
2817
2818 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2819
2820 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2821 reference syntax.
2822
2823 (g) (?| introduces a group in which the numbering of parentheses in each
2824 alternative starts with the same number.
2825
2826 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2827
2828 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2829 PCRE_INFO_JCHANGED.
2830
2831 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2832 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2833 for detecting groups that can match an empty string.
2834
2835 9. A pattern with a very large number of alternatives (more than several
2836 hundred) was running out of internal workspace during the pre-compile
2837 phase, where pcre_compile() figures out how much memory will be needed. A
2838 bit of new cunning has reduced the workspace needed for groups with
2839 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2840 workspace instead of running out of the 4096 that are available.
2841
2842 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2843
2844 11. Applied patch from Google to remove an optimization that didn't quite work.
2845 The report of the bug said:
2846
2847 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2848 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2849 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2850
2851 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2852 it matched the wrong number of bytes.
2853
2854
2855 Version 7.1 24-Apr-07
2856 ---------------------
2857
2858 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2859 that is more "standard", making use of automake and other Autotools. There
2860 is some re-arrangement of the files and adjustment of comments consequent
2861 on this.
2862
2863 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2864 for recursive directory scanning broke on some systems because the files
2865 are not scanned in any specific order and on different systems the order
2866 was different. A call to "sort" has been inserted into RunGrepTest for the
2867 approprate test as a short-term fix. In the longer term there may be an
2868 alternative.
2869
2870 3. I had an email from Eric Raymond about problems translating some of PCRE's
2871 man pages to HTML (despite the fact that I distribute HTML pages, some
2872 people do their own conversions for various reasons). The problems
2873 concerned the use of low-level troff macros .br and .in. I have therefore
2874 removed all such uses from the man pages (some were redundant, some could
2875 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2876 HTML has been updated to handle .nf/.fi and to complain if it encounters
2877 .br or .in.
2878
2879 4. Updated comments in configure.ac that get placed in config.h.in and also
2880 arranged for config.h to be included in the distribution, with the name
2881 config.h.generic, for the benefit of those who have to compile without
2882 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2883
2884 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2885 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2886 makevp.bat for the current PCRE, using the additional files
2887 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2888
2889 6. A Windows user reported a minor discrepancy with test 2, which turned out
2890 to be caused by a trailing space on an input line that had got lost in his
2891 copy. The trailing space was an accident, so I've just removed it.
2892
2893 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2894 that is needed.
2895
2896 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2897 as "const" (a) because they are and (b) because it helps the PHP
2898 maintainers who have recently made a script to detect big data structures
2899 in the php code that should be moved to the .rodata section. I remembered
2900 to update Builducptable as well, so it won't revert if ucptable.h is ever
2901 re-created.
2902
2903 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2904 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2905 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2906 support is not required. This saves 1.5-2K of code, which is important in
2907 some applications.
2908
2909 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2910 so as not to refer to the tables, even though these functions will never be
2911 called when UTF-8 support is disabled. Otherwise there are problems with a
2912 shared library.
2913
2914 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2915
2916 (a) It was defining its arguments as char * instead of void *.
2917
2918 (b) It was assuming that all moves were upwards in memory; this was true
2919 a long time ago when I wrote it, but is no longer the case.
2920
2921 The emulated memove() is provided for those environments that have neither
2922 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2923 is clearly not the case, as these two bugs were recently reported.
2924
2925 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2926 and Detrail to create the HTML documentation, the .txt form of the man
2927 pages, and it removes trailing spaces from listed files. It also creates
2928 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2929 case, it wraps all the #defines with #ifndefs. This script should be run
2930 before "make dist".
2931
2932 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2933 with Unicode property support.
2934
2935 (a) For a maximizing quantifier, if the two different cases of the
2936 character were of different lengths in their UTF-8 codings (there are
2937 some cases like this - I found 11), and the matching function had to
2938 back up over a mixture of the two cases, it incorrectly assumed they
2939 were both the same length.
2940
2941 (b) When PCRE was configured to use the heap rather than the stack for
2942 recursion during matching, it was not correctly preserving the data for
2943 the other case of a UTF-8 character when checking ahead for a match
2944 while processing a minimizing repeat. If the check also involved
2945 matching a wide character, but failed, corruption could cause an
2946 erroneous result when trying to check for a repeat of the original
2947 character.
2948
2949 13. Some tidying changes to the testing mechanism:
2950
2951 (a) The RunTest script now detects the internal link size and whether there
2952 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2953 values substituted by "configure". (The RunGrepTest script already did
2954 this for UTF-8.) The configure.ac script no longer substitutes the
2955 relevant variables.
2956
2957 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2958 with length and offset values. This means that the output is different
2959 for different internal link sizes. Test 2 is skipped for link sizes
2960 other than 2 because of this, bypassing the problem. Unfortunately,
2961 there was also a test in test 3 (the locale tests) that used /B and
2962 failed for link sizes other than 2. Rather than cut the whole test out,
2963 I have added a new /Z option to pcretest that replaces the length and
2964 offset values with spaces. This is now used to make test 3 independent
2965 of link size. (Test 2 will be tidied up later.)
2966
2967 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2968 segmentation fault instead of returning the appropriate error message.
2969
2970 15. In multiline mode when the newline sequence was set to "any", the pattern
2971 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2972 This doesn't seem right; it now treats the CRLF combination as the line
2973 ending, and so does not match in that case. It's only a pattern such as ^$
2974 that would hit this one: something like ^ABC$ would have failed after \r
2975 and then tried again after \r\n.
2976
2977 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2978 in an attempt to make files that differ only in their line terminators
2979 compare equal. This works on Linux.
2980
2981 17. Under certain error circumstances pcregrep might try to free random memory
2982 as it exited. This is now fixed, thanks to valgrind.
2983
2984 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2985 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2986 was because its rules for how to advance for /g after matching an empty
2987 string at the end of a line did not allow for this case. They now check for
2988 it specially.
2989
2990 20. pcretest is supposed to handle patterns and data of any length, by
2991 extending its buffers when necessary. It was getting this wrong when the
2992 buffer for a data line had to be extended.
2993
2994 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2995 CRLF as a newline sequence.
2996
2997 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2998 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2999 I have nevertheless tidied it up.
3000
3001 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3002
3003 24. Added a man page for pcre-config.
3004
3005
3006 Version 7.0 19-Dec-06
3007 ---------------------
3008
3009 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3010 moving to gcc 4.1.1.
3011
3012 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3013 sys/time.h, which is documented as needed for this function. It doesn't
3014 seem to matter on Linux, but it showed up on some releases of OS X.
3015
3016 3. It seems that there are systems where bytes whose values are greater than
3017 127 match isprint() in the "C" locale. The "C" locale should be the
3018 default when a C program starts up. In most systems, only ASCII printing
3019 characters match isprint(). This difference caused the output from pcretest
3020 to vary, making some of the tests fail. I have changed pcretest so that:
3021
3022 (a) When it is outputting text in the compiled version of a pattern, bytes
3023 other than 32-126 are always shown as hex escapes.
3024
3025 (b) When it is outputting text that is a matched part of a subject string,
3026 it does the same, unless a different locale has been set for the match
3027 (using the /L modifier). In this case, it uses isprint() to decide.
3028
3029 4. Fixed a major bug that caused incorrect computation of the amount of memory
3030 required for a compiled pattern when options that changed within the
3031 pattern affected the logic of the preliminary scan that determines the
3032 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3033 that the computed length was too small. The symptoms of this bug were
3034 either the PCRE error "internal error: code overflow" from pcre_compile(),
3035 or a glibc crash with a message such as "pcretest: free(): invalid next
3036 size (fast)". Examples of patterns that provoked this bug (shown in
3037 pcretest format) are:
3038
3039 /(?-x: )/x
3040 /(?x)(?-x: \s*#\s*)/
3041 /((?i)[\x{c0}])/8
3042 /(?i:[\x{c0}])/8
3043
3044 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3045 is now done differently.
3046
3047 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3048 wrapper classes; (b) implement a new function in the C++ scanner that is
3049 more efficient than the old way of doing things because it avoids levels of
3050 recursion in the regex matching; (c) add a paragraph to the documentation
3051 for the FullMatch() function.
3052
3053 6. The escape sequence \n was being treated as whatever was defined as
3054 "newline". Not only was this contrary to the documentation, which states
3055 that \n is character 10 (hex 0A), but it also went horribly wrong when
3056 "newline" was defined as CRLF. This has been fixed.
3057
3058 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3059 was being set to -1 for the "end of line" case (supposedly a value that no
3060 character can have). Though this value is never used (the check for end of
3061 line is "zero bytes in current character"), it caused compiler complaints.
3062 I've changed it to 0xffffffff.
3063
3064 8. In pcre_version.c, the version string was being built by a sequence of
3065 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3066 string (as it is for production releases) called a macro with an empty
3067 argument. The C standard says the result of this is undefined. The gcc
3068 compiler treats it as an empty string (which was what was wanted) but it is
3069 reported that Visual C gives an error. The source has been hacked around to
3070 avoid this problem.
3071
3072 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3073 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3074 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3075 of them did).
3076
3077 10. Originally, pcretest opened its input and output without "b"; then I was
3078 told that "b" was needed in some environments, so it was added for release
3079 5.0 to both the input and output. (It makes no difference on Unix-like
3080 systems.) Later I was told that it is wrong for the input on Windows. I've
3081 now abstracted the modes into two macros, to make it easier to fiddle with
3082 them, and removed "b" from the input mode under Windows.
3083
3084 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3085
3086 12. Added -help and --help to pcretest as an official way of being reminded
3087 of the options.
3088
3089 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3090 and pcrecpp.cc because they annoy compilers at high warning levels.
3091
3092 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3093
3094 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3095 scripts are not C programs :-) and which was not noticed because it works
3096 on Linux.
3097
3098 16. pcretest is supposed to handle any length of pattern and data line (as one
3099 line or as a continued sequence of lines) by extending its input buffer if
3100 necessary. This feature was broken for very long pattern lines, leading to
3101 a string of junk being passed to pcre_compile() if the pattern was longer
3102 than about 50K.
3103
3104 17. I have done a major re-factoring of the way pcre_compile() computes the
3105 amount of memory needed for a compiled pattern. Previously, there was code
3106 that made a preliminary scan of the pattern in order to do this. That was
3107 OK when PCRE was new, but as the facilities have expanded, it has become
3108 harder and harder to keep it in step with the real compile phase, and there
3109 have been a number of bugs (see for example, 4 above). I have now found a
3110 cunning way of running the real compile function in a "fake" mode that
3111 enables it to compute how much memory it would need, while actually only
3112 ever using a few hundred bytes of working memory and without too many
3113 tests of the mode. This should make future maintenance and development
3114 easier. A side effect of this work is that the limit of 200 on the nesting
3115 depth of parentheses has been removed (though this was never a serious
3116 limitation, I suspect). However, there is a downside: pcre_compile() now
3117 runs more slowly than before (30% or more, depending on the pattern). I
3118 hope this isn't a big issue. There is no effect on runtime performance.
3119
3120 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3121 newline (only possible for the last line of a file) and it was a
3122 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3123
3124 19. Added additional timing features to pcretest. (1) The -tm option now times
3125 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3126 separate command line item, by a number that specifies the number of
3127 repeats to use when timing. The default is 50000; this gives better
3128 precision, but takes uncomfortably long for very large patterns.
3129
3130 20. Extended pcre_study() to be more clever in cases where a branch of a
3131 subpattern has no definite first character. For example, (a*|b*)[cd] would
3132 previously give no result from pcre_study(). Now it recognizes that the
3133 first character must be a, b, c, or d.
3134
3135 21. There was an incorrect error "recursive call could loop indefinitely" if
3136 a subpattern (or the entire pattern) that was being tested for matching an
3137 empty string contained only one non-empty item after a nested subpattern.
