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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1486 - (show annotations)
Wed Jun 18 16:48:57 2014 UTC (5 years, 3 months ago) by ph10
File size: 267124 byte(s)
Fix not including VT in starting characters for \s.
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
73 Version 8.35 04-April-2014
74 --------------------------
75
76 1. A new flag is set, when property checks are present in an XCLASS.
77 When this flag is not set, PCRE can perform certain optimizations
78 such as studying these XCLASS-es.
79
80 2. The auto-possessification of character sets were improved: a normal
81 and an extended character set can be compared now. Furthermore
82 the JIT compiler optimizes more character set checks.
83
84 3. Got rid of some compiler warnings for potentially uninitialized variables
85 that show up only when compiled with -O2.
86
87 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
88 of a match later then the end of the match. The pcretest program was not
89 handling the case sensibly - it was outputting from the start to the next
90 binary zero. It now reports this situation in a message, and outputs the
91 text from the end to the start.
92
93 5. Fast forward search is improved in JIT. Instead of the first three
94 characters, any three characters with fixed position can be searched.
95 Search order: first, last, middle.
96
97 6. Improve character range checks in JIT. Characters are read by an inprecise
98 function now, which returns with an unknown value if the character code is
99 above a certain treshold (e.g: 256). The only limitation is that the value
100 must be bigger than the treshold as well. This function is useful, when
101 the characters above the treshold are handled in the same way.
102
103 7. The macros whose names start with RAWUCHAR are placeholders for a future
104 mode in which only the bottom 21 bits of 32-bit data items are used. To
105 make this more memorable for those maintaining the code, the names have
106 been changed to start with UCHAR21, and an extensive comment has been added
107 to their definition.
108
109 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
110 to the export list in Makefile.am (they were accidentally omitted from the
111 8.34 tarball).
112
113 9. The informational output from pcretest used the phrase "starting byte set"
114 which is inappropriate for the 16-bit and 32-bit libraries. As the output
115 for "first char" and "need char" really means "non-UTF-char", I've changed
116 "byte" to "char", and slightly reworded the output. The documentation about
117 these values has also been (I hope) clarified.
118
119 10. Another JIT related optimization: use table jumps for selecting the correct
120 backtracking path, when more than four alternatives are present inside a
121 bracket.
122
123 11. Empty match is not possible, when the minimum length is greater than zero,
124 and there is no \K in the pattern. JIT should avoid empty match checks in
125 such cases.
126
127 12. In a caseless character class with UCP support, when a character with more
128 than one alternative case was not the first character of a range, not all
129 the alternative cases were added to the class. For example, s and \x{17f}
130 are both alternative cases for S: the class [RST] was handled correctly,
131 but [R-T] was not.
132
133 13. The configure.ac file always checked for pthread support when JIT was
134 enabled. This is not used in Windows, so I have put this test inside a
135 check for the presence of windows.h (which was already tested for).
136
137 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
138 The algorithm provides a way to skip certain starting offsets, and usually
139 faster than linear prefix searches.
140
141 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
142 as for 'fr_FR' and 'french'. For some reason, however, it then used the
143 Windows-specific input and output files, which have 'french' screwed in.
144 So this could never have worked. One of the problems with locales is that
145 they aren't always the same. I have now updated RunTest so that it checks
146 the output of the locale test (test 3) against three different output
147 files, and it allows the test to pass if any one of them matches. With luck
148 this should make the test pass on some versions of Solaris where it was
149 failing. Because of the uncertainty, the script did not used to stop if
150 test 3 failed; it now does. If further versions of a French locale ever
151 come to light, they can now easily be added.
152
153 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
154 there was a message during ./configure, but it did not stop. This now
155 provokes an error. The invalid example in README has been corrected.
156 If a value less than the minimum is given, the minimum value has always
157 been used, but now a warning is given.
158
159 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
160 was a bug in the test system, which is now fixed. Also, the list of various
161 configurations that are tested for each release did not have one with both
162 16/32 bits and --enable-bar-anycrlf. It now does.
163
164 18. pcretest was missing "-C bsr" for displaying the \R default setting.
165
166 19. Little endian PowerPC systems are supported now by the JIT compiler.
167
168 20. The fast forward newline mechanism could enter to an infinite loop on
169 certain invalid UTF-8 input. Although we don't support these cases
170 this issue can be fixed by a performance optimization.
171
172 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
173 not take account if existing stack usage. There is now a new global
174 variable called pcre_stack_guard that can be set to point to an external
175 function to check stack availability. It is called at the start of
176 processing every parenthesized group.
177
178 22. A typo in the code meant that in ungreedy mode the max/min qualifier
179 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
180 not match "ab".
181
182 23. When UTF was disabled, the JIT program reported some incorrect compile
183 errors. These messages are silenced now.
184
185 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
186 compiler.
187
188 25. Change all the temporary files used in RunGrepTest to be different to those
189 used by RunTest so that the tests can be run simultaneously, for example by
190 "make -j check".
191
192
193 Version 8.34 15-December-2013
194 -----------------------------
195
196 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
197 executable memory. Patch inspired by Carsten Klein.
198
199 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
200 this macro is never tested and has no effect, because the work to support
201 coverage involves only compiling and linking options and special targets in
202 the Makefile. The comment in config.h implied that defining the macro would
203 enable coverage support, which is totally false. There was also support for
204 setting this macro in the CMake files (my fault, I just copied it from
205 configure). SUPPORT_GCOV has now been removed.
206
207 3. Make a small performance improvement in strlen16() and strlen32() in
208 pcretest.
209
210 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
211 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
212 these cases). There was also one in pcretest.c.
213
214 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
215
216 6. In UTF mode, the code for checking whether a group could match an empty
217 string (which is used for indefinitely repeated groups to allow for
218 breaking an infinite loop) was broken when the group contained a repeated
219 negated single-character class with a character that occupied more than one
220 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
221 UTF-8 mode). The effect was undefined: the group might or might not be
222 deemed as matching an empty string, or the program might have crashed.
223
224 7. The code for checking whether a group could match an empty string was not
225 recognizing that \h, \H, \v, \V, and \R must match a character.
226
227 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
228 an empty string. If it can, pcretest shows this in its information output.
229
230 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
231 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
232 matched by pcre_exec() without using JIT:
233
234 (a) If the rest of the pattern did not match after a maximal run of
235 grapheme clusters, the code for backing up to try with fewer of them
236 did not always back up over a full grapheme when characters that do not
237 have the modifier quality were involved, e.g. Hangul syllables.
238
239 (b) If the match point in a subject started with modifier character, and
240 there was no match, the code could incorrectly back up beyond the match
241 point, and potentially beyond the first character in the subject,
242 leading to a segfault or an incorrect match result.
243
244 10. A conditional group with an assertion condition could lead to PCRE
245 recording an incorrect first data item for a match if no other first data
246 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
247 first data item, and therefore matched "ca" after "c" instead of at the
248 start.
249
250 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
251 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
252
253 12. The source of pcregrep now includes z/OS-specific code so that it can be
254 compiled for z/OS as part of the special z/OS distribution.
255
256 13. Added the -T and -TM options to pcretest.
257
258 14. The code in pcre_compile.c for creating the table of named capturing groups
259 has been refactored. Instead of creating the table dynamically during the
260 actual compiling pass, the information is remembered during the pre-compile
261 pass (on the stack unless there are more than 20 named groups, in which
262 case malloc() is used) and the whole table is created before the actual
263 compile happens. This has simplified the code (it is now nearly 150 lines
264 shorter) and prepared the way for better handling of references to groups
265 with duplicate names.
266
267 15. A back reference to a named subpattern when there is more than one of the
268 same name now checks them in the order in which they appear in the pattern.
269 The first one that is set is used for the reference. Previously only the
270 first one was inspected. This change makes PCRE more compatible with Perl.
271
272 16. Unicode character properties were updated from Unicode 6.3.0.
273
274 17. The compile-time code for auto-possessification has been refactored, based
275 on a patch by Zoltan Herczeg. It now happens after instead of during
276 compilation. The code is cleaner, and more cases are handled. The option
277 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
278 options in pcretest are provided to set it. It can also be set by
279 (*NO_AUTO_POSSESS) at the start of a pattern.
280
281 18. The character VT has been added to the default ("C" locale) set of
282 characters that match \s and are generally treated as white space,
283 following this same change in Perl 5.18. There is now no difference between
284 "Perl space" and "POSIX space". Whether VT is treated as white space in
285 other locales depends on the locale.