3138 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3139 incorrectly, because the \d was being skipped in the check.
3140
3141 22. The pcretest program now has a new pattern option /B and a command line
3142 option -b, which is equivalent to adding /B to every pattern. This causes
3143 it to show the compiled bytecode, without the additional information that
3144 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3145 is the same as /B/I).
3146
3147 23. A new optimization is now able automatically to treat some sequences such
3148 as a*b as a*+b. More specifically, if something simple (such as a character
3149 or a simple class like \d) has an unlimited quantifier, and is followed by
3150 something that cannot possibly match the quantified thing, the quantifier
3151 is automatically "possessified".
3152
3153 24. A recursive reference to a subpattern whose number was greater than 39
3154 went wrong under certain circumstances in UTF-8 mode. This bug could also
3155 have affected the operation of pcre_study().
3156
3157 25. Realized that a little bit of performance could be had by replacing
3158 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3159
3160 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3161
3162 27. Possessive quantifiers such as a++ were previously implemented by turning
3163 them into atomic groups such as ($>a+). Now they have their own opcodes,
3164 which improves performance. This includes the automatically created ones
3165 from 23 above.
3166
3167 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3168 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3169 the first matched character to be a colon. This applied both to named and
3170 numbered groups.
3171
3172 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3173
3174 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3175 building PCRE on Windows easier, so I have included it in the distribution.
3176
3177 31. There is now a check in pcretest against a ridiculously large number being
3178 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3179 loop, the loop is abandoned.
3180
3181 32. Forward references to subpatterns in conditions such as (?(2)...) where
3182 subpattern 2 is defined later cause pcre_compile() to search forwards in
3183 the pattern for the relevant set of parentheses. This search went wrong
3184 when there were unescaped parentheses in a character class, parentheses
3185 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3186
3187 33. "Subroutine" calls and backreferences were previously restricted to
3188 referencing subpatterns earlier in the regex. This restriction has now
3189 been removed.
3190
3191 34. Added a number of extra features that are going to be in Perl 5.10. On the
3192 whole, these are just syntactic alternatives for features that PCRE had
3193 previously implemented using the Python syntax or my own invention. The
3194 other formats are all retained for compatibility.
3195
3196 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3197 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3198 also .NET compatible.
3199
3200 (b) A recursion or subroutine call to a named group can now be defined as
3201 (?&name) as well as (?P>name).
3202
3203 (c) A backreference to a named group can now be defined as \k<name> or
3204 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3205 5.10, are also .NET compatible.
3206
3207 (d) A conditional reference to a named group can now use the syntax
3208 (?(<name>) or (?('name') as well as (?(name).
3209
3210 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3211 groups (named and numbered) that are never evaluated inline, but can be
3212 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3213 is always false. There may be only one alternative in such a group.
3214
3215 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3216 as the simple (?(R). The condition is true only if the most recent
3217 recursion is that of the given number or name. It does not search out
3218 through the entire recursion stack.
3219
3220 (g) The escape \gN or \g{N} has been added, where N is a positive or
3221 negative number, specifying an absolute or relative reference.
3222
3223 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3224 some "unreachable code" warnings.
3225
3226 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3227 things, this adds five new scripts.
3228
3229 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3230 There were also incompatibilities regarding the handling of \Q..\E inside
3231 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3232 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3233
3234 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3235 matches an empty string, and forcibly breaks the loop. There were bugs in
3236 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3237 against aaaa the result was just "a" rather than "aaaa", for example. Two
3238 separate and independent bugs (that affected different cases) have been
3239 fixed.
3240
3241 39. Refactored the code to abolish the use of different opcodes for small
3242 capturing bracket numbers. This is a tidy that I avoided doing when I
3243 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3244 The new approach is not only tidier, it makes it possible to reduce the
3245 memory needed to fix the previous bug (38).
3246
3247 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3248 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3249 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3250 mode.
3251
3252 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3253 report.
3254
3255 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3256 copy construction and assignment in the C++ wrapper.
3257
3258 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3259 couple of bugs that could have given wrong results in the "--newline=crlf"
3260 case.
3261
3262 44. Added a number of casts and did some reorganization of signed/unsigned int
3263 variables following suggestions from Dair Grant. Also renamed the variable
3264 "this" as "item" because it is a C++ keyword.
3265
3266 45. Arranged for dftables to add
3267
3268 #include "pcre_internal.h"
3269
3270 to pcre_chartables.c because without it, gcc 4.x may remove the array
3271 definition from the final binary if PCRE is built into a static library and
3272 dead code stripping is activated.
3273
3274 46. For an unanchored pattern, if a match attempt fails at the start of a
3275 newline sequence, and the newline setting is CRLF or ANY, and the next two
3276 characters are CRLF, advance by two characters instead of one.
3277
3278
3279 Version 6.7 04-Jul-06
3280 ---------------------
3281
3282 1. In order to handle tests when input lines are enormously long, pcretest has
3283 been re-factored so that it automatically extends its buffers when
3284 necessary. The code is crude, but this _is_ just a test program. The
3285 default size has been increased from 32K to 50K.
3286
3287 2. The code in pcre_study() was using the value of the re argument before
3288 testing it for NULL. (Of course, in any sensible call of the function, it
3289 won't be NULL.)
3290
3291 3. The memmove() emulation function in pcre_internal.h, which is used on
3292 systems that lack both memmove() and bcopy() - that is, hardly ever -
3293 was missing a "static" storage class specifier.
3294
3295 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3296 containing an extended class (one that cannot be represented by a bitmap
3297 because it contains high-valued characters or Unicode property items, e.g.
3298 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3299 pattern, but PCRE should not loop if you do not (it no longer does).
3300 [Detail: two cases were found: (a) a repeated subpattern containing an
3301 extended class; (b) a recursive reference to a subpattern that followed a
3302 previous extended class. It wasn't skipping over the extended class
3303 correctly when UTF-8 mode was not set.]
3304
3305 5. A negated single-character class was not being recognized as fixed-length
3306 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3307 compile error "lookbehind assertion is not fixed length".
3308
3309 6. The RunPerlTest auxiliary script was showing an unexpected difference
3310 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3311 write a Perl script that can interpret lines of an input file either as
3312 byte characters or as UTF-8, which is what "perltest" was being required to
3313 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3314 can't do is switch easily at run time between having the "use utf8;" pragma
3315 or not. In the end, I fudged it by using the RunPerlTest script to insert
3316 "use utf8;" explicitly for the UTF-8 tests.
3317
3318 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3319 the end of the subject string, contrary to the documentation and to what
3320 Perl does. This was true of both matching functions. Now it matches only at
3321 the start of the subject and immediately after *internal* newlines.
3322
3323 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3324 a pointer to an int instead of a pointer to an unsigned long int. This
3325 caused problems on 64-bit systems.
3326
3327 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3328 instance of the 'standard' template library not being so standard".
3329
3330 10. There was no check on the number of named subpatterns nor the maximum
3331 length of a subpattern name. The product of these values is used to compute
3332 the size of the memory block for a compiled pattern. By supplying a very
3333 long subpattern name and a large number of named subpatterns, the size
3334 computation could be caused to overflow. This is now prevented by limiting
3335 the length of names to 32 characters, and the number of named subpatterns
3336 to 10,000.
3337
3338 11. Subpatterns that are repeated with specific counts have to be replicated in
3339 the compiled pattern. The size of memory for this was computed from the
3340 length of the subpattern and the repeat count. The latter is limited to
3341 65535, but there was no limit on the former, meaning that integer overflow
3342 could in principle occur. The compiled length of a repeated subpattern is
3343 now limited to 30,000 bytes in order to prevent this.
3344
3345 12. Added the optional facility to have named substrings with the same name.
3346
3347 13. Added the ability to use a named substring as a condition, using the
3348 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3349 are numbers (not recommended). Forward references are permitted.
3350
3351 14. Added forward references in named backreferences (if you see what I mean).
3352
3353 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3354 pattern could run off the end of the subject. For example, the pattern
3355 "(?s)(.{1,5})"8 did this with the subject "ab".
3356
3357 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3358 PCRE_CASELESS was set when matching characters that were quantified with ?
3359 or *.
3360
3361 17. A character class other than a single negated character that had a minimum
3362 but no maximum quantifier - for example [ab]{6,} - was not handled
3363 correctly by pce_dfa_exec(). It would match only one character.
3364
3365 18. A valid (though odd) pattern that looked like a POSIX character
3366 class but used an invalid character after [ (for example [[,abc,]]) caused
3367 pcre_compile() to give the error "Failed: internal error: code overflow" or
3368 in some cases to crash with a glibc free() error. This could even happen if
3369 the pattern terminated after [[ but there just happened to be a sequence of
3370 letters, a binary zero, and a closing ] in the memory that followed.
3371
3372 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3373 over the years. Originally (before any Unicode support), just the bottom 8
3374 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3375 output from "man perlunicode" includes this:
3376
3377 The regular expression compiler produces polymorphic opcodes. That
3378 is, the pattern adapts to the data and automatically switches to
3379 the Unicode character scheme when presented with Unicode data--or
3380 instead uses a traditional byte scheme when presented with byte
3381 data.
3382
3383 Sadly, a wide octal escape does not cause a switch, and in a string with
3384 no other multibyte characters, these octal escapes are treated as before.
3385 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3386 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3387 Unicode string.
3388
3389 I have not perpetrated such confusion in PCRE. Up till now, it took just
3390 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3391 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3392 translate to the appropriate multibyte character.
3393
3394 29. Applied some refactoring to reduce the number of warnings from Microsoft
3395 and Borland compilers. This has included removing the fudge introduced
3396 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3397 a warning about an unused variable.
3398
3399 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3400 characters since release 4.0, because Perl (from release 5.004) does not.
3401 [Or at least, is documented not to: some releases seem to be in conflict
3402 with the documentation.] However, when a pattern was studied with
3403 pcre_study() and all its branches started with \s, PCRE still included VT
3404 as a possible starting character. Of course, this did no harm; it just
3405 caused an unnecessary match attempt.
3406
3407 22. Removed a now-redundant internal flag bit that recorded the fact that case
3408 dependency changed within the pattern. This was once needed for "required
3409 byte" processing, but is no longer used. This recovers a now-scarce options
3410 bit. Also moved the least significant internal flag bit to the most-
3411 significant bit of the word, which was not previously used (hangover from
3412 the days when it was an int rather than a uint) to free up another bit for
3413 the future.
3414
3415 23. Added support for CRLF line endings as well as CR and LF. As well as the
3416 default being selectable at build time, it can now be changed at runtime
3417 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3418 specify that it is scanning data with non-default line endings.
3419
3420 24. Changed the definition of CXXLINK to make it agree with the definition of
3421 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3422
3423 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3424 recursions. This makes a big different to stack usage for some patterns.
3425
3426 26. If a subpattern containing a named recursion or subroutine reference such
3427 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3428 the space required for the compiled pattern went wrong and gave too small a
3429 value. Depending on the environment, this could lead to "Failed: internal
3430 error: code overflow at offset 49" or "glibc detected double free or
3431 corruption" errors.
3432
3433 27. Applied patches from Google (a) to support the new newline modes and (b) to
3434 advance over multibyte UTF-8 characters in GlobalReplace.
3435
3436 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3437 difference for some implementation of PCRE in some Windows version.
3438
3439 29. Added some extra testing facilities to pcretest:
3440
3441 \q<number> in a data line sets the "match limit" value
3442 \Q<number> in a data line sets the "match recursion limt" value
3443 -S <number> sets the stack size, where <number> is in megabytes
3444
3445 The -S option isn't available for Windows.
3446
3447
3448 Version 6.6 06-Feb-06
3449 ---------------------
3450
3451 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3452 in pcreposix.h. I have copied the definition from pcre.h.