286
287 19. The code for checking named groups as conditions, either for being set or
288 for being recursed, has been refactored (this is related to 14 and 15
289 above). Processing unduplicated named groups should now be as fast at
290 numerical groups, and processing duplicated groups should be faster than
291 before.
292
293 20. Two patches to the CMake build system, by Alexander Barkov:
294
295 (1) Replace the "source" command by "." in CMakeLists.txt because
296 "source" is a bash-ism.
297
298 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
299 without these the CMake build does not work on Solaris.
300
301 21. Perl has changed its handling of \8 and \9. If there is no previously
302 encountered capturing group of those numbers, they are treated as the
303 literal characters 8 and 9 instead of a binary zero followed by the
304 literals. PCRE now does the same.
305
306 22. Following Perl, added \o{} to specify codepoints in octal, making it
307 possible to specify values greater than 0777 and also making them
308 unambiguous.
309
310 23. Perl now gives an error for missing closing braces after \x{... instead of
311 treating the string as literal. PCRE now does the same.
312
313 24. RunTest used to grumble if an inappropriate test was selected explicitly,
314 but just skip it when running all tests. This make it awkward to run ranges
315 of tests when one of them was inappropriate. Now it just skips any
316 inappropriate tests, as it always did when running all tests.
317
318 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
319 character types such as \d or \w, too many callouts were inserted, and the
320 data that they returned was rubbish.
321
322 26. In UCP mode, \s was not matching two of the characters that Perl matches,
323 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
324 were matched by \h. The code has now been refactored so that the lists of
325 the horizontal and vertical whitespace characters used for \h and \v (which
326 are defined only in one place) are now also used for \s.
327
328 27. Add JIT support for the 64 bit TileGX architecture.
329 Patch by Jiong Wang (Tilera Corporation).
330
331 28. Possessive quantifiers for classes (both explicit and automatically
332 generated) now use special opcodes instead of wrapping in ONCE brackets.
333
334 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
335 (because it's meaningless), this was not happening when PCRE_CASELESS was
336 set. Not wrong, but inefficient.
337
338 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
339 properties for \w, \d, etc) is present in a test regex. Otherwise if the
340 test contains no characters greater than 255, Perl doesn't realise it
341 should be using Unicode semantics.
342
343 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
344 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
345 does in Unicode mode.
346
347 32. Added the "forbid" facility to pcretest so that putting tests into the
348 wrong test files can sometimes be quickly detected.
349
350 33. There is now a limit (default 250) on the depth of nesting of parentheses.
351 This limit is imposed to control the amount of system stack used at compile
352 time. It can be changed at build time by --with-parens-nest-limit=xxx or
353 the equivalent in CMake.
354
355 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
356 errors. Perl warns for these when in warning mode, but PCRE has no facility
357 for giving warnings.
358
359 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
360 However, this was not working for (?!) because it is optimized to (*FAIL),
361 for which PCRE does not allow quantifiers. The optimization is now disabled
362 when a quantifier follows (?!). I can't see any use for this, but it makes
363 things uniform.
364
365 36. Perl no longer allows group names to start with digits, so I have made this
366 change also in PCRE. It simplifies the code a bit.
367
368 37. In extended mode, Perl ignores spaces before a + that indicates a
369 possessive quantifier. PCRE allowed a space before the quantifier, but not
370 before the possessive +. It now does.
371
372 38. The use of \K (reset reported match start) within a repeated possessive
373 group such as (a\Kb)*+ was not working.
374
375 40. Document that the same character tables must be used at compile time and
376 run time, and that the facility to pass tables to pcre_exec() and
377 pcre_dfa_exec() is for use only with saved/restored patterns.
378
379 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
380 features for Builds with MSVC:
381
382 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
383 builds). The .rc files can be used to set FileDescription and many other
384 attributes.
385
386 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
387 This allows higher-level build scripts which want .pdb files to avoid
388 hard-coding the exact files needed."
389
390 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
391 mean "start of word" and "end of word", respectively, as a transition aid.
392
393 43. A minimizing repeat of a class containing codepoints greater than 255 in
394 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
395 compiled to use the heap for recursion.
396
397 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
398 is configured.
399
400
401 Version 8.33 28-May-2013
402 ------------------------
403
404 1. Added 'U' to some constants that are compared to unsigned integers, to
405 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
406 variables that are added to signed variables, to ensure the result is
407 signed and can be negated.
408
409 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
410 CMake config files.
411
412 3. Revise the creation of config.h.generic so that all boolean macros are
413 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
414 overriding via -D on the command line possible.
415
416 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
417 to unsigned int is reported to make a quite noticeable speed difference in
418 a specific Windows environment. Testing on Linux did also appear to show
419 some benefit (and it is clearly not harmful). Also fixed the definition of
420 Xop which should be unsigned.
421
422 5. Related to (4), changing the definition of the intermediate variable cc
423 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
424 improvements.
425
426 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
427 unnecessary spaces.
428
429 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
430 and later.
431
432 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
433
434 9. Optimizing fast_forward_start_bits in JIT.
435
436 10. Adding support for callouts in JIT, and fixing some issues revealed
437 during this work. Namely:
438
439 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
440
441 (b) Minimum length was not checked before the matching is started.
442
443 11. The value of capture_last that is passed to callouts was incorrect in some
444 cases when there was a capture on one path that was subsequently abandoned
445 after a backtrack. Also, the capture_last value is now reset after a
446 recursion, since all captures are also reset in this case.
447
448 12. The interpreter no longer returns the "too many substrings" error in the
449 case when an overflowing capture is in a branch that is subsequently
450 abandoned after a backtrack.
451
452 13. In the pathological case when an offset vector of size 2 is used, pcretest
453 now prints out the matched string after a yield of 0 or 1.
454
455 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
456 Only supported by the JIT compiler at the moment.
457
458 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
459
460 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
461 offset of the starting point of the matching process, provided the offsets
462 vector is large enough.
463
464 17. The \A escape now records a lookbehind value of 1, though its execution
465 does not actually inspect the previous character. This is to ensure that,
466 in partial multi-segment matching, at least one character from the old
467 segment is retained when a new segment is processed. Otherwise, if there
468 are no lookbehinds in the pattern, \A might match incorrectly at the start
469 of a new segment.
470
471 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
472
473 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
474 gives some modest performance improvement in 8-bit mode.
475
476 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
477 be expressed in certain programming languages using Universal Character
478 Names.
479
480 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
481 which points out that "non characters" are not "characters that may not
482 appear in Unicode strings" but rather "characters that are reserved for
483 internal use and have only local meaning".
484
485 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
486 there was a conditional group that depended on an assertion, if the
487 assertion was false, the callout that immediately followed the alternation
488 in the condition was skipped when pcre_exec() was used for matching.
489
490 23. Allow an explicit callout to be inserted before an assertion that is the
491 condition for a conditional group, for compatibility with automatic
492 callouts, which always insert a callout at this point.
493
494 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
495 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
496
497 25. (*PRUNE) is now supported by the JIT compiler.
498
499 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
500
501 27. Fix the case where there are two or more SKIPs with arguments that may be
502 ignored.
503
504 28. (*SKIP) is now supported by the JIT compiler.
505
506 29. (*THEN) is now supported by the JIT compiler.
507
508 30. Update RunTest with additional test selector options.
509
510 31. The way PCRE handles backtracking verbs has been changed in two ways.
511
512 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
513 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
514 backtracking. In some cases this makes it more Perl-compatible, but Perl's
515 rather obscure rules do not always do the same thing.
516
517 (2) Previously, backtracking verbs were confined within assertions. This is
518 no longer the case for positive assertions, except for (*ACCEPT). Again,
519 this sometimes improves Perl compatibility, and sometimes does not.
520
521 32. A number of tests that were in test 2 because Perl did things differently
522 have been moved to test 1, because either Perl or PCRE has changed, and
523 these tests are now compatible.
524
525 32. Backtracking control verbs are now handled in the same way in JIT and
526 interpreter.
527
528 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
529 contained a forward subroutine reference caused a compile error.
530
531 34. Auto-detect and optimize limited repetitions in JIT.
532
533 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
534 blocking (*UTF) etc.
535
536 36. In the interpreter, maximizing pattern repetitions for characters and
537 character types now use tail recursion, which reduces stack usage.
538
539 37. The value of the max lookbehind was not correctly preserved if a compiled
540 and saved regex was reloaded on a host of different endianness.
541
542 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
543 of the compiled pattern block, expand the flags field from 16 to 32 bits
544 because it was almost full.
545
546 39. Try madvise first before posix_madvise.
547
548 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
549 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
550 original change locked out all matches of empty strings. This has been
551 changed so that one match of an empty string per line is recognized.
552 Subsequent searches on the same line (for colouring or for --only-matching,
553 for example) do not recognize empty strings.
554
555 41. Applied a user patch to fix a number of spelling mistakes in comments.
556
557 42. Data lines longer than 65536 caused pcretest to crash.
558
559 43. Clarified the data type for length and startoffset arguments for pcre_exec
560 and pcre_dfa_exec in the function-specific man pages, where they were
561 explicitly stated to be in bytes, never having been updated. I also added
562 some clarification to the pcreapi man page.
563
564 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
565 a segmentation fault.
566
567
568 Version 8.32 30-November-2012
569 -----------------------------
570
571 1. Improved JIT compiler optimizations for first character search and single
572 character iterators.
573
574 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
575 Patch by Daniel Richard G.
576
577 3. Single character iterator optimizations in the JIT compiler.
578
579 4. Improved JIT compiler optimizations for character ranges.
580
581 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
582 Reported by Giuseppe D'Angelo.
583
584 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
585 of a line, was being set incorrectly in cases where .* appeared inside
586 atomic brackets at the start of a pattern, or where there was a subsequent
587 *PRUNE or *SKIP.
588
589 7. Improved instruction cache flush for POWER/PowerPC.
590 Patch by Daniel Richard G.
591
592 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
593 grep:
594
595 (a) There is now no limit to the number of patterns to be matched.
596
597 (b) An error is given if a pattern is too long.
598
599 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
600 are now supported.
601
602 (d) --exclude-from and --include-from (multiple use) have been added.
603
604 (e) Exclusions and inclusions now apply to all files and directories, not
605 just to those obtained from scanning a directory recursively.
606
607 (f) Multiple uses of -f and --file-list are now supported.
608
609 (g) In a Windows environment, the default for -d has been changed from
610 "read" (the GNU grep default) to "skip", because otherwise the presence
611 of a directory in the file list provokes an error.
612
613 (h) The documentation has been revised and clarified in places.
614
615 9. Improve the matching speed of capturing brackets.
616
617 10. Changed the meaning of \X so that it now matches a Unicode extended
618 grapheme cluster.
619
620 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
621 out POSIX threads when JIT support is configured.
622
623 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
624
625 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
626 structure could go wrong in environments where size_t is not the same size
627 as int.
628
629 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
630 set.
631
632 15. The EBCDIC support had decayed; later updates to the code had included
633 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
634 general tidy up of EBCDIC-related issues, and the documentation was also
635 not quite right. There is now a test that can be run on ASCII systems to
636 check some of the EBCDIC-related things (but is it not a full test).
637
638 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
639 in a small tidy to the code.
640
641 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
642
643 18. If the --only-matching (-o) option in pcregrep is specified multiple
644 times, each one causes appropriate output. For example, -o1 -o2 outputs the
645 substrings matched by the 1st and 2nd capturing parentheses. A separating
646 string can be specified by --om-separator (default empty).
647
648 19. Improving the first n character searches.
649
650 20. Turn case lists for horizontal and vertical white space into macros so that
651 they are defined only once.
652
653 21. This set of changes together give more compatible Unicode case-folding
654 behaviour for characters that have more than one other case when UCP
655 support is available.
656
657 (a) The Unicode property table now has offsets into a new table of sets of
658 three or more characters that are case-equivalent. The MultiStage2.py
659 script that generates these tables (the pcre_ucd.c file) now scans
660 CaseFolding.txt instead of UnicodeData.txt for character case
661 information.
662
663 (b) The code for adding characters or ranges of characters to a character
664 class has been abstracted into a generalized function that also handles
665 case-independence. In UTF-mode with UCP support, this uses the new data
666 to handle characters with more than one other case.
667
668 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
669 whose other case is greater than 256 are now correctly matched
670 caselessly. Previously, the high codepoint matched the low one, but not
671 vice versa.
672
673 (d) The processing of \h, \H, \v, and \ in character classes now makes use
674 of the new class addition function, using character lists defined as
675 macros alongside the case definitions of 20 above.
676
677 (e) Caseless back references now work with characters that have more than
678 one other case.
679
680 (f) General caseless matching of characters with more than one other case
681 is supported.
682
683 22. Unicode character properties were updated from Unicode 6.2.0
684
685 23. Improved CMake support under Windows. Patch by Daniel Richard G.
686
687 24. Add support for 32-bit character strings, and UTF-32
688
689 25. Major JIT compiler update (code refactoring and bugfixing).
690 Experimental Sparc 32 support is added.
691
692 26. Applied a modified version of Daniel Richard G's patch to create
693 pcre.h.generic and config.h.generic by "make" instead of in the
694 PrepareRelease script.
695
696 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
697 pcre_compile.c when checking for a zero character.
698
699 28. Introducing a native interface for JIT. Through this interface, the compiled
700 machine code can be directly executed. The purpose of this interface is to
701 provide fast pattern matching, so several sanity checks are not performed.
702 However, feature tests are still performed. The new interface provides
703 1.4x speedup compared to the old one.
704
705 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
706 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
707 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
708
709 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
710 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
711 the "old" RFC 2279). Instead, it ended up passing a negative length to
712 pcre_exec().
713
714 31. Add support for GCC's visibility feature to hide internal functions.
715
716 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
717 "unknown -C option" after outputting 0 or 1.
718
719 33. There is now support for generating a code coverage report for the test
720 suite in environments where gcc is the compiler and lcov is installed. This
721 is mainly for the benefit of the developers.
722
723 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
724 unaddressable using valgrind annotations, allowing valgrind to detect
725 invalid memory accesses. This is mainly for the benefit of the developers.
726
727 25. (*UTF) can now be used to start a pattern in any of the three libraries.
728
729 26. Give configure error if --enable-cpp but no C++ compiler found.
730
731
732 Version 8.31 06-July-2012
733 -------------------------
734
735 1. Fixing a wrong JIT test case and some compiler warnings.
736
737 2. Removed a bashism from the RunTest script.
738
739 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
740 to unsigned type, result still unsigned" that was given by an MS compiler
741 on encountering the code "-sizeof(xxx)".
742
743 4. Partial matching support is added to the JIT compiler.
744
745 5. Fixed several bugs concerned with partial matching of items that consist
746 of more than one character:
747
748 (a) /^(..)\1/ did not partially match "aba" because checking references was
749 done on an "all or nothing" basis. This also applied to repeated
750 references.
751
752 (b) \R did not give a hard partial match if \r was found at the end of the
753 subject.
754
755 (c) \X did not give a hard partial match after matching one or more
756 characters at the end of the subject.
757
758 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
759 a partial match for the string "\r".
760
761 (e) When newline was set to CRLF, the metacharacter "." did not recognize
762 a partial match for a CR character at the end of the subject string.
763
764 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
765 running pcretest, the text "(JIT)" added to the output whenever JIT is
766 actually used to run the match.
767
768 7. Individual JIT compile options can be set in pcretest by following -s+[+]
769 or /S+[+] with a digit between 1 and 7.
770
771 8. OP_NOT now supports any UTF character not just single-byte ones.
772
773 9. (*MARK) control verb is now supported by the JIT compiler.
774
775 10. The command "./RunTest list" lists the available tests without actually
776 running any of them. (Because I keep forgetting what they all are.)
777
778 11. Add PCRE_INFO_MAXLOOKBEHIND.
779
780 12. Applied a (slightly modified) user-supplied patch that improves performance
781 when the heap is used for recursion (compiled with --disable-stack-for-
782 recursion). Instead of malloc and free for each heap frame each time a
783 logical recursion happens, frames are retained on a chain and re-used where
784 possible. This sometimes gives as much as 30% improvement.
785
786 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
787 call.
788
789 14. As documented, (*COMMIT) is now confined to within a positive assertion.
790
791 15. It is now possible to link pcretest with libedit as an alternative to
792 libreadline.
793
794 16. (*COMMIT) control verb is now supported by the JIT compiler.
795
796 17. The Unicode data tables have been updated to Unicode 6.1.0.
797
798 18. Added --file-list option to pcregrep.
799
800 19. Added binary file support to pcregrep, including the -a, --binary-files,
801 -I, and --text options.
802
803 20. The madvise function is renamed for posix_madvise for QNX compatibility
804 reasons. Fixed by Giuseppe D'Angelo.