3453
3454 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3455 because pcre.h is no longer a built file.
3456
3457 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3458 not normally included in the compiled code.
3459
3460
3461 Version 6.5 01-Feb-06
3462 ---------------------
3463
3464 1. When using the partial match feature with pcre_dfa_exec(), it was not
3465 anchoring the second and subsequent partial matches at the new starting
3466 point. This could lead to incorrect results. For example, with the pattern
3467 /1234/, partially matching against "123" and then "a4" gave a match.
3468
3469 2. Changes to pcregrep:
3470
3471 (a) All non-match returns from pcre_exec() were being treated as failures
3472 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3473 error message is output. Some extra information is given for the
3474 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3475 probably the only errors that are likely to be caused by users (by
3476 specifying a regex that has nested indefinite repeats, for instance).
3477 If there are more than 20 of these errors, pcregrep is abandoned.
3478
3479 (b) A binary zero was treated as data while matching, but terminated the
3480 output line if it was written out. This has been fixed: binary zeroes
3481 are now no different to any other data bytes.
3482
3483 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3484 used to set a locale for matching. The --locale=xxxx long option has
3485 been added (no short equivalent) to specify a locale explicitly on the
3486 pcregrep command, overriding the environment variables.
3487
3488 (d) When -B was used with -n, some line numbers in the output were one less
3489 than they should have been.
3490
3491 (e) Added the -o (--only-matching) option.
3492
3493 (f) If -A or -C was used with -c (count only), some lines of context were
3494 accidentally printed for the final match.
3495
3496 (g) Added the -H (--with-filename) option.
3497
3498 (h) The combination of options -rh failed to suppress file names for files
3499 that were found from directory arguments.
3500
3501 (i) Added the -D (--devices) and -d (--directories) options.
3502
3503 (j) Added the -F (--fixed-strings) option.
3504
3505 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3506
3507 (l) Added the --colo(u)r option.
3508
3509 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3510 is not present by default.
3511
3512 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3513 items such as (?R) or (?1), when the recursion could match a number of
3514 alternatives. If it matched one of the alternatives, but subsequently,
3515 outside the recursion, there was a failure, the code tried to back up into
3516 the recursion. However, because of the way PCRE is implemented, this is not
3517 possible, and the result was an incorrect result from the match.
3518
3519 In order to prevent this happening, the specification of recursion has
3520 been changed so that all such subpatterns are automatically treated as
3521 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3522
3523 4. I had overlooked the fact that, in some locales, there are characters for
3524 which isalpha() is true but neither isupper() nor islower() are true. In
3525 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3526 and ordfeminine) are like this. This affected the treatment of \w and \W
3527 when they appeared in character classes, but not when they appeared outside
3528 a character class. The bit map for "word" characters is now created
3529 separately from the results of isalnum() instead of just taking it from the
3530 upper, lower, and digit maps. (Plus the underscore character, of course.)
3531
3532 5. The above bug also affected the handling of POSIX character classes such as
3533 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3534 permanent tables. Instead, the bit maps for such a class were previously
3535 created as the appropriate unions of the upper, lower, and digit bitmaps.
3536 Now they are created by subtraction from the [[:word:]] class, which has
3537 its own bitmap.
3538
3539 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3540 It is created by subtracting the vertical space characters (\x09, \x0a,
3541 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3542 subtraction was done in the overall bitmap for a character class, meaning
3543 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3544 be recognized. This bug has been fixed.
3545
3546 7. Patches from the folks at Google:
3547
3548 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3549 real life, but is still worth protecting against".
3550
3551 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3552 regular expressions".
3553
3554 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3555 have it.
3556
3557 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3558 "configure" and the latter not, in order to fix a problem somebody had
3559 with compiling the Arg class on HP-UX.
3560
3561 (e) Improve the error-handling of the C++ wrapper a little bit.
3562
3563 (f) New tests for checking recursion limiting.
3564
3565 8. The pcre_memmove() function, which is used only if the environment does not
3566 have a standard memmove() function (and is therefore rarely compiled),
3567 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3568 returning a result (though PCRE never actually uses the result).
3569
3570 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3571 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3572 returned instead of calling malloc() with an overflowing number that would
3573 most likely cause subsequent chaos.
3574
3575 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3576
3577 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3578 with this option is matched, the nmatch and pmatch options of regexec() are
3579 ignored.
3580
3581 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3582 provided in case anyone wants to the the POSIX interface with UTF-8
3583 strings.
3584
3585 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3586 C++ linking (needed for some HP-UX environments).
3587
3588 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3589 (unused parameter) and in the pcre_printint() function (omitted "default"
3590 switch label when the default is to do nothing).
3591
3592 15. Added some code to make it possible, when PCRE is compiled as a C++
3593 library, to replace subject pointers for pcre_exec() with a smart pointer
3594 class, thus making it possible to process discontinuous strings.
3595
3596 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3597 much the same function. They were added by different people who were trying
3598 to make PCRE easy to compile on non-Unix systems. It has been suggested
3599 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3600 for compiling on Windows systems. I have therefore replaced it with
3601 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3602 defaults to "extern" for C or "extern C" for C++, which works fine on
3603 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3604 SCOPE with something explicit in config.h. In addition:
3605
3606 (a) pcreposix.h still had just "extern" instead of either of these macros;
3607 I have replaced it with PCRE_DATA_SCOPE.
3608
3609 (b) Functions such as _pcre_xclass(), which are internal to the library,
3610 but external in the C sense, all had PCRE_EXPORT in their definitions.
3611 This is apparently wrong for the Windows case, so I have removed it.
3612 (It makes no difference on Unix-like systems.)
3613
3614 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3615 of recursive calls to match(). This is different to MATCH_LIMIT because
3616 that limits the total number of calls to match(), not all of which increase
3617 the depth of recursion. Limiting the recursion depth limits the amount of
3618 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3619 when PCRE is compiled, and changed at run time. A patch from Google adds
3620 this functionality to the C++ interface.
3621
3622 18. Changes to the handling of Unicode character properties:
3623
3624 (a) Updated the table to Unicode 4.1.0.
3625
3626 (b) Recognize characters that are not in the table as "Cn" (undefined).
3627
3628 (c) I revised the way the table is implemented to a much improved format
3629 which includes recognition of ranges. It now supports the ranges that
3630 are defined in UnicodeData.txt, and it also amalgamates other
3631 characters into ranges. This has reduced the number of entries in the
3632 table from around 16,000 to around 3,000, thus reducing its size
3633 considerably. I realized I did not need to use a tree structure after
3634 all - a binary chop search is just as efficient. Having reduced the
3635 number of entries, I extended their size from 6 bytes to 8 bytes to
3636 allow for more data.
3637
3638 (d) Added support for Unicode script names via properties such as \p{Han}.
3639
3640 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3641 matching that character.
3642
3643 20. When matching a repeated Unicode property with a minimum greater than zero,
3644 (for example \pL{2,}), PCRE could look past the end of the subject if it
3645 reached it while seeking the minimum number of characters. This could
3646 happen only if some of the characters were more than one byte long, because
3647 there is a check for at least the minimum number of bytes.
3648
3649 21. Refactored the implementation of \p and \P so as to be more general, to
3650 allow for more different types of property in future. This has changed the
3651 compiled form incompatibly. Anybody with saved compiled patterns that use
3652 \p or \P will have to recompile them.
3653
3654 22. Added "Any" and "L&" to the supported property types.
3655
3656 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3657 but give a compile time error if the value is greater than 0xff.
3658
3659 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3660 accidentally not being installed or uninstalled.
3661
3662 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3663 made were to insert the current release number. This seemed silly, because
3664 it made things harder for people building PCRE on systems that don't run
3665 "configure". I have turned pcre.h into a distributed file, no longer built
3666 by "configure", with the version identification directly included. There is
3667 no longer a pcre.h.in file.
3668
3669 However, this change necessitated a change to the pcre-config script as
3670 well. It is built from pcre-config.in, and one of the substitutions was the
3671 release number. I have updated configure.ac so that ./configure now finds
3672 the release number by grepping pcre.h.
3673
3674 26. Added the ability to run the tests under valgrind.
3675
3676
3677 Version 6.4 05-Sep-05
3678 ---------------------
3679
3680 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3681 "--" to be printed when multiple files were scanned, even when none of the
3682 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3683 consider it to be a bug, and have restored the previous behaviour.
3684
3685 2. A couple of code tidies to get rid of compiler warnings.
3686
3687 3. The pcretest program used to cheat by referring to symbols in the library
3688 whose names begin with _pcre_. These are internal symbols that are not
3689 really supposed to be visible externally, and in some environments it is
3690 possible to suppress them. The cheating is now confined to including
3691 certain files from the library's source, which is a bit cleaner.
3692
3693 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3694 file's purpose clearer.
3695
3696 5. Reorganized pcre_ucp_findchar().
3697
3698
3699 Version 6.3 15-Aug-05
3700 ---------------------
3701
3702 1. The file libpcre.pc.in did not have general read permission in the tarball.
3703
3704 2. There were some problems when building without C++ support:
3705
3706 (a) If C++ support was not built, "make install" and "make test" still
3707 tried to test it.
3708
3709 (b) There were problems when the value of CXX was explicitly set. Some
3710 changes have been made to try to fix these, and ...
3711
3712 (c) --disable-cpp can now be used to explicitly disable C++ support.
3713
3714 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3715 backslash in a target when C++ was disabled. This confuses some
3716 versions of "make", apparently. Using an intermediate variable solves
3717 this. (Same for CPP_LOBJ.)
3718
3719 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3720 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3721 necessary on certain architectures.
3722
3723 4. Added a setting of -export-symbols-regex to the link command to remove
3724 those symbols that are exported in the C sense, but actually are local
3725 within the library, and not documented. Their names all begin with
3726 "_pcre_". This is not a perfect job, because (a) we have to except some
3727 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3728 available (and never for static libraries). I have made a note to try to
3729 find a way round (a) in the future.
3730
3731
3732 Version 6.2 01-Aug-05
3733 ---------------------
3734
3735 1. There was no test for integer overflow of quantifier values. A construction
3736 such as {1111111111111111} would give undefined results. What is worse, if
3737 a minimum quantifier for a parenthesized subpattern overflowed and became
3738 negative, the calculation of the memory size went wrong. This could have
3739 led to memory overwriting.
3740
3741 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3742
3743 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3744 operating environments where this matters.
3745
3746 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3747 PCRE options from within the C++ wrapper.
3748
3749 5. Named capturing subpatterns were not being correctly counted when a pattern
3750 was compiled. This caused two problems: (a) If there were more than 100
3751 such subpatterns, the calculation of the memory needed for the whole
3752 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3753 back references of the form \12, where the number was greater than 9, were
3754 not recognized as back references, even though there were sufficient
3755 previous subpatterns.
3756
3757 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3758 versions of gcc, e.g. 2.95.4.
3759
3760
3761 Version 6.1 21-Jun-05
3762 ---------------------
3763
3764 1. There was one reference to the variable "posix" in pcretest.c that was not
3765 surrounded by "#if !defined NOPOSIX".
3766
3767 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3768 the cross-check on the old pcre_info() function, for the benefit of the
3769 cut-down version of PCRE that is currently imported into Exim.
3770
3771 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3772 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3773 patterns that start (?i) but I don't think that matters. The use of (?i) is
3774 just an example; this all applies to the other options as well.
3775
3776 4. Since libtool seems to echo the compile commands it is issuing, the output
3777 from "make" can be reduced a bit by putting "@" in front of each libtool
3778 compile command.
3779
3780 5. Patch from the folks at Google for configure.in to be a bit more thorough
3781 in checking for a suitable C++ installation before trying to compile the
3782 C++ stuff. This should fix a reported problem when a compiler was present,
3783 but no suitable headers.