805
806 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
807
808 22. Changed the option for creating symbolic links for 16-bit man pages from
809 -s to -sf so that re-installing does not cause issues.
810
811 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
812
813 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
814 matches in certain environments (the workspace was not being correctly
815 retained). Also added to pcre_dfa_exec() a simple plausibility check on
816 some of the workspace data at the beginning of a restart.
817
818 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
819 was not doing so when it should - probably a typo introduced by SVN 528
820 (change 8.10/14).
821
822 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
823 \w+ when the character tables indicated that \x{c4} was a word character.
824 There were several related cases, all because the tests for doing a table
825 lookup were testing for characters less than 127 instead of 255.
826
827 27. If a pattern contains capturing parentheses that are not used in a match,
828 their slots in the ovector are set to -1. For those that are higher than
829 any matched groups, this happens at the end of processing. In the case when
830 there were back references that the ovector was too small to contain
831 (causing temporary malloc'd memory to be used during matching), and the
832 highest capturing number was not used, memory off the end of the ovector
833 was incorrectly being set to -1. (It was using the size of the temporary
834 memory instead of the true size.)
835
836 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
837 ovector size, it uses memory at the end of the block that it has got.
838
839 29. Check for an overlong MARK name and give an error at compile time. The
840 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
841
842 30. JIT compiler update.
843
844 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
845 Rill for the patch.
846
847 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
848
849 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
850
851 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
852 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
853 was enabled.
854
855 35. Improve JIT code generation for greedy plus quantifier.
856
857 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
858 1 to "aa" instead of to an empty string. The bug affected repeated groups
859 that could potentially match an empty string.
860
861 37. Optimizing single character iterators in JIT.
862
863 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
864 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
865 codepoints that are too big for the mode are faulted, and in a UTF mode,
866 disallowed codepoints are also faulted.
867
868 39. If PCRE was compiled with UTF support, in three places in the DFA
869 matcher there was code that should only have been obeyed in UTF mode, but
870 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
871 processing when bytes with values greater than 127 were present. In 16-bit
872 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
873 both cases the values are those that cannot be the first data item in a UTF
874 character. The three items that might have provoked this were recursions,
875 possessively repeated groups, and atomic groups.
876
877 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
878 and pcregrep, because some OS require shared objects to be explicitly
879 passed to ld, causing the link step to fail if they are not.
880
881 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
882 mode, patterns that started with \h* or \R* might be incorrectly matched.
883
884
885 Version 8.30 04-February-2012
886 -----------------------------
887
888 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
889 name is defined in ctype.h.
890
891 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
892 only in quite long subpatterns.
893
894 3. Removed the function pcre_info(), which has been obsolete and deprecated
895 since it was replaced by pcre_fullinfo() in February 2000.
896
897 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
898 match a (*MARK), and the match failed at the start of the subject, a
899 reference to memory before the start of the subject could occur. This bug
900 was introduced by fix 17 of release 8.21.
901
902 5. A reference to an unset group with zero minimum repetition was giving
903 totally wrong answers (in non-JavaScript-compatibility mode). For example,
904 /(another)?(\1?)test/ matched against "hello world test". This bug was
905 introduced in release 8.13.
906
907 6. Add support for 16-bit character strings (a large amount of work involving
908 many changes and refactorings).
909
910 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
911 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
912 from a file.
913
914 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
915 rounding is not applied in this particular case).
916
917 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
918 if they appear, or are escaped, in patterns.
919
920 10. Get rid of a number of -Wunused-but-set-variable warnings.
921
922 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
923 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
924 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
925 also returns the mark "x". This bug applied to capturing parentheses,
926 non-capturing parentheses, and atomic parentheses. It also applied to some
927 assertions.
928
929 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
930 information out of configure.ac instead of relying on pcre.h.generic, which
931 is not stored in the repository.
932
933 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
934 -lreadline.
935
936 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
937
938 15. Applied Graycode's patch to put the top-level frame on the stack rather
939 than the heap when not using the stack for recursion. This gives a
940 performance improvement in many cases when recursion is not deep.
941
942 16. Experimental code added to "pcretest -C" to output the stack frame size.
943
944
945 Version 8.21 12-Dec-2011
946 ------------------------
947
948 1. Updating the JIT compiler.
949
950 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
951 are added as well.
952
953 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
954 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
955 calling _pcre_jit_exec. Some extra comments are added.
956
957 4. (*MARK) settings inside atomic groups that do not contain any capturing
958 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
959 was introduced by change 18 for 8.20.
960
961 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
962 ECMA-262 standard.
963
964 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
965 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
966 This bug was probably introduced by change 9 of 8.13.
967
968 7. While fixing 6 above, I noticed that a number of other items were being
969 incorrectly rejected as "not fixed length". This arose partly because newer
970 opcodes had not been added to the fixed-length checking code. I have (a)
971 corrected the bug and added tests for these items, and (b) arranged for an
972 error to occur if an unknown opcode is encountered while checking for fixed
973 length instead of just assuming "not fixed length". The items that were
974 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
975 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
976 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
977
978 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
979 being incorrectly compiled and would have given unpredicatble results.
980
981 9. A possessively repeated subpattern with minimum repeat count greater than
982 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
983 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
984 the first (A) could occur when it should not.
985
986 10. Add a cast and remove a redundant test from the code.
987
988 11. JIT should use pcre_malloc/pcre_free for allocation.
989
990 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
991 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
992 is anything other than /usr, -L is still shown).
993
994 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
995
996 14. Perl does not support \N without a following name in a [] class; PCRE now
997 also gives an error.
998
999 15. If a forward reference was repeated with an upper limit of around 2000,
1000 it caused the error "internal error: overran compiling workspace". The
1001 maximum number of forward references (including repeats) was limited by the
1002 internal workspace, and dependent on the LINK_SIZE. The code has been
1003 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1004 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1005 of around 200,000 forward references. While doing this, I also speeded up
1006 the filling in of repeated forward references.
1007
1008 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1009 incorrectly expecting the subject to contain another "a" after the start.
1010
1011 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1012 in the match, the SKIP should be ignored. This was not happening; instead
1013 the SKIP was being treated as NOMATCH. For patterns such as
1014 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1015 tested.
1016
1017 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1018 now much more compatible with Perl, in particular in cases where the result
1019 is a non-match for a non-anchored pattern. For example, if
1020 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1021 "m", where previously it did not return a name. A side effect of this
1022 change is that for partial matches, the last encountered mark name is
1023 returned, as for non matches. A number of tests that were previously not
1024 Perl-compatible have been moved into the Perl-compatible test files. The
1025 refactoring has had the pleasing side effect of removing one argument from
1026 the match() function, thus reducing its stack requirements.
1027
1028 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1029 subsequent uses of /S (without +) incorrectly behaved like /S+.
1030
1031 21. Retrieve executable code size support for the JIT compiler and fixing
1032 some warnings.
1033
1034 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1035 not work when the shorter character appeared right at the end of the
1036 subject string.
1037
1038 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1039 systems.
1040
1041 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1042 output it when the /M option is used in pcretest.
1043
1044 25. The CheckMan script was not being included in the distribution. Also, added
1045 an explicit "perl" to run Perl scripts from the PrepareRelease script
1046 because this is reportedly needed in Windows.
1047
1048 26. If study data was being save in a file and studying had not found a set of
1049 "starts with" bytes for the pattern, the data written to the file (though
1050 never used) was taken from uninitialized memory and so caused valgrind to
1051 complain.
1052
1053 27. Updated RunTest.bat as provided by Sheri Pierce.
1054
1055 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1056
1057 29. Computation of memory usage for the table of capturing group names was
1058 giving an unnecessarily large value.
1059
1060
1061 Version 8.20 21-Oct-2011
1062 ------------------------
1063
1064 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1065 a POSIX class. After further experiments with Perl, which convinced me that
1066 Perl has bugs and confusions, a closing square bracket is no longer allowed
1067 in a POSIX name. This bug also affected patterns with classes that started
1068 with full stops.
1069
1070 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1071 captured substring, but while checking the failing first alternative,
1072 substring 1 is temporarily captured. If the output vector supplied to
1073 pcre_exec() was not big enough for this capture, the yield of the function
1074 was still zero ("insufficient space for captured substrings"). This cannot
1075 be totally fixed without adding another stack variable, which seems a lot
1076 of expense for a edge case. However, I have improved the situation in cases
1077 such as /(a)(b)x|abc/ matched against "abc", where the return code
1078 indicates that fewer than the maximum number of slots in the ovector have
1079 been set.
1080
1081 3. Related to (2) above: when there are more back references in a pattern than
1082 slots in the output vector, pcre_exec() uses temporary memory during
1083 matching, and copies in the captures as far as possible afterwards. It was
1084 using the entire output vector, but this conflicts with the specification
1085 that only 2/3 is used for passing back captured substrings. Now it uses
1086 only the first 2/3, for compatibility. This is, of course, another edge
1087 case.
1088
1089 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1090 main code base, and can be used by building with --enable-jit. When this is
1091 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1092 runtime --no-jit option is given.
1093
1094 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1095 ovector, the return from the function was zero, implying that there were
1096 other matches that did not fit. The correct "exactly full" value is now
1097 returned.
1098
1099 6. If a subpattern that was called recursively or as a subroutine contained
1100 (*PRUNE) or any other control that caused it to give a non-standard return,
1101 invalid errors such as "Error -26 (nested recursion at the same subject
1102 position)" or even infinite loops could occur.
1103
1104 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1105 computing the minimum length on reaching *ACCEPT, and so ended up with the
1106 wrong value of 1 rather than 0. Further investigation indicates that
1107 computing a minimum subject length in the presence of *ACCEPT is difficult
1108 (think back references, subroutine calls), and so I have changed the code
1109 so that no minimum is registered for a pattern that contains *ACCEPT.
1110
1111 8. If (*THEN) was present in the first (true) branch of a conditional group,
1112 it was not handled as intended. [But see 16 below.]
1113
1114 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1115 Sheri Pierce.
1116
1117 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1118 the first byte in a match must be "a".
1119
1120 11. Change 17 for 8.13 increased the recursion depth for patterns like
1121 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1122 pattern contains any instances of (*THEN). If it does not, the old
1123 optimizations are restored. It would be nice to do this on a per-group
1124 basis, but at the moment that is not feasible.
1125
1126 12. In some environments, the output of pcretest -C is CRLF terminated. This
1127 broke RunTest's code that checks for the link size. A single white space
1128 character after the value is now allowed for.
1129
1130 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1131 For "fr", it uses the Windows-specific input and output files.
1132
1133 14. If (*THEN) appeared in a group that was called recursively or as a
1134 subroutine, it did not work as intended. [But see next item.]
1135
1136 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1137 pattern fragments (but not containing any | characters). If A and B are
1138 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1139 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1140 D. In other words, Perl considers parentheses that do not contain any |
1141 characters to be part of a surrounding alternative, whereas PCRE was
1142 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1143 differently. PCRE now behaves in the same way as Perl, except in the case
1144 of subroutine/recursion calls such as (?1) which have in any case always
1145 been different (but PCRE had them first :-).
1146
1147 16. Related to 15 above: Perl does not treat the | in a conditional group as
1148 creating alternatives. Such a group is treated in the same way as an
1149 ordinary group without any | characters when processing (*THEN). PCRE has
1150 been changed to match Perl's behaviour.
1151
1152 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1153 RunGrepTest script failed.
1154
1155 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1156 inevitable for groups that contain captures, but it can lead to a lot of
1157 stack use in large patterns. The old behaviour has been restored for atomic
1158 groups that do not contain any capturing parentheses.
1159
1160 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1161 suppress the check for a minimum subject length at run time. (If it was
1162 given to pcre_exec() or pcre_dfa_exec() it did work.)
1163
1164 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1165 fail to work when decoding hex characters in data strings in EBCDIC
1166 environments.
1167
1168 21. It appears that in at least one Mac OS environment, the isxdigit() function
1169 is implemented as a macro that evaluates to its argument more than once,
1170 contravening the C 90 Standard (I haven't checked a later standard). There
1171 was an instance in pcretest which caused it to go wrong when processing
1172 \x{...} escapes in subject strings. The has been rewritten to avoid using
1173 things like p++ in the argument of isxdigit().
1174
1175
1176 Version 8.13 16-Aug-2011
1177 ------------------------
1178
1179 1. The Unicode data tables have been updated to Unicode 6.0.0.
1180
1181 2. Two minor typos in pcre_internal.h have been fixed.
1182
1183 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1184 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1185 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1186
1187 4. There were a number of related bugs in the code for matching backrefences
1188 caselessly in UTF-8 mode when codes for the characters concerned were
1189 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1190 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1191 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1192 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1193 2-byte code at the end of the subject (it thought there wasn't enough data
1194 left).
1195
1196 5. Comprehensive information about what went wrong is now returned by
1197 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1198 as the output vector has at least 2 elements. The offset of the start of
1199 the failing character and a reason code are placed in the vector.
1200
1201 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1202 now returned is for the first byte of the failing character, instead of the
1203 last byte inspected. This is an incompatible change, but I hope it is small
1204 enough not to be a problem. It makes the returned offset consistent with
1205 pcre_exec() and pcre_dfa_exec().
1206
1207 7. pcretest now gives a text phrase as well as the error number when
1208 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1209 failure, the offset and reason code are output.
1210
1211 8. When \R was used with a maximizing quantifier it failed to skip backwards
1212 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1213 back over a single character (\n). This seems wrong (because it treated the
1214 two characters as a single entity when going forwards), conflicts with the
1215 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1216 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1217 has been changed.
1218
1219 9. Some internal refactoring has changed the processing so that the handling
1220 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1221 time (the PCRE_DOTALL option was changed this way some time ago: version
1222 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1223 which was always a bit of a fudge. It also means that there is one less
1224 argument for the match() function, which reduces its stack requirements
1225 slightly. This change also fixes an incompatibility with Perl: the pattern
1226 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1227
1228 10. More internal refactoring has drastically reduced the number of recursive
1229 calls to match() for possessively repeated groups such as (abc)++ when
1230 using pcre_exec().
1231
1232 11. While implementing 10, a number of bugs in the handling of groups were
1233 discovered and fixed:
1234
1235 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1236 (a|)*(?1) gave a compile-time internal error.
1237 ((a|)+)+ did not notice that the outer group could match an empty string.
1238 (^a|^)+ was not marked as anchored.
1239 (.*a|.*)+ was not marked as matching at start or after a newline.
1240
1241 12. Yet more internal refactoring has removed another argument from the match()
1242 function. Special calls to this function are now indicated by setting a
1243 value in a variable in the "match data" data block.
1244
1245 13. Be more explicit in pcre_study() instead of relying on "default" for
1246 opcodes that mean there is no starting character; this means that when new
1247 ones are added and accidentally left out of pcre_study(), testing should
1248 pick them up.
1249
1250 14. The -s option of pcretest has been documented for ages as being an old
1251 synonym of -m (show memory usage). I have changed it to mean "force study
1252 for every regex", that is, assume /S for every regex. This is similar to -i
1253 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1254 using it. It makes it easier to run collections of tests with and without
1255 study enabled, and thereby test pcre_study() more easily. All the standard
1256 tests are now run with and without -s (but some patterns can be marked as
1257 "never study" - see 20 below).
1258
1259 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1260 restoration of the capturing data to the outer values was not happening
1261 correctly.
1262
1263 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1264 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1265 pattern had matched an empty string, and so incorrectly returned a no
1266 match.
1267
1268 17. There was optimizing code for the last branch of non-capturing parentheses,
1269 and also for the obeyed branch of a conditional subexpression, which used
1270 tail recursion to cut down on stack usage. Unfortunately, now that there is
1271 the possibility of (*THEN) occurring in these branches, tail recursion is
1272 no longer possible because the return has to be checked for (*THEN). These
1273 two optimizations have therefore been removed. [But see 8.20/11 above.]
1274
1275 18. If a pattern containing \R was studied, it was assumed that \R always
1276 matched two bytes, thus causing the minimum subject length to be
1277 incorrectly computed because \R can also match just one byte.
1278
1279 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1280 was incorrectly computed.
1281
1282 20. If /S is present twice on a test pattern in pcretest input, it now
1283 *disables* studying, thereby overriding the use of -s on the command line
1284 (see 14 above). This is necessary for one or two tests to keep the output
1285 identical in both cases.
1286
1287 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1288 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1289
1290 22. When an atomic group that contained a capturing parenthesis was
1291 successfully matched, but the branch in which it appeared failed, the
1292 capturing was not being forgotten if a higher numbered group was later
1293 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1294 group 1 to "a", when in fact it should be unset. This applied to multi-
1295 branched capturing and non-capturing groups, repeated or not, and also to
1296 positive assertions (capturing in negative assertions does not happen
1297 in PCRE) and also to nested atomic groups.