3784
3785 6. The man pages all had just "PCRE" as their title. I have changed them to
3786 be the relevant file name. I have also arranged that these names are
3787 retained in the file doc/pcre.txt, which is a concatenation in text format
3788 of all the man pages except the little individual ones for each function.
3789
3790 7. The NON-UNIX-USE file had not been updated for the different set of source
3791 files that come with release 6. I also added a few comments about the C++
3792 wrapper.
3793
3794
3795 Version 6.0 07-Jun-05
3796 ---------------------
3797
3798 1. Some minor internal re-organization to help with my DFA experiments.
3799
3800 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3801 didn't matter for the library itself when fully configured, but did matter
3802 when compiling without UCP support, or within Exim, where the ucp files are
3803 not imported.
3804
3805 3. Refactoring of the library code to split up the various functions into
3806 different source modules. The addition of the new DFA matching code (see
3807 below) to a single monolithic source would have made it really too
3808 unwieldy, quite apart from causing all the code to be include in a
3809 statically linked application, when only some functions are used. This is
3810 relevant even without the DFA addition now that patterns can be compiled in
3811 one application and matched in another.
3812
3813 The downside of splitting up is that there have to be some external
3814 functions and data tables that are used internally in different modules of
3815 the library but which are not part of the API. These have all had their
3816 names changed to start with "_pcre_" so that they are unlikely to clash
3817 with other external names.
3818
3819 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3820 a different (DFA) algorithm. Although it is slower than the original
3821 function, it does have some advantages for certain types of matching
3822 problem.
3823
3824 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3825 including restarting after a partial match.
3826
3827 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3828 defined when compiling for Windows was sent to me. I have put it into the
3829 code, though I have no means of testing or verifying it.
3830
3831 7. Added the pcre_refcount() auxiliary function.
3832
3833 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3834 match before or at the first newline in the subject string. In pcretest,
3835 the /f option on a pattern can be used to set this.
3836
3837 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3838 would behave wrongly. This has been present in PCRE since release 4.0.
3839
3840 10. A number of changes to the pcregrep command:
3841
3842 (a) Refactored how -x works; insert ^(...)$ instead of setting
3843 PCRE_ANCHORED and checking the length, in preparation for adding
3844 something similar for -w.
3845
3846 (b) Added the -w (match as a word) option.
3847
3848 (c) Refactored the way lines are read and buffered so as to have more
3849 than one at a time available.
3850
3851 (d) Implemented a pcregrep test script.
3852
3853 (e) Added the -M (multiline match) option. This allows patterns to match
3854 over several lines of the subject. The buffering ensures that at least
3855 8K, or the rest of the document (whichever is the shorter) is available
3856 for matching (and similarly the previous 8K for lookbehind assertions).
3857
3858 (f) Changed the --help output so that it now says
3859
3860 -w, --word-regex(p)
3861
3862 instead of two lines, one with "regex" and the other with "regexp"
3863 because that confused at least one person since the short forms are the
3864 same. (This required a bit of code, as the output is generated
3865 automatically from a table. It wasn't just a text change.)
3866
3867 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3868 option but starts with a hyphen. Could be a pattern or a path name
3869 starting with a hyphen, for instance.
3870
3871 (h) "-" can be given as a file name to represent stdin.
3872
3873 (i) When file names are being printed, "(standard input)" is used for
3874 the standard input, for compatibility with GNU grep. Previously
3875 "<stdin>" was used.
3876
3877 (j) The option --label=xxx can be used to supply a name to be used for
3878 stdin when file names are being printed. There is no short form.
3879
3880 (k) Re-factored the options decoding logic because we are going to add
3881 two more options that take data. Such options can now be given in four
3882 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3883
3884 (l) Added the -A, -B, and -C options for requesting that lines of context
3885 around matches be printed.
3886
3887 (m) Added the -L option to print the names of files that do not contain
3888 any matching lines, that is, the complement of -l.
3889
3890 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3891 continue to scan other files.
3892
3893 (o) The -s option was incorrectly implemented. For compatibility with other
3894 greps, it now suppresses the error message for a non-existent or non-
3895 accessible file (but not the return code). There is a new option called
3896 -q that suppresses the output of matching lines, which was what -s was
3897 previously doing.
3898
3899 (p) Added --include and --exclude options to specify files for inclusion
3900 and exclusion when recursing.
3901
3902 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3903 Hopefully, it now does.
3904
3905 12. Missing cast in pcre_study().
3906
3907 13. Added an "uninstall" target to the makefile.
3908
3909 14. Replaced "extern" in the function prototypes in Makefile.in with
3910 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3911 world, but is set differently for Windows.
3912
3913 15. Added a second compiling function called pcre_compile2(). The only
3914 difference is that it has an extra argument, which is a pointer to an
3915 integer error code. When there is a compile-time failure, this is set
3916 non-zero, in addition to the error test pointer being set to point to an
3917 error message. The new argument may be NULL if no error number is required
3918 (but then you may as well call pcre_compile(), which is now just a
3919 wrapper). This facility is provided because some applications need a
3920 numeric error indication, but it has also enabled me to tidy up the way
3921 compile-time errors are handled in the POSIX wrapper.
3922
3923 16. Added VPATH=.libs to the makefile; this should help when building with one
3924 prefix path and installing with another. (Or so I'm told by someone who
3925 knows more about this stuff than I do.)
3926
3927 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3928 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3929 match everything, including newlines. This is not POSIX-compatible, but
3930 somebody wanted the feature. From pcretest it can be activated by using
3931 both the P and the s flags.
3932
3933 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3934
3935 19. libpcre.pc was being incorrectly installed as executable.
3936
3937 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3938 it now also looks for '\r' so that it will work unmodified on Windows.
3939
3940 21. Added Google's contributed C++ wrapper to the distribution.
3941
3942 22. Added some untidy missing memory free() calls in pcretest, to keep
3943 Electric Fence happy when testing.
3944
3945
3946
3947 Version 5.0 13-Sep-04
3948 ---------------------
3949
3950 1. Internal change: literal characters are no longer packed up into items
3951 containing multiple characters in a single byte-string. Each character
3952 is now matched using a separate opcode. However, there may be more than one
3953 byte in the character in UTF-8 mode.
3954
3955 2. The pcre_callout_block structure has two new fields: pattern_position and
3956 next_item_length. These contain the offset in the pattern to the next match
3957 item, and its length, respectively.
3958
3959 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3960 insertion of callouts before each pattern item. Added the /C option to
3961 pcretest to make use of this.
3962
3963 4. On the advice of a Windows user, the lines
3964
3965 #if defined(_WIN32) || defined(WIN32)
3966 _setmode( _fileno( stdout ), 0x8000 );
3967 #endif /* defined(_WIN32) || defined(WIN32) */
3968
3969 have been added to the source of pcretest. This apparently does useful
3970 magic in relation to line terminators.
3971
3972 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3973 for the benefit of those environments where the "b" makes a difference.
3974
3975 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3976 to know about it. I have put a hack into configure.in that adds in code
3977 to set GCC=yes if CC=icc. This seems to end up at a point in the
3978 generated configure script that is early enough to affect the setting of
3979 compiler options, which is what is needed, but I have no means of testing
3980 whether it really works. (The user who reported this had patched the
3981 generated configure script, which of course I cannot do.)
3982
3983 LATER: After change 22 below (new libtool files), the configure script
3984 seems to know about icc (and also ecc). Therefore, I have commented out
3985 this hack in configure.in.
3986
3987 7. Added support for pkg-config (2 patches were sent in).
3988
3989 8. Negated POSIX character classes that used a combination of internal tables
3990 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3991 [[:^ascii]]. Typically, they would match almost any characters. The other
3992 POSIX classes were not broken in this way.
3993
3994 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3995 to find the match, as PCRE was deluded into thinking that the match had to
3996 start at the start point or following a newline. The same bug applied to
3997 patterns with negative forward assertions or any backward assertions
3998 preceding ".*" at the start, unless the pattern required a fixed first
3999 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4000
4001 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4002 starting at the last subject character, bytes beyond the end of the subject
4003 string were read.
4004
4005 11. Renamed the variable "class" as "classbits" to make life easier for C++
4006 users. (Previously there was a macro definition, but it apparently wasn't
4007 enough.)
4008
4009 12. Added the new field "tables" to the extra data so that tables can be passed
4010 in at exec time, or the internal tables can be re-selected. This allows
4011 a compiled regex to be saved and re-used at a later time by a different
4012 program that might have everything at different addresses.
4013
4014 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4015 -R library as well as a -L library.
4016
4017 14. The debugging options of pcretest (-d on the command line or D on a
4018 pattern) showed incorrect output for anything following an extended class
4019 that contained multibyte characters and which was followed by a quantifier.
4020
4021 15. Added optional support for general category Unicode character properties
4022 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4023 support. It adds about 90K to the size of the library. The meanings of the
4024 inbuilt class escapes such as \d and \s have NOT been changed.
4025
4026 16. Updated pcredemo.c to include calls to free() to release the memory for the
4027 compiled pattern.
4028
4029 17. The generated file chartables.c was being created in the source directory
4030 instead of in the building directory. This caused the build to fail if the
4031 source directory was different from the building directory, and was
4032 read-only.
4033
4034 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4035 file. No doubt somebody will tell me if they don't make sense... Also added
4036 Dan Mooney's comments about building on OpenVMS.
4037
4038 19. Added support for partial matching via the PCRE_PARTIAL option for
4039 pcre_exec() and the \P data escape in pcretest.
4040
4041 20. Extended pcretest with 3 new pattern features:
4042
4043 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4044 write the compiled pattern to the file whose name is "rest-of-line".
4045 This is a straight binary dump of the data, with the saved pointer to
4046 the character tables forced to be NULL. The study data, if any, is
4047 written too. After writing, pcretest reads a new pattern.
4048
4049 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4050 compiled pattern from the given file. There must not be any
4051 occurrences of "<" in the file name (pretty unlikely); if there are,
4052 pcretest will instead treat the initial "<" as a pattern delimiter.
4053 After reading in the pattern, pcretest goes on to read data lines as
4054 usual.
4055
4056 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4057 and 16-bit fields in a compiled pattern, to simulate a pattern that
4058 was compiled on a host of opposite endianness.
4059
4060 21. The pcre-exec() function can now cope with patterns that were compiled on
4061 hosts of opposite endianness, with this restriction:
4062
4063 As for any compiled expression that is saved and used later, the tables
4064 pointer field cannot be preserved; the extra_data field in the arguments
4065 to pcre_exec() should be used to pass in a tables address if a value
4066 other than the default internal tables were used at compile time.
4067
4068 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4069 now diagnosed as an error. Previously, most of the time, a negative number
4070 would have been treated as zero, but if in addition "ovector" was passed as
4071 NULL, a crash could occur.
4072
4073 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4074 new versions from the libtool 1.5 distribution (the last one is a copy of
4075 a file called libtool.m4). This seems to have fixed the need to patch
4076 "configure" to support Darwin 1.3 (which I used to do). However, I still
4077 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4078 workstation).
4079
4080 24. Changed the PCRE licence to be the more standard "BSD" licence.
4081
4082
4083 Version 4.5 01-Dec-03
4084 ---------------------
4085
4086 1. There has been some re-arrangement of the code for the match() function so
4087 that it can be compiled in a version that does not call itself recursively.
4088 Instead, it keeps those local variables that need separate instances for
4089 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4090 needs to "recurse". Keeping track of where control must go is done by means
4091 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4092 hide most of the details from the main code, and operates only if
4093 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4094 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4095 operating.
4096
4097 To make it easier for callers to provide specially tailored get/free
4098 functions for this usage, two new functions, pcre_stack_malloc, and
4099 pcre_stack_free, are used. They are always called in strict stacking order,
4100 and the size of block requested is always the same.