1298
1299 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1300 subject after a captured substring, to make it easier to tell which of a
1301 number of identical substrings has been captured.
1302
1303 24. The way atomic groups are processed by pcre_exec() has been changed so that
1304 if they are repeated, backtracking one repetition now resets captured
1305 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1306 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1307 "aaa". Previously, it would have been "a". As part of this code
1308 refactoring, the way recursive calls are handled has also been changed.
1309
1310 25. If an assertion condition captured any substrings, they were not passed
1311 back unless some other capturing happened later. For example, if
1312 (?(?=(a))a) was matched against "a", no capturing was returned.
1313
1314 26. When studying a pattern that contained subroutine calls or assertions,
1315 the code for finding the minimum length of a possible match was handling
1316 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1317 group 1 called group 2 while simultaneously a separate group 2 called group
1318 1). A stack overflow occurred in this case. I have fixed this by limiting
1319 the recursion depth to 10.
1320
1321 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1322 Fortmann. This supports explicit test numbers on the command line, and has
1323 argument validation and error reporting.
1324
1325 28. An instance of \X with an unlimited repeat could fail if at any point the
1326 first character it looked at was a mark character.
1327
1328 29. Some minor code refactoring concerning Unicode properties and scripts
1329 should reduce the stack requirement of match() slightly.
1330
1331 30. Added the '=' option to pcretest to check the setting of unused capturing
1332 slots at the end of the pattern, which are documented as being -1, but are
1333 not included in the return count.
1334
1335 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1336 compiled something random. Now it gives a compile-time error (as does
1337 Perl).
1338
1339 32. A *MARK encountered during the processing of a positive assertion is now
1340 recorded and passed back (compatible with Perl).
1341
1342 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1343 had alternative anchored branches, the search for a second match in a line
1344 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1345 matched the line "0102" twice. The same bug affected patterns that started
1346 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1347 twice.
1348
1349 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1350 does, and because of capturing effects, quantifying parenthesized
1351 assertions may at times be useful. Quantifiers are now allowed for
1352 parenthesized assertions.
1353
1354 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1355
1356 36. \g was being checked for fancy things in a character class, when it should
1357 just be a literal "g".
1358
1359 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1360 appearance of a nested POSIX class supersedes an apparent external class.
1361 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1362 unescaped square brackets may also appear as part of class names. For
1363 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1364 more like Perl. (But see 8.20/1 above.)
1365
1366 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1367 was because it thought it was \N{name}, which is not supported).
1368
1369 39. Add minix to OS list not supporting the -S option in pcretest.
1370
1371 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1372 cannot analyze patterns in sufficient detail to catch mutual recursions
1373 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1374 subgroup is called recursively as a subpattern for a second time at the
1375 same position in the subject string. In previous releases this might have
1376 been caught by the recursion limit, or it might have run out of stack.
1377
1378 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1379 happen only once. PCRE was, however incorrectly giving a compile time error
1380 "recursive call could loop indefinitely" because it cannot analyze the
1381 pattern in sufficient detail. The compile time test no longer happens when
1382 PCRE is compiling a conditional subpattern, but actual runaway loops are
1383 now caught at runtime (see 40 above).
1384
1385 42. It seems that Perl allows any characters other than a closing parenthesis
1386 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1387 has been changed to be the same.
1388
1389 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1390 as not to get warnings when autogen.sh is called. Also changed
1391 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1392
1393 44. To help people who use pcregrep to scan files containing exceedingly long
1394 lines, the following changes have been made:
1395
1396 (a) The default value of the buffer size parameter has been increased from
1397 8K to 20K. (The actual buffer used is three times this size.)
1398
1399 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1400 PCRE is built.
1401
1402 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1403 to be set at run time.
1404
1405 (d) Numerical values in pcregrep options can be followed by K or M, for
1406 example --buffer-size=50K.
1407
1408 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1409 given and the return code is set to 2.
1410
1411 45. Add a pointer to the latest mark to the callout data block.
1412
1413 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1414 partial match of an empty string instead of no match. This was specific to
1415 the use of ".".
1416
1417 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1418 complete match instead of a partial match. This bug was dependent on both
1419 the PCRE_UTF8 and PCRE_DOTALL options being set.
1420
1421 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1422 starting byte set, because \b was not being ignored.
1423
1424
1425 Version 8.12 15-Jan-2011
1426 ------------------------
1427
1428 1. Fixed some typos in the markup of the man pages, and wrote a script that
1429 checks for such things as part of the documentation building process.
1430
1431 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1432 --match-limit and --recursion-limit options (added for 8.11). In
1433 particular, this made one of the standard tests fail. (The integer value
1434 went into the wrong half of a long int.)
1435
1436 3. If the --colour option was given to pcregrep with -v (invert match), it
1437 did strange things, either producing crazy output, or crashing. It should,
1438 of course, ignore a request for colour when reporting lines that do not
1439 match.
1440
1441 4. Another pcregrep bug caused similar problems if --colour was specified with
1442 -M (multiline) and the pattern match finished with a line ending.
1443
1444 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1445 matched in multiline mode, the following line was shown as part of the
1446 match. This seems wrong, so I have changed it.
1447
1448 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1449 the check for further matches in the same line (so they could be coloured)
1450 to overrun the end of the current line. If another match was found, it was
1451 incorrectly shown (and then shown again when found in the next line).
1452
1453 7. If pcregrep was compiled under Windows, there was a reference to the
1454 function pcregrep_exit() before it was defined. I am assuming this was
1455 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1456 reported by a user. I've moved the definition above the reference.
1457
1458
1459 Version 8.11 10-Dec-2010
1460 ------------------------
1461
1462 1. (*THEN) was not working properly if there were untried alternatives prior
1463 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1464 backtracked to try for "b" instead of moving to the next alternative branch
1465 at the same level (in this case, to look for "c"). The Perl documentation
1466 is clear that when (*THEN) is backtracked onto, it goes to the "next
1467 alternative in the innermost enclosing group".
1468
1469 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1470 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1471 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1472 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1473 (*THEN).
1474
1475 3. If \s appeared in a character class, it removed the VT character from
1476 the class, even if it had been included by some previous item, for example
1477 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1478 of \s, but is part of the POSIX "space" class.)
1479
1480 4. A partial match never returns an empty string (because you can always
1481 match an empty string at the end of the subject); however the checking for
1482 an empty string was starting at the "start of match" point. This has been
1483 changed to the "earliest inspected character" point, because the returned
1484 data for a partial match starts at this character. This means that, for
1485 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1486 (previously it gave "no match").
1487
1488 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1489 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1490 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1491 has an implication that the given string is incomplete (because a partial
1492 match is preferred over a full match). For this reason, these items now
1493 give a partial match in this situation. [Aside: previously, the one case
1494 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1495 match rather than a full match, which was wrong by the old rules, but is
1496 now correct.]
1497
1498 6. There was a bug in the handling of #-introduced comments, recognized when
1499 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1500 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1501 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1502 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1503 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1504 places in pcre_compile().
1505
1506 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1507 comments when looking ahead for named forward references to subpatterns,
1508 the only newline sequence it recognized was NL. It now handles newlines
1509 according to the set newline convention.
1510
1511 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1512 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1513 cater for a lack of strerror(). These oversights have been fixed.
1514
1515 9. Added --match-limit and --recursion-limit to pcregrep.
1516
1517 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1518
1519 11. When the -o option was used, pcregrep was setting a return code of 1, even
1520 when matches were found, and --line-buffered was not being honoured.
1521
1522 12. Added an optional parentheses number to the -o and --only-matching options
1523 of pcregrep.
1524
1525 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1526 can match an empty string. The code to do it in pcretest and pcredemo
1527 needed fixing:
1528
1529 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1530 only one byte after an empty string match just before CRLF (this case
1531 just got forgotten; "any" and "anycrlf" were OK).
1532
1533 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1534 mode when an empty string match preceded an ASCII character followed by
1535 a non-ASCII character. (The code for advancing by one character rather
1536 than one byte was nonsense.)
1537
1538 (c) The pcredemo.c sample program did not have any code at all to handle
1539 the cases when CRLF is a valid newline sequence.
1540
1541 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1542 as a starting offset was within the subject string. There is now a new
1543 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1544 negative or greater than the length of the string. In order to test this,
1545 pcretest is extended to allow the setting of negative starting offsets.