4101
4102 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4103 PCRE has been compiled to use the stack or the heap for recursion. The
4104 -C option of pcretest uses this to show which version is compiled.
4105
4106 A new data escape \S, is added to pcretest; it causes the amounts of store
4107 obtained and freed by both kinds of malloc/free at match time to be added
4108 to the output.
4109
4110 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4111 what's available on my current Linux desktop machine.
4112
4113 3. When matching a UTF-8 string, the test for a valid string at the start has
4114 been extended. If start_offset is not zero, PCRE now checks that it points
4115 to a byte that is the start of a UTF-8 character. If not, it returns
4116 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4117 this is necessary because there may be backward assertions in the pattern.
4118 When matching the same subject several times, it may save resources to use
4119 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4120
4121 4. The code for checking the validity of UTF-8 strings has been tightened so
4122 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4123 containing "overlong sequences".
4124
4125 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4126 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4127 should have been "&", but it just so happened that all the cases this let
4128 through by mistake were picked up later in the function.
4129
4130 6. I had used a variable called "isblank" - this is a C99 function, causing
4131 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4132
4133 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4134 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4135 so the version is shown; (c) stop "make test" echoing "./RunTest".
4136
4137 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4138
4139 9. The prototype for memmove() for systems that don't have it was using
4140 size_t, but the inclusion of the header that defines size_t was later. I've
4141 moved the #includes for the C headers earlier to avoid this.
4142
4143 10. Added some adjustments to the code to make it easier to compiler on certain
4144 special systems:
4145
4146 (a) Some "const" qualifiers were missing.
4147 (b) Added the macro EXPORT before all exported functions; by default this
4148 is defined to be empty.
4149 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4150 that it reads its output file name as an argument instead of writing
4151 to the standard output and assuming this can be redirected.
4152
4153 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4154 class containing characters with values greater than 255, PCRE compilation
4155 went into a loop.
4156
4157 12. A recursive reference to a subpattern that was within another subpattern
4158 that had a minimum quantifier of zero caused PCRE to crash. For example,
4159 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4160 recursion. If the recursively-called subpattern itself had a zero repeat,
4161 that was OK.
4162
4163 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4164 buffer into which it was copied (for escape processing) was still set at
4165 1024, so long lines caused crashes.
4166
4167 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4168 "internal error: code overflow...". This applied to any character class
4169 that was followed by a possessive quantifier.
4170
4171 15. Modified the Makefile to add libpcre.la as a prerequisite for
4172 libpcreposix.la because I was told this is needed for a parallel build to
4173 work.
4174
4175 16. If a pattern that contained .* following optional items at the start was
4176 studied, the wrong optimizing data was generated, leading to matching
4177 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4178 matching string must start with a or b or c. The correct conclusion for
4179 this pattern is that a match can start with any character.
4180
4181
4182 Version 4.4 13-Aug-03
4183 ---------------------
4184
4185 1. In UTF-8 mode, a character class containing characters with values between
4186 127 and 255 was not handled correctly if the compiled pattern was studied.
4187 In fixing this, I have also improved the studying algorithm for such
4188 classes (slightly).
4189
4190 2. Three internal functions had redundant arguments passed to them. Removal
4191 might give a very teeny performance improvement.
4192
4193 3. Documentation bug: the value of the capture_top field in a callout is *one
4194 more than* the number of the hightest numbered captured substring.
4195
4196 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4197 in incorrectly linking with a previously installed version. They now link
4198 explicitly with libpcre.la.
4199
4200 5. configure.in no longer needs to recognize Cygwin specially.
4201
4202 6. A problem in pcre.in for Windows platforms is fixed.
4203
4204 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4205 pcretest, it used to include the size of the study block as part of its
4206 output. Unfortunately, the structure contains a field that has a different
4207 size on different hardware architectures. This meant that the tests that
4208 showed this size failed. As the block is currently always of a fixed size,
4209 this information isn't actually particularly useful in pcretest output, so
4210 I have just removed it.
4211
4212 8. Three pre-processor statements accidentally did not start in column 1.
4213 Sadly, there are *still* compilers around that complain, even though
4214 standard C has not required this for well over a decade. Sigh.
4215
4216 9. In pcretest, the code for checking callouts passed small integers in the
4217 callout_data field, which is a void * field. However, some picky compilers
4218 complained about the casts involved for this on 64-bit systems. Now
4219 pcretest passes the address of the small integer instead, which should get
4220 rid of the warnings.
4221
4222 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4223 both compile and run time, and gives an error if an invalid UTF-8 sequence
4224 is found. There is a option for disabling this check in cases where the
4225 string is known to be correct and/or the maximum performance is wanted.
4226
4227 11. In response to a bug report, I changed one line in Makefile.in from
4228
4229 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4230 to
4231 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4232
4233 to look similar to other lines, but I have no way of telling whether this
4234 is the right thing to do, as I do not use Windows. No doubt I'll get told
4235 if it's wrong...
4236
4237
4238 Version 4.3 21-May-03
4239 ---------------------
4240
4241 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4242 Makefile.
4243
4244 2. Some refactoring to improve the quality of the code:
4245
4246 (i) The utf8_table... variables are now declared "const".
4247
4248 (ii) The code for \cx, which used the "case flipping" table to upper case
4249 lower case letters, now just substracts 32. This is ASCII-specific,
4250 but the whole concept of \cx is ASCII-specific, so it seems
4251 reasonable.
4252
4253 (iii) PCRE was using its character types table to recognize decimal and
4254 hexadecimal digits in the pattern. This is silly, because it handles
4255 only 0-9, a-f, and A-F, but the character types table is locale-
4256 specific, which means strange things might happen. A private
4257 table is now used for this - though it costs 256 bytes, a table is
4258 much faster than multiple explicit tests. Of course, the standard
4259 character types table is still used for matching digits in subject
4260 strings against \d.
4261
4262 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4263 ending in _t are). So I've renamed it as ESC_tee.
4264
4265 3. The first argument for regexec() in the POSIX wrapper should have been
4266 defined as "const".
4267
4268 4. Changed pcretest to use malloc() for its buffers so that they can be
4269 Electric Fenced for debugging.
4270
4271 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4272 to read one or more bytes before the start of the subject string. Often this
4273 had no effect on PCRE's behaviour, but in some circumstances it could
4274 provoke a segmentation fault.
4275
4276 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4277 to try to read one or more bytes before the start of the subject string.
4278
4279 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4280 UTF-8 support could misbehave in various ways if the subject string
4281 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4282 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4283 back over UTF-8 characters.)
4284
4285
4286 Version 4.2 14-Apr-03
4287 ---------------------
4288
4289 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4290
4291 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4292 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4293 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4294 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4295 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4296 and BUILD_EXEEXT
4297 Note: automatic setting of the BUILD variables is not yet working
4298 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4299 compile-time but not at link-time
4300 [LINK]: use for linking executables only
4301 make different versions for Windows and non-Windows
4302 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4303 libraries
4304 [LINK_FOR_BUILD]: new variable
4305 [OBJEXT]: use throughout
4306 [EXEEXT]: use throughout
4307 <winshared>: new target
4308 <wininstall>: new target
4309 <dftables.o>: use native compiler
4310 <dftables>: use native linker
4311 <install>: handle Windows platform correctly
4312 <clean>: ditto
4313 <check>: ditto
4314 copy DLL to top builddir before testing
4315
4316 As part of these changes, -no-undefined was removed again. This was reported
4317 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4318 in any case.
4319
4320 3. Some tidies to get rid of compiler warnings:
4321
4322 . In the match_data structure, match_limit was an unsigned long int, whereas
4323 match_call_count was an int. I've made them both unsigned long ints.
4324
4325 . In pcretest the fact that a const uschar * doesn't automatically cast to
4326 a void * provoked a warning.
4327
4328 . Turning on some more compiler warnings threw up some "shadow" variables
4329 and a few more missing casts.
4330
4331 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4332 option, a class that contained a single character with a value between 128
4333 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4334
4335 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4336 option, a class that contained several characters, but with at least one
4337 whose value was between 128 and 255 caused PCRE to crash.
4338
4339
4340 Version 4.1 12-Mar-03
4341 ---------------------
4342
4343 1. Compiling with gcc -pedantic found a couple of places where casts were
4344 needed, and a string in dftables.c that was longer than standard compilers are
4345 required to support.
4346
4347 2. Compiling with Sun's compiler found a few more places where the code could
4348 be tidied up in order to avoid warnings.
4349
4350 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4351 first of these names is deprecated in the latest Autoconf in favour of the name
4352 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4353 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4354 analogy I have changed it to CFLAGS_FOR_BUILD.
4355
4356 4. Added -no-undefined to the linking command in the Makefile, because this is
4357 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4358 linking step for the pcreposix library.
4359
4360 5. PCRE was failing to diagnose the case of two named groups with the same
4361 name.
4362
4363 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4364 literal character that is needed in the subject for a match, and scans along to
4365 ensure that it is present before embarking on the full matching process. This
4366 saves time in cases of nested unlimited repeats that are never going to match.
4367 Problem: the scan can take a lot of time if the subject is very long (e.g.
4368 megabytes), thus penalizing straightforward matches. It is now done only if the
4369 amount of subject to be scanned is less than 1000 bytes.
4370
4371 7. A lesser problem with the same optimization is that it was recording the
4372 first character of an anchored pattern as "needed", thus provoking a search
4373 right along the subject, even when the first match of the pattern was going to
4374 fail. The "needed" character is now not set for anchored patterns, unless it
4375 follows something in the pattern that is of non-fixed length. Thus, it still
4376 fulfils its original purpose of finding quick non-matches in cases of nested
4377 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4378
4379
4380 Version 4.0 17-Feb-03
4381 ---------------------
4382
4383 1. If a comment in an extended regex that started immediately after a meta-item
4384 extended to the end of string, PCRE compiled incorrect data. This could lead to
4385 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4386
4387 2. Moved to autoconf 2.53 and libtool 1.4.2.
4388
4389 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4390 the special perltest8 script is no longer needed - all the tests can be run
4391 from a single perltest script.
4392
4393 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4394 by \s. It has now been removed in PCRE. This means it isn't recognized as
4395 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4396 class [:space:] *does* include VT, thereby creating a mess.
4397
4398 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4399 space and tab.
4400
4401 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4402 its new features into the main test script, reducing the number of scripts.
4403
4404 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4405 were backward compatible, and made the (?i) apply to the whole pattern, as if
4406 /i were given. Now it behaves more logically, and applies the option setting
4407 only to what follows. PCRE has been changed to follow suit. However, if it
4408 finds options settings right at the start of the pattern, it extracts them into
4409 the global options, as before. Thus, they show up in the info data.
4410
4411 8. Added support for the \Q...\E escape sequence. Characters in between are
4412 treated as literals. This is slightly different from Perl in that $ and @ are
4413 also handled as literals inside the quotes. In Perl, they will cause variable
4414 interpolation. Note the following examples:
4415
4416 Pattern PCRE matches Perl matches
4417
4418 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4419 \Qabc\$xyz\E abc\$xyz abc\$xyz
4420 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4421
4422 For compatibility with Perl, \Q...\E sequences are recognized inside character
4423 classes as well as outside them.
4424
4425 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4426 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4427 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4428 signed/unsigned warnings.
4429
4430 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4431 option for pcretest, so I've replaced it by a simple function that does just
4432 that job.
4433
4434 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4435 "pcregrep -".
4436
4437 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4438 Java package. This provides some syntactic sugar for simple cases of what my
4439 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4440 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4441 item, you can use this simplified notation. Note that only makes sense with
4442 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4443 greediness, whatever the setting of the PCRE_UNGREEDY option.
4444
4445 13. A change of greediness default within a pattern was not taking effect at
4446 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4447 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4448 was abstracted outside.
4449
4450 14. PCRE now supports the \G assertion. It is true when the current matching
4451 position is at the start point of the match. This differs from \A when the
4452 starting offset is non-zero. Used with the /g option of pcretest (or similar
4453 code), it works in the same way as it does for Perl's /g option. If all
4454 alternatives of a regex begin with \G, the expression is anchored to the start
4455 match position, and the "anchored" flag is set in the compiled expression.
4456
4457 15. Some bugs concerning the handling of certain option changes within patterns
4458 have been fixed. These applied to options other than (?ims). For example,
4459 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4460 been the other way round. Some of this was related to change 7 above.
4461
4462 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4463 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4464 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4465 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4466
4467 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4468 mode. Unlike ".", it always matches newline, whatever the setting of
4469 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4470 assertions. Perl allows it, but it doesn't (in general) work because it can't
4471 calculate the length of the lookbehind. At least, that's the case for Perl
4472 5.8.0 - I've been told they are going to document that it doesn't work in
4473 future.
4474
4475 18. Added an error diagnosis for escapes that PCRE does not support: these are
4476 \L, \l, \N, \P, \p, \U, \u, and \X.
4477
4478 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4479 reading past the end of the pattern in cases such as /[abcd/.
4480
4481 20. PCRE was getting more memory than necessary for patterns with classes that
4482 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4483
4484 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4485 compiling PCRE for use with Virtual Pascal.
4486
4487 22. Small fix to the Makefile to make it work properly if the build is done
4488 outside the source tree.
4489
4490 23. Added a new extension: a condition to go with recursion. If a conditional
4491 subpattern starts with (?(R) the "true" branch is used if recursion has
4492 happened, whereas the "false" branch is used only at the top level.
4493
4494 24. When there was a very long string of literal characters (over 255 bytes
4495 without UTF support, over 250 bytes with UTF support), the computation of how
4496 much memory was required could be incorrect, leading to segfaults or other
4497 strange effects.
4498
4499 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4500 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4501 there was a subsequent back reference to those brackets. This meant that, for
4502 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4503 possible to check for precisely this case. All we can do is abandon the
4504 optimization if .* occurs inside capturing brackets when there are any back
4505 references whatsoever. (See below for a better fix that came later.)
4506
4507 26. The handling of the optimization for finding the first character of a
4508 non-anchored pattern, and for finding a character that is required later in the
4509 match were failing in some cases. This didn't break the matching; it just
4510 failed to optimize when it could. The way this is done has been re-implemented.
4511
4512 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4513
4514 28. Added a new feature that provides some of the functionality that Perl
4515 provides with (?{...}). The facility is termed a "callout". The way it is done
4516 in PCRE is for the caller to provide an optional function, by setting
4517 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4518 global variable. By default it is unset, which disables all calling out. To get
4519 the function called, the regex must include (?C) at appropriate points. This
4520 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4521 This provides a means of identifying different callout points. When PCRE
4522 reaches such a point in the regex, if pcre_callout has been set, the external
4523 function is called. It is provided with data in a structure called
4524 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4525 matching continues; if it returns a non-zero value, the match at the current
4526 point fails. However, backtracking will occur if possible. [This was changed
4527 later and other features added - see item 49 below.]
4528
4529 29. pcretest is upgraded to test the callout functionality. It provides a
4530 callout function that displays information. By default, it shows the start of
4531 the match and the current position in the text. There are some new data escapes
4532 to vary what happens:
4533
4534 \C+ in addition, show current contents of captured substrings
4535 \C- do not supply a callout function
4536 \C!n return 1 when callout number n is reached
4537 \C!n!m return 1 when callout number n is reached for the mth time
4538
4539 30. If pcregrep was called with the -l option and just a single file name, it
4540 output "<stdin>" if a match was found, instead of the file name.
4541
4542 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4543 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4544 pcre_exec(). This saves a malloc/free per call. The default value of
4545 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4546 when configuring.
4547
4548 32. The default maximum size of a compiled pattern is 64K. There have been a
4549 few cases of people hitting this limit. The code now uses macros to handle the
4550 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4551 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4552 configuring. Tests 2 and 5 work only with 2-byte links because they output
4553 debugging information about compiled patterns.
4554
4555 33. Internal code re-arrangements:
4556
4557 (a) Moved the debugging function for printing out a compiled regex into
4558 its own source file (printint.c) and used #include to pull it into
4559 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4560 separate copies.
4561
4562 (b) Defined the list of op-code names for debugging as a macro in
4563 internal.h so that it is next to the definition of the opcodes.
4564
4565 (c) Defined a table of op-code lengths for simpler skipping along compiled
4566 code. This is again a macro in internal.h so that it is next to the
4567 definition of the opcodes.
4568
4569 34. Added support for recursive calls to individual subpatterns, along the
4570 lines of Robin Houston's patch (but implemented somewhat differently).
4571
4572 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4573 allow it to read and process whole directories in Win32. This code was
4574 contributed by Lionel Fourquaux; it has not been tested by me.
4575
4576 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4577 used to name a group. Names consist of alphanumerics and underscores, and must
4578 be unique. Back references use the syntax (?P=name) and recursive calls use
4579 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4580 numbers. The function pcre_fullinfo() can be used after compilation to extract
4581 a name/number map. There are three relevant calls:
4582
4583 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4584 PCRE_INFO_NAMECOUNT yields the number of entries
4585 PCRE_INFO_NAMETABLE yields a pointer to the map.
4586
4587 The map is a vector of fixed-size entries. The size of each entry depends on
4588 the length of the longest name used. The first two bytes of each entry are the
4589 group number, most significant byte first. There follows the corresponding
4590 name, zero terminated. The names are in alphabetical order.
4591
4592 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4593 case instead of 255. Making it the same both with and without UTF-8 support
4594 means that the same test output works with both.
4595
4596 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4597 calling malloc() with a zero argument.
4598
4599 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4600 optimization. I've improved things by keeping a bitmap of backreferences with
4601 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4602 fact referenced, the optimization can be applied. It is unlikely that a
4603 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4604 the match to follow \n) will appear inside brackets with a number greater than
4605 31, but if it does, any back reference > 31 suppresses the optimization.
4606
4607 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4608 of disabling numbered capturing parentheses. Any opening parenthesis that is
4609 not followed by ? behaves as if it were followed by ?: but named parentheses
4610 can still be used for capturing (and they will acquire numbers in the usual
4611 way).
4612
4613 41. Redesigned the return codes from the match() function into yes/no/error so
4614 that errors can be passed back from deep inside the nested calls. A malloc
4615 failure while inside a recursive subpattern call now causes the
4616 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4617
4618 42. It is now possible to set a limit on the number of times the match()
4619 function is called in a call to pcre_exec(). This facility makes it possible to
4620 limit the amount of recursion and backtracking, though not in a directly
4621 obvious way, because the match() function is used in a number of different
4622 circumstances. The count starts from zero for each position in the subject
4623 string (for non-anchored patterns). The default limit is, for compatibility, a
4624 large number, namely 10 000 000. You can change this in two ways:
4625
4626 (a) When configuring PCRE before making, you can use --with-match-limit=n
4627 to set a default value for the compiled library.
4628
4629 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4630 a different value is set. See 45 below.
4631
4632 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4633
4634 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4635 of things that can be changed at compile time. The first argument specifies
4636 what is wanted and the second points to where the information is to be placed.
4637 The current list of available information is:
4638
4639 PCRE_CONFIG_UTF8
4640
4641 The output is an integer that is set to one if UTF-8 support is available;
4642 otherwise it is set to zero.
4643
4644 PCRE_CONFIG_NEWLINE
4645
4646 The output is an integer that it set to the value of the code that is used for
4647 newline. It is either LF (10) or CR (13).
4648
4649 PCRE_CONFIG_LINK_SIZE
4650
4651 The output is an integer that contains the number of bytes used for internal
4652 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4653
4654 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4655
4656 The output is an integer that contains the threshold above which the POSIX
4657 interface uses malloc() for output vectors. See item 31 above.
4658
4659 PCRE_CONFIG_MATCH_LIMIT
4660
4661 The output is an unsigned integer that contains the default limit of the number
4662 of match() calls in a pcre_exec() execution. See 42 above.
4663
4664 44. pcretest has been upgraded by the addition of the -C option. This causes it
4665 to extract all the available output from the new pcre_config() function, and to
4666 output it. The program then exits immediately.
4667
4668 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4669 order to support additional features. One way would have been to define
4670 pcre_exec2() (for example) with extra arguments, but this would not have been
4671 extensible, and would also have required all calls to the original function to
4672 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4673 is used for passing in "extra" data from pcre_study().
4674
4675 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4676 contains the following fields:
4677
4678 flags a bitmap indicating which of the following fields are set
4679 study_data opaque data from pcre_study()
4680 match_limit a way of specifying a limit on match() calls for a specific
4681 call to pcre_exec()
4682 callout_data data for callouts (see 49 below)
4683
4684 The flag bits are also defined in pcre.h, and are
4685
4686 PCRE_EXTRA_STUDY_DATA
4687 PCRE_EXTRA_MATCH_LIMIT
4688 PCRE_EXTRA_CALLOUT_DATA
4689
4690 The pcre_study() function now returns one of these new pcre_extra blocks, with
4691 the actual study data pointed to by the study_data field, and the
4692 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4693 before. That is, this change is entirely upwards-compatible and requires no
4694 change to existing code.
4695
4696 If you want to pass in additional data to pcre_exec(), you can either place it
4697 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4698 block.
4699
4700 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4701 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4702 times with different match limits, until it finds the minimum value needed for
4703 pcre_exec() to complete. The value is then output. This can be instructive; for
4704 most simple matches the number is quite small, but for pathological cases it
4705 gets very large very quickly.
4706
4707 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4708 returns the size of the data block pointed to by the study_data field in a
4709 pcre_extra block, that is, the value that was passed as the argument to
4710 pcre_malloc() when PCRE was getting memory in which to place the information
4711 created by pcre_study(). The fourth argument should point to a size_t variable.
4712 pcretest has been extended so that this information is shown after a successful
4713 pcre_study() call when information about the compiled regex is being displayed.
4714
4715 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4716 because what follows is always an absolute path. (Later: it turns out that this
4717 is more than cosmetic for MinGW, because it doesn't like empty path
4718 components.)
4719
4720 49. Some changes have been made to the callout feature (see 28 above):
4721
4722 (i) A callout function now has three choices for what it returns:
4723
4724 0 => success, carry on matching
4725 > 0 => failure at this point, but backtrack if possible
4726 < 0 => serious error, return this value from pcre_exec()
4727
4728 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4729 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4730 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4731 use by callout functions. It will never be used by PCRE itself.
4732
4733 (ii) The pcre_extra structure (see 45 above) has a void * field called
4734 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4735 pcre_callout_block structure has a field of the same name. The contents of
4736 the field passed in the pcre_extra structure are passed to the callout
4737 function in the corresponding field in the callout block. This makes it
4738 easier to use the same callout-containing regex from multiple threads. For
4739 testing, the pcretest program has a new data escape
4740
4741 \C*n pass the number n (may be negative) as callout_data
4742
4743 If the callout function in pcretest receives a non-zero value as
4744 callout_data, it returns that value.
4745
4746 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4747 there were some redundant $(CFLAGS) in commands that are now specified as
4748 $(LINK), which already includes $(CFLAGS).
4749
4750 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4751 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4752 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4753 one-byte characters throughout. Note that case-insensitive matching applies
4754 only to characters whose values are less than 256. PCRE doesn't support the
4755 notion of cases for higher-valued characters.