1546
1547 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1548 starting offset points to the beginning of a UTF-8 character was
1549 unnecessarily clumsy. I tidied it up.
1550
1551 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1552 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1553
1554 17. Nobody had reported that the --include_dir option, which was added in
1555 release 7.7 should have been called --include-dir (hyphen, not underscore)
1556 for compatibility with GNU grep. I have changed it to --include-dir, but
1557 left --include_dir as an undocumented synonym, and the same for
1558 --exclude-dir, though that is not available in GNU grep, at least as of
1559 release 2.5.4.
1560
1561 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1562 characters from a string of bytes) have been redefined so as not to use
1563 loops, in order to improve performance in some environments. At the same
1564 time, I abstracted some of the common code into auxiliary macros to save
1565 repetition (this should not affect the compiled code).
1566
1567 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1568 compile-time error is now given if \c is not followed by an ASCII
1569 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1570 different, and any byte value is allowed.)
1571
1572 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1573 START_OPTIMIZE option, which is now allowed at compile time - but just
1574 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1575 to pcregrep and other applications that have no direct access to PCRE
1576 options. The new /Y option in pcretest sets this option when calling
1577 pcre_compile().
1578
1579 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1580 back references. Groups containing recursive back references were forced to
1581 be atomic by that change, but in the case of named groups, the amount of
1582 memory required was incorrectly computed, leading to "Failed: internal
1583 error: code overflow". This has been fixed.
1584
1585 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1586 pcretest.c, to avoid build problems in some Borland environments.
1587
1588
1589 Version 8.10 25-Jun-2010
1590 ------------------------
1591
1592 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1593 THEN.
1594
1595 2. (*ACCEPT) was not working when inside an atomic group.
1596
1597 3. Inside a character class, \B is treated as a literal by default, but
1598 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1599 causes the error). The code is unchanged, but I tidied the documentation.
1600
1601 4. Inside a character class, PCRE always treated \R and \X as literals,
1602 whereas Perl faults them if its -w option is set. I have changed PCRE so
1603 that it faults them when PCRE_EXTRA is set.
1604
1605 5. Added support for \N, which always matches any character other than
1606 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1607
1608 6. When compiling pcregrep with newer versions of gcc which may have
1609 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1610 declared with attribute warn_unused_result" were given. Just casting the
1611 result to (void) does not stop the warnings; a more elaborate fudge is
1612 needed. I've used a macro to implement this.
1613
1614 7. Minor change to pcretest.c to avoid a compiler warning.
1615
1616 8. Added four artifical Unicode properties to help with an option to make
1617 \s etc use properties (see next item). The new properties are: Xan
1618 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1619
1620 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1621 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1622 this option. Modified pcretest to add /W to test this facility. Added
1623 REG_UCP to make it available via the POSIX interface.
1624
1625 10. Added --line-buffered to pcregrep.
1626
1627 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1628 studied, and the match started with a letter with a code point greater than
1629 127 whose first byte was different to the first byte of the other case of
1630 the letter, the other case of this starting letter was not recognized
1631 (#976).
1632
1633 12. If a pattern that was studied started with a repeated Unicode property
1634 test, for example, \p{Nd}+, there was the theoretical possibility of
1635 setting up an incorrect bitmap of starting bytes, but fortunately it could
1636 not have actually happened in practice until change 8 above was made (it
1637 added property types that matched character-matching opcodes).
1638
1639 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1640 possible starting bytes for non-anchored patterns.
1641
1642 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1643 \R, and also a number of cases that involve Unicode properties, both
1644 explicit and implicit when PCRE_UCP is set.
1645
1646 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1647 input, it could crash or give wrong results if characters with values
1648 greater than 0xc0 were present in the subject string. (Detail: it assumed
1649 UTF-8 input when processing these items.)
1650
1651 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1652 size_t is 64-bit (#991).
1653
1654 17. Added a check for running out of memory when PCRE is compiled with
1655 --disable-stack-for-recursion (#990).
1656
1657 18. If the last data line in a file for pcretest does not have a newline on
1658 the end, a newline was missing in the output.
1659
1660 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1661 less than 128) in its various bitmaps. However, there is a facility for
1662 generating tables according to the current locale when PCRE is compiled. It
1663 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1664 characters, are recognized by isspace() and therefore were getting set in
1665 these tables, and indeed these tables seem to approximate to ISO 8859. This
1666 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1667 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1668 which of course cannot start UTF-8 characters. I have changed the code so
1669 that only real ASCII characters (less than 128) and the correct starting
1670 bytes for UTF-8 encodings are set for characters greater than 127 when in
1671 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1672 altogether.)
1673
1674 20. Added the /T option to pcretest so as to be able to run tests with non-
1675 standard character tables, thus making it possible to include the tests
1676 used for 19 above in the standard set of tests.
1677
1678 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1679 reference to a subpattern the other side of a comment that contains an
1680 opening parenthesis caused either an internal compiling error, or a
1681 reference to the wrong subpattern.
1682
1683
1684 Version 8.02 19-Mar-2010
1685 ------------------------
1686
1687 1. The Unicode data tables have been updated to Unicode 5.2.0.
1688
1689 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1690 configured.
1691
1692 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1693 original author of that file, following a query about its status.
1694
1695 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1696 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1697
1698 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1699 quantifier applied to a forward-referencing subroutine call, could compile
1700 incorrect code or give the error "internal error: previously-checked
1701 referenced subpattern not found".
1702
1703 6. Both MS Visual Studio and Symbian OS have problems with initializing
1704 variables to point to external functions. For these systems, therefore,
1705 pcre_malloc etc. are now initialized to local functions that call the
1706 relevant global functions.
1707
1708 7. There were two entries missing in the vectors called coptable and poptable
1709 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1710 I've fixed the data, and added a kludgy way of testing at compile time that
1711 the lengths are correct (equal to the number of opcodes).
1712
1713 8. Following on from 7, I added a similar kludge to check the length of the
1714 eint vector in pcreposix.c.
1715
1716 9. Error texts for pcre_compile() are held as one long string to avoid too
1717 much relocation at load time. To find a text, the string is searched,
1718 counting zeros. There was no check for running off the end of the string,
1719 which could happen if a new error number was added without updating the
1720 string.
1721
1722 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1723
1724 11. \K was not working if it appeared in an atomic group or in a group that
1725 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1726 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1727 the assertion is positive, but not if it is negative.
1728
1729 12. Change 11 fortuitously reduced the size of the stack frame used in the
1730 "match()" function of pcre_exec.c by one pointer. Forthcoming
1731 implementation of support for (*MARK) will need an extra pointer on the
1732 stack; I have reserved it now, so that the stack frame size does not
1733 decrease.
1734
1735 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1736 item in branch that calls a recursion is a subroutine call - as in the
1737 second branch in the above example - was incorrectly given the compile-
1738 time error "recursive call could loop indefinitely" because pcre_compile()
1739 was not correctly checking the subroutine for matching a non-empty string.
1740
1741 14. The checks for overrunning compiling workspace could trigger after an
1742 overrun had occurred. This is a "should never occur" error, but it can be
1743 triggered by pathological patterns such as hundreds of nested parentheses.
1744 The checks now trigger 100 bytes before the end of the workspace.
1745
1746 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1747
1748
1749 Version 8.01 19-Jan-2010
1750 ------------------------
1751
1752 1. If a pattern contained a conditional subpattern with only one branch (in
1753 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1754 computed the wrong minimum data length (which is of course zero for such
1755 subpatterns). This could cause incorrect "no match" results.
1756
1757 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1758 the pattern is reset in the first branch, pcre_compile() failed with
1759 "internal error: code overflow at offset...". This happened only when
1760 the reset was to the original external option setting. (An optimization
1761 abstracts leading options settings into an external setting, which was the
1762 cause of this.)
1763
1764 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1765 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1766 assertion pattern did not match (meaning that the assertion was true), it
1767 was incorrectly treated as false if the SKIP had been reached during the
1768 matching. This also applied to assertions used as conditions.
1769
1770 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1771 assertion subpattern, including such a pattern used as a condition,
1772 unpredictable results occurred, instead of the error return
1773 PCRE_ERROR_DFA_UITEM.
1774
1775 5. The C++ GlobalReplace function was not working like Perl for the special
1776 situation when an empty string is matched. It now does the fancy magic
1777 stuff that is necessary.
1778
1779 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1780 removed. (These were left over from very, very early versions of PCRE.)