4756
4757 (i) A character class whose characters are all within 0-255 is handled as
4758 a bit map, and the map is inverted for negative classes. Previously, a
4759 character > 255 always failed to match such a class; however it should
4760 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4761
4762 (ii) A negated character class with a single character < 255 is coded as
4763 "not this character" (OP_NOT). This wasn't working properly when the test
4764 character was multibyte, either singly or repeated.
4765
4766 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4767 mode, for example: \x{100}{2,3}.
4768
4769 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4770 singly or repeated) now correctly test multibyte characters. However,
4771 PCRE doesn't recognize any characters with values greater than 255 as
4772 digits, spaces, or word characters. Such characters always match \D, \S,
4773 and \W, and never match \d, \s, or \w.
4774
4775 (v) Classes may now contain characters and character ranges with values
4776 greater than 255. For example: [ab\x{100}-\x{400}].
4777
4778 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4779 PCRE in UTF-8 mode.
4780
4781 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4782 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4783 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4784 value.)
4785
4786 53. The single man page has become too large. I have therefore split it up into
4787 a number of separate man pages. These also give rise to individual HTML pages;
4788 these are now put in a separate directory, and there is an index.html page that
4789 lists them all. Some hyperlinking between the pages has been installed.
4790
4791 54. Added convenience functions for handling named capturing parentheses.
4792
4793 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4794 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4795 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4796 are faulted.
4797
4798 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4799 calling configure. These values are used when compiling the dftables.c program
4800 which is run to generate the source of the default character tables. They
4801 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4802 you will need to set these values.
4803
4804 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4805
4806
4807 Version 3.9 02-Jan-02
4808 ---------------------
4809
4810 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4811
4812 2. If --disable-static was given, the building process failed when trying to
4813 build pcretest and pcregrep. (For some reason it was using libtool to compile
4814 them, which is not right, as they aren't part of the library.)
4815
4816
4817 Version 3.8 18-Dec-01
4818 ---------------------
4819
4820 1. The experimental UTF-8 code was completely screwed up. It was packing the
4821 bytes in the wrong order. How dumb can you get?
4822
4823
4824 Version 3.7 29-Oct-01
4825 ---------------------
4826
4827 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4828 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4829 this didn't happen under Solaris 8, where I normally test things.
4830
4831 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4832 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4833 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4834 this fix an hour or so after the initial 3.7 release.)
4835
4836
4837 Version 3.6 23-Oct-01
4838 ---------------------
4839
4840 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4841 offsets passed as NULL with zero offset count.
4842
4843 2. The config.guess and config.sub files had not been updated when I moved to
4844 the latest autoconf.
4845
4846
4847 Version 3.5 15-Aug-01
4848 ---------------------
4849
4850 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4851 had been forgotten.
4852
4853 2. By using declared but undefined structures, we can avoid using "void"
4854 definitions in pcre.h while keeping the internal definitions of the structures
4855 private.
4856
4857 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4858 user point of view, this means that both static and shared libraries are built
4859 by default, but this can be individually controlled. More of the work of
4860 handling this static/shared cases is now inside libtool instead of PCRE's make
4861 file.
4862
4863 4. The pcretest utility is now installed along with pcregrep because it is
4864 useful for users (to test regexs) and by doing this, it automatically gets
4865 relinked by libtool. The documentation has been turned into a man page, so
4866 there are now .1, .txt, and .html versions in /doc.
4867
4868 5. Upgrades to pcregrep:
4869 (i) Added long-form option names like gnu grep.
4870 (ii) Added --help to list all options with an explanatory phrase.
4871 (iii) Added -r, --recursive to recurse into sub-directories.
4872 (iv) Added -f, --file to read patterns from a file.
4873
4874 6. pcre_exec() was referring to its "code" argument before testing that
4875 argument for NULL (and giving an error if it was NULL).
4876
4877 7. Upgraded Makefile.in to allow for compiling in a different directory from
4878 the source directory.
4879
4880 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4881 options bits, the pointer it was passed was to an int instead of to an unsigned
4882 long int. This mattered only on 64-bit systems.
4883
4884 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4885 generated) instead of pcre.in, which it its source. Also made the same change
4886 in several of the .c files.
4887
4888 10. A new release of gcc defines printf() as a macro, which broke pcretest
4889 because it had an ifdef in the middle of a string argument for printf(). Fixed
4890 by using separate calls to printf().
4891
4892 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4893 script, to force use of CR or LF instead of \n in the source. On non-Unix
4894 systems, the value can be set in config.h.
4895
4896 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4897 absolute limit. Changed the text of the error message to make this clear, and
4898 likewise updated the man page.
4899
4900 13. The limit of 99 on the number of capturing subpatterns has been removed.
4901 The new limit is 65535, which I hope will not be a "real" limit.
4902
4903
4904 Version 3.4 22-Aug-00
4905 ---------------------
4906
4907 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4908
4909 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4910
4911
4912 Version 3.3 01-Aug-00
4913 ---------------------
4914
4915 1. If an octal character was given, but the value was greater than \377, it
4916 was not getting masked to the least significant bits, as documented. This could
4917 lead to crashes in some systems.
4918
4919 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4920 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4921
4922 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4923 These just pass their arguments on to (pcre_free)(), but they are provided
4924 because some uses of PCRE bind it to non-C systems that can call its functions,
4925 but cannot call free() or pcre_free() directly.
4926
4927 4. Add "make test" as a synonym for "make check". Corrected some comments in
4928 the Makefile.
4929
4930 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4931 Makefile.
4932
4933 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4934 command called pgrep for grepping around the active processes.
4935
4936 7. Added the beginnings of support for UTF-8 character strings.
4937
4938 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4939 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4940 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4941 out for the ar command.)
4942
4943
4944 Version 3.2 12-May-00
4945 ---------------------
4946
4947 This is purely a bug fixing release.
4948
4949 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4950 of ZA. This was just one example of several cases that could provoke this bug,
4951 which was introduced by change 9 of version 2.00. The code for breaking
4952 infinite loops after an iteration that matches an empty string was't working
4953 correctly.
4954
4955 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4956 when matched against abbab (for example). After matching an empty string, it
4957 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4958 caused it to match further down the string than it should.
4959
4960 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4961 was there because it doesn't seem to be needed, and it causes trouble on some
4962 systems, as it is not a Standard C header. It has been removed.
4963
4964 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4965 were reported on the Macintosh. The changes were from
4966
4967 while ((c = *(++ptr)) != 0 && c != '\n');
4968 to
4969 while ((c = *(++ptr)) != 0 && c != '\n') ;
4970
4971 Totally extraordinary, but if that's what it takes...
4972
4973 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4974 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4975 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4976 assumes the way PCRE uses memmove() (always moving upwards).
4977
4978 6. PCRE is being used in one environment where strchr() is not available. There
4979 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4980 faster code anyway.
4981
4982
4983 Version 3.1 09-Feb-00
4984 ---------------------
4985
4986 The only change in this release is the fixing of some bugs in Makefile.in for
4987 the "install" target:
4988
4989 (1) It was failing to install pcreposix.h.
4990
4991 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4992
4993
4994 Version 3.0 01-Feb-00
4995 ---------------------
4996
4997 1. Add support for the /+ modifier to perltest (to output $` like it does in
4998 pcretest).
4999
5000 2. Add support for the /g modifier to perltest.
5001
5002 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5003 matches null strings.
5004
5005 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5006 pattern. Perl treats empty patterns specially - it reuses the most recent
5007 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5008 effect.
5009
5010 5. The POSIX interface was broken in that it was just handing over the POSIX
5011 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5012 required a bigger vector, with some working space on the end. This means that
5013 the POSIX wrapper now has to get and free some memory, and copy the results.
5014
5015 6. Added some simple autoconf support, placing the test data and the
5016 documentation in separate directories, re-organizing some of the
5017 information files, and making it build pcre-config (a GNU standard). Also added
5018 libtool support for building PCRE as a shared library, which is now the
5019 default.
5020
5021 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5022 09 are not valid octal constants. Single digits will be used for minor values
5023 less than 10.
5024
5025 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5026 existing programs that set these in the POSIX interface can use PCRE without
5027 modification.
5028
5029 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5030 return all that pcre_info() returns, plus additional data. The pcre_info()
5031 function is retained for compatibility, but is considered to be obsolete.
5032
5033 10. Added experimental recursion feature (?R) to handle one common case that
5034 Perl 5.6 will be able to do with (?p{...}).
5035
5036 11. Added support for POSIX character classes like [:alpha:], which Perl is
5037 adopting.
5038
5039
5040 Version 2.08 31-Aug-99
5041 ----------------------
5042
5043 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5044 trying to match at the startoffset position, but instead was moving forward to
5045 the next newline as if a previous match had failed.
5046
5047 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5048 and could get into a loop if a null string was matched other than at the start
5049 of the subject.
5050
5051 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5052 be distinguished at compile time, and for completeness also added PCRE_DATE.
5053
5054 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5055 in GnuWin32 environments.
5056
5057
5058 Version 2.07 29-Jul-99
5059 ----------------------
5060
5061 1. The documentation is now supplied in plain text form and HTML as well as in
5062 the form of man page sources.
5063
5064 2. C++ compilers don't like assigning (void *) values to other pointer types.
5065 In particular this affects malloc(). Although there is no problem in Standard
5066 C, I've put in casts to keep C++ compilers happy.
5067
5068 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5069 should be (const char *).
5070
5071 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5072 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5073 However, I haven't made this a standard facility. The documentation doesn't
5074 mention it, and the Makefile doesn't support it.
5075
5076 5. The Makefile now contains an "install" target, with editable destinations at
5077 the top of the file. The pcretest program is not installed.
5078
5079 6. pgrep -V now gives the PCRE version number and date.
5080
5081 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5082 causing the entire string to be ignored, instead of just the last character.
5083
5084 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5085 non-matching string, it can take a very, very long time, even for strings of
5086 quite modest length, because of the nested recursion. PCRE now does better in
5087 some of these cases. It does this by remembering the last required literal
5088 character in the pattern, and pre-searching the subject to ensure it is present
5089 before running the real match. In other words, it applies a heuristic to detect
5090 some types of certain failure quickly, and in the above example, if presented
5091 with a string that has no trailing " it gives "no match" very quickly.
5092
5093 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5094 other alternatives are tried instead.
5095
5096
5097 Version 2.06 09-Jun-99
5098 ----------------------
5099
5100 1. Change pcretest's output for amount of store used to show just the code
5101 space, because the remainder (the data block) varies in size between 32-bit and
5102 64-bit systems.
5103
5104 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5105 start matching at. This allows lookbehinds to work when searching for multiple
5106 occurrences in a string.
5107
5108 3. Added additional options to pcretest for testing multiple occurrences:
5109
5110 /+ outputs the rest of the string that follows a match
5111 /g loops for multiple occurrences, using the new startoffset argument
5112 /G loops for multiple occurrences by passing an incremented pointer
5113
5114 4. PCRE wasn't doing the "first character" optimization for patterns starting
5115 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5116 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5117 the letter 'x'. On long subject strings, this gives a significant speed-up.
5118
5119
5120 Version 2.05 21-Apr-99
5121 ----------------------
5122
5123 1. Changed the type of magic_number from int to long int so that it works
5124 properly on 16-bit systems.
5125
5126 2. Fixed a bug which caused patterns starting with .* not to work correctly
5127 when the subject string contained newline characters. PCRE was assuming
5128 anchoring for such patterns in all cases, which is not correct because .* will
5129 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5130 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5131 must be retried after every newline in the subject.
5132
5133
5134 Version 2.04 18-Feb-99
5135 ----------------------
5136
5137 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5138 computation of the store needed to hold the pattern was incorrect (too large).
5139 If such patterns were nested a few deep, this could multiply and become a real
5140 problem.
5141
5142 2. Added /M option to pcretest to show the memory requirement of a specific
5143 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5144
5145 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5146 compiled in such a way that the backtracking after subsequent failure was
5147 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5148 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5149
5150
5151 Version 2.03 02-Feb-99
5152 ----------------------
5153
5154 1. Fixed typo and small mistake in man page.
5155
5156 2. Added 4th condition (GPL supersedes if conflict) and created separate
5157 LICENCE file containing the conditions.
5158
5159 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5160 Perl, that is the internal \ allows the delimiter to be included in the
5161 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5162 the final delimiter, add \ to the end of the pattern (to test the error).
5163
5164 4. Added the convenience functions for extracting substrings after a successful
5165 match. Updated pcretest to make it able to test these functions.
5166
5167
5168 Version 2.02 14-Jan-99
5169 ----------------------
5170
5171 1. Initialized the working variables associated with each extraction so that
5172 their saving and restoring doesn't refer to uninitialized store.
5173
5174 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5175 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5176 fix the problem.
5177
5178 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5179 calls, and wasn't printing the correct value for compiling calls. Increased the
5180 default value of LOOPREPEAT, and the number of significant figures in the
5181 times.
5182
5183 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5184
5185 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5186 a building problem on Windows NT with a FAT file system.
5187
5188
5189 Version 2.01 21-Oct-98
5190 ----------------------
5191
5192 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5193 to character tables built by pcre_maketables() in the current locale. If NULL
5194 is passed, the default tables are used.
5195
5196
5197 Version 2.00 24-Sep-98
5198 ----------------------
5199
5200 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5201 it any more.
5202
5203 2. Allow quantification of (?>) groups, and make it work correctly.
5204
5205 3. The first character computation wasn't working for (?>) groups.
5206
5207 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5208 end of the subject) and add 5.005's \z, which really does match only at the
5209 very end of the subject.
5210
5211 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5212
5213 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5214 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5215 localized options. All options to pcre_study() were also removed.
5216
5217 7. Add other new features from 5.005:
5218
5219 $(?<= positive lookbehind
5220 $(?<! negative lookbehind
5221 (?imsx-imsx) added the unsetting capability
5222 such a setting is global if at outer level; local otherwise
5223 (?imsx-imsx:) non-capturing groups with option setting
5224 (?(cond)re|re) conditional pattern matching
5225
5226 A backreference to itself in a repeated group matches the previous
5227 captured string.
5228
5229 8. General tidying up of studying (both automatic and via "study")
5230 consequential on the addition of new assertions.
5231
5232 9. As in 5.005, unlimited repeated groups that could match an empty substring
5233 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5234 runtime if any iteration does actually match an empty substring.
5235
5236 10. Include the RunTest script in the distribution.
5237
5238 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5239 discrepancies, some of which were old and were also with respect to 5.004. They
5240 have now been fixed.
5241
5242
5243 Version 1.09 28-Apr-98
5244 ----------------------
5245
5246 1. A negated single character class followed by a quantifier with a minimum
5247 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5248 program crashes, or just wrong answers. This did not apply to negated classes
5249 containing more than one character, or to minima other than one.
5250
5251
5252 Version 1.08 27-Mar-98
5253 ----------------------
5254
5255 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5256
5257 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5258 latter must appear before anything that relies on it in the pattern.
5259
5260
5261 Version 1.07 16-Feb-98
5262 ----------------------
5263
5264 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5265 repeat of a potentially empty string).
5266
5267
5268 Version 1.06 23-Jan-98
5269 ----------------------
5270
5271 1. Added Markus Oberhumer's little patches for C++.
5272
5273 2. Literal strings longer than 255 characters were broken.
5274
5275
5276 Version 1.05 23-Dec-97
5277 ----------------------
5278
5279 1. Negated character classes containing more than one character were failing if
5280 PCRE_CASELESS was set at run time.
5281
5282
5283 Version 1.04 19-Dec-97
5284 ----------------------
5285
5286 1. Corrected the man page, where some "const" qualifiers had been omitted.
5287
5288 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5289 input syntax.
5290
5291 3. Fixed memory leak which occurred when a regex with back references was
5292 matched with an offsets vector that wasn't big enough. The temporary memory
5293 that is used in this case wasn't being freed if the match failed.
5294
5295 4. Tidied pcretest to ensure it frees memory that it gets.
5296
5297 5. Temporary memory was being obtained in the case where the passed offsets
5298 vector was exactly big enough.
5299
5300 6. Corrected definition of offsetof() from change 5 below.
5301
5302 7. I had screwed up change 6 below and broken the rules for the use of
5303 setjmp(). Now fixed.
5304
5305
5306 Version 1.03 18-Dec-97
5307 ----------------------
5308
5309 1. A erroneous regex with a missing opening parenthesis was correctly
5310 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5311 on some systems.
5312
5313 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5314 it was reported that one broken compiler failed on the former because "code" is
5315 also an independent variable.
5316
5317 3. The erroneous regex a[]b caused an array overrun reference.
5318
5319 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5320 fail on data ending with that character. (It was going on too far, and checking
5321 the next character, typically a binary zero.) This was specific to the
5322 optimized code for single-character negative classes.
5323
5324 5. Added a contributed patch from the TIN world which does the following:
5325
5326 + Add an undef for memmove, in case the the system defines a macro for it.
5327
5328 + Add a definition of offsetof(), in case there isn't one. (I don't know
5329 the reason behind this - offsetof() is part of the ANSI standard - but
5330 it does no harm).
5331
5332 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5333 most of the places where whitespace preceded '#'. I have given up and
5334 allowed the remaining 2 cases to be at the margin.
5335
5336 + Rename some variables in pcre to eliminate shadowing. This seems very
5337 pedantic, but does no harm, of course.
5338
5339 6. Moved the call to setjmp() into its own function, to get rid of warnings
5340 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5341
5342 7. Constructs such as \d{8,} were compiling into the equivalent of
5343 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5344 outcome, but in this particular case used more store than had been allocated,
5345 which caused the bug to be discovered because it threw up an internal error.
5346
5347 8. The debugging code in both pcre and pcretest for outputting the compiled
5348 form of a regex was going wrong in the case of back references followed by
5349 curly-bracketed repeats.
5350
5351
5352 Version 1.02 12-Dec-97
5353 ----------------------
5354
5355 1. Typos in pcre.3 and comments in the source fixed.
5356
5357 2. Applied a contributed patch to get rid of places where it used to remove
5358 'const' from variables, and fixed some signed/unsigned and uninitialized
5359 variable warnings.
5360
5361 3. Added the "runtest" target to Makefile.
5362
5363 4. Set default compiler flag to -O2 rather than just -O.
5364
5365
5366 Version 1.01 19-Nov-97
5367 ----------------------
5368
5369 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5370 like /([ab]*)*/, that is, for classes with more than one character in them.
5371
5372 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5373 as /((?>a*))*/ (a PCRE_EXTRA facility).
5374
5375
5376 Version 1.00 18-Nov-97
5377 ----------------------
5378
5379 1. Added compile-time macros to support systems such as SunOS4 which don't have
5380 memmove() or strerror() but have other things that can be used instead.
5381
5382 2. Arranged that "make clean" removes the executables.
5383
5384
5385 Version 0.99 27-Oct-97
5386 ----------------------
5387
5388 1. Fixed bug in code for optimizing classes with only one character. It was
5389 initializing a 32-byte map regardless, which could cause it to run off the end
5390 of the memory it had got.
5391
5392 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5393
5394
5395 Version 0.98 22-Oct-97
5396 ----------------------
5397
5398 1. Fixed bug in code for handling temporary memory usage when there are more
5399 back references than supplied space in the ovector. This could cause segfaults.
5400
5401
5402 Version 0.97 21-Oct-97
5403 ----------------------
5404
5405 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5406
5407 2. Optimized negated single characters not to use a bit map.
5408
5409 3. Brought error texts together as macro definitions; clarified some of them;
5410 fixed one that was wrong - it said "range out of order" when it meant "invalid
5411 escape sequence".
5412
5413 4. Changed some char * arguments to const char *.
5414
5415 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5416
5417 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5418 pcretest.
5419
5420
5421 Version 0.96 16-Oct-97
5422 ----------------------
5423
5424 1. Added a simple "pgrep" utility to the distribution.
5425
5426 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5427 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5428 where "ddd" means "one or more decimal digits".
5429
5430 3. Fixed serious bug. If a pattern had a back reference, but the call to
5431 pcre_exec() didn't supply a large enough ovector to record the related
5432 identifying subpattern, the match always failed. PCRE now remembers the number
5433 of the largest back reference, and gets some temporary memory in which to save
5434 the offsets during matching if necessary, in order to ensure that
5435 backreferences always work.
5436
5437 4. Increased the compatibility with Perl in a number of ways:
5438
5439 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5440 to request this handling. The option can be set at compile or exec time.
5441
5442 (b) $ matches before a terminating newline by default; an option
5443 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5444 mode). The option can be set at compile or exec time.
5445
5446 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5447 the same as Perl's. If the decimal number it represents is less than 10
5448 or there aren't that many previous left capturing parentheses, an octal
5449 escape is read. Inside a character class, it's always an octal escape,
5450 even if it is a single digit.
5451
5452 (d) An escaped but undefined alphabetic character is taken as a literal,
5453 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5454 escapes.
5455
5456 (e) {0} is now permitted. (The previous item is removed from the compiled
5457 pattern).
5458
5459 5. Changed all the names of code files so that the basic parts are no longer
5460 than 10 characters, and abolished the teeny "globals.c" file.
5461
5462 6. Changed the handling of character classes; they are now done with a 32-byte
5463 bit map always.
5464
5465 7. Added the -d and /D options to pcretest to make it possible to look at the
5466 internals of compilation without having to recompile pcre.
5467
5468
5469 Version 0.95 23-Sep-97
5470 ----------------------
5471
5472 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5473 \x20 at the start of a run of normal characters. These were being treated as
5474 real characters, instead of the source characters being re-checked.
5475
5476
5477 Version 0.94 18-Sep-97
5478 ----------------------
5479
5480 1. The functions are now thread-safe, with the caveat that the global variables
5481 containing pointers to malloc() and free() or alternative functions are the
5482 same for all threads.
5483
5484 2. Get pcre_study() to generate a bitmap of initial characters for non-
5485 anchored patterns when this is possible, and use it if passed to pcre_exec().
5486
5487
5488 Version 0.93 15-Sep-97
5489 ----------------------
5490
5491 1. /(b)|(:+)/ was computing an incorrect first character.
5492
5493 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5494 but not actually doing anything yet.
5495
5496 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5497 as Perl does (e.g. [-az] or [az-]).
5498
5499 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5500 all possible positions.
5501
5502 5. Split up into different modules to avoid including unneeded functions in a
5503 compiled binary. However, compile and exec are still in one module. The "study"
5504 function is split off.
5505
5506 6. The character tables are now in a separate module whose source is generated
5507 by an auxiliary program - but can then be edited by hand if required. There are
5508 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5509 toupper() in the code.
5510
5511 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5512 make them global. Abolish the function for setting them, as the caller can now
5513 set them directly.
5514
5515
5516 Version 0.92 11-Sep-97
5517 ----------------------
5518
5519 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5520 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5521
5522 2. Caseless matching was not working in character classes if the characters in
5523 the pattern were in upper case.
5524
5525 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5526
5527 4. Make PCRE_ANCHORED public and accept as a compile option.
5528
5529 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5530 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5531 pass them.
5532
5533 6. Give an error if bad option bits passed at compile or run time.
5534
5535 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5536 pcretest to cause it to pass that flag.
5537
5538 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5539 options, and the first character, if set.
5540
5541 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5542
5543
5544 Version 0.91 10-Sep-97
5545 ----------------------
5546
5547 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5548 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5549
5550 2. PCRE was looping on encountering an indefinitely repeated back reference to
5551 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5552 Perl does - treats the match as successful.
5553
5554 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5