1781
1782 7. Some cosmetic changes to the code to make life easier when compiling it
1783 as part of something else:
1784
1785 (a) Change DEBUG to PCRE_DEBUG.
1786
1787 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1788 called "current" as "current_branch", to prevent a collision with the
1789 Linux macro when compiled as a kernel module.
1790
1791 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1792 prevent a collision with the Linux macro when compiled as a kernel
1793 module.
1794
1795 8. In pcre_compile() there are some checks for integer overflows that used to
1796 cast potentially large values to (double). This has been changed to that
1797 when building, a check for int64_t is made, and if it is found, it is used
1798 instead, thus avoiding the use of floating point arithmetic. (There is no
1799 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1800 double.
1801
1802 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1803 2005 (difference between two addresses compared to an unsigned value).
1804
1805 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1806 custom one, because of the following reported problem in Windows:
1807
1808 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1809 under Win32.
1810 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1811 therefore missing the function definition.
1812 - The compiler thus generates a "C" signature for the test function.
1813 - The linker fails to find the "C" function.
1814 - PCRE fails to configure if asked to do so against libbz2.
1815
1816 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1817 messages were output:
1818
1819 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1820 rerunning libtoolize, to keep the correct libtool macros in-tree.
1821 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1822
1823 I have done both of these things.
1824
1825 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1826 most of the time, it *can* run out if it is given a pattern that contains a
1827 runaway infinite recursion. I updated the discussion in the pcrestack man
1828 page.
1829
1830 13. Now that we have gone to the x.xx style of version numbers, the minor
1831 version may start with zero. Using 08 or 09 is a bad idea because users
1832 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1833 interpreted as invalid octal numbers. I've updated the previous comment in
1834 configure.ac, and also added a check that gives an error if 08 or 09 are
1835 used.
1836
1837 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1838 causing partial matching to fail when the end of the subject matched \W
1839 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1840
1841 15. There were some discrepancies between the declarations in pcre_internal.h
1842 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1843 their definitions. The declarations used "const uschar *" and the
1844 definitions used USPTR. Even though USPTR is normally defined as "const
1845 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1846 reported that: "This difference in casting confuses some C++ compilers, for
1847 example, SunCC recognizes above declarations as different functions and
1848 generates broken code for hbpcre." I have changed the declarations to use
1849 USPTR.
1850
1851 16. GNU libtool is named differently on some systems. The autogen.sh script now
1852 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1853 (FreeBSD).
1854
1855 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1856 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1857 comment: "Figure out how to create a longlong from a string: strtoll and
1858 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1859 instance, but it only takes 2 args instead of 3!"
1860
1861 18. A subtle bug concerned with back references has been fixed by a change of
1862 specification, with a corresponding code fix. A pattern such as
1863 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1864 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1865 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1866 same bug. Such groups have to be quantified to be useful, or contained
1867 inside another quantified group. (If there's no repetition, the reference
1868 can never match.) The problem arises because, having left the group and
1869 moved on to the rest of the pattern, a later failure that backtracks into
1870 the group uses the captured value from the final iteration of the group
1871 rather than the correct earlier one. I have fixed this in PCRE by forcing
1872 any group that contains a reference to itself to be an atomic group; that
1873 is, there cannot be any backtracking into it once it has completed. This is
1874 similar to recursive and subroutine calls.
1875
1876
1877 Version 8.00 19-Oct-09
1878 ----------------------
1879
1880 1. The table for translating pcre_compile() error codes into POSIX error codes
1881 was out-of-date, and there was no check on the pcre_compile() error code
1882 being within the table. This could lead to an OK return being given in
1883 error.
1884
1885 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1886 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1887 in a Windows environment.
1888
1889 3. The pcregrep --count option prints the count for each file even when it is
1890 zero, as does GNU grep. However, pcregrep was also printing all files when
1891 --files-with-matches was added. Now, when both options are given, it prints
1892 counts only for those files that have at least one match. (GNU grep just
1893 prints the file name in this circumstance, but including the count seems
1894 more useful - otherwise, why use --count?) Also ensured that the
1895 combination -clh just lists non-zero counts, with no names.
1896
1897 4. The long form of the pcregrep -F option was incorrectly implemented as
1898 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1899 but it seems right to fix it, and I didn't think it was worth preserving
1900 the old behaviour.
1901
1902 5. The command line items --regex=pattern and --regexp=pattern were not
1903 recognized by pcregrep, which required --regex pattern or --regexp pattern
1904 (with a space rather than an '='). The man page documented the '=' forms,
1905 which are compatible with GNU grep; these now work.
1906
1907 6. No libpcreposix.pc file was created for pkg-config; there was just
1908 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1909
1910 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1911 when UCP support is not needed, by modifying the Python script that
1912 generates it from Unicode data files. This should not matter if the module
1913 is correctly used as a library, but I received one complaint about 50K of
1914 unwanted data. My guess is that the person linked everything into his
1915 program rather than using a library. Anyway, it does no harm.
1916
1917 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1918 was a minimum greater than 1 for a wide character in a possessive
1919 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1920 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1921 character. Chaos in the form of incorrect output or a compiling loop could
1922 result.
1923
1924 9. The restrictions on what a pattern can contain when partial matching is
1925 requested for pcre_exec() have been removed. All patterns can now be
1926 partially matched by this function. In addition, if there are at least two
1927 slots in the offset vector, the offset of the earliest inspected character
1928 for the match and the offset of the end of the subject are set in them when
1929 PCRE_ERROR_PARTIAL is returned.
1930
1931 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1932 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1933 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1934 and may be more useful for multi-segment matching.
1935
1936 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1937 used to be given if ever the end of the subject was reached; now it is
1938 given only if matching could not proceed because another character was
1939 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1940 string "Z", which now yields "no match" instead of "partial match". In the
1941 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1942 final character ended with (*FAIL).
1943
1944 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1945 if the pattern had a "must contain" character that was already found in the
1946 earlier partial match, unless partial matching was again requested. For
1947 example, with the pattern /dog.(body)?/, the "must contain" character is
1948 "g". If the first part-match was for the string "dog", restarting with
1949 "sbody" failed. This bug has been fixed.
1950
1951 13. The string returned by pcre_dfa_exec() after a partial match has been
1952 changed so that it starts at the first inspected character rather than the
1953 first character of the match. This makes a difference only if the pattern
1954 starts with a lookbehind assertion or \b or \B (\K is not supported by
1955 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1956 matching functions compatible, and I think it's the right thing to do.
1957
1958 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1959 so that the demonstration program is easily available in environments where
1960 PCRE has not been installed from source.
1961
1962 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1963 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1964 library.
1965
1966 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1967 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1968 is not the first non-POSIX option to be added. Clearly some people find
1969 these options useful.
1970
1971 17. If a caller to the POSIX matching function regexec() passes a non-zero
1972 value for nmatch with a NULL value for pmatch, the value of
1973 nmatch is forced to zero.
1974
1975 18. RunGrepTest did not have a test for the availability of the -u option of
1976 the diff command, as RunTest does. It now checks in the same way as
1977 RunTest, and also checks for the -b option.
1978
1979 19. If an odd number of negated classes containing just a single character
1980 interposed, within parentheses, between a forward reference to a named
1981 subpattern and the definition of the subpattern, compilation crashed with
1982 an internal error, complaining that it could not find the referenced
1983 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1984 [The bug was that it was starting one character too far in when skipping
1985 over the character class, thus treating the ] as data rather than
1986 terminating the class. This meant it could skip too much.]
1987
1988 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1989 /g option in pcretest when the pattern contains \K, which makes it possible
1990 to have an empty string match not at the start, even when the pattern is
1991 anchored. Updated pcretest and pcredemo to use this option.
1992
1993 21. If the maximum number of capturing subpatterns in a recursion was greater
1994 than the maximum at the outer level, the higher number was returned, but
1995 with unset values at the outer level. The correct (outer level) value is
1996 now given.
1997
1998 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1999 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2000 make it do so. The string so far is captured, making this feature
2001 compatible with Perl.
2002
2003 23. The tests have been re-organized, adding tests 11 and 12, to make it
2004 possible to check the Perl 5.10 features against Perl 5.10.
2005
2006 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2007 pattern matches a fixed length string. PCRE did not allow this; now it
2008 does. Neither allows recursion.
2009
2010 25. I finally figured out how to implement a request to provide the minimum
2011 length of subject string that was needed in order to match a given pattern.
2012 (It was back references and recursion that I had previously got hung up
2013 on.) This code has now been added to pcre_study(); it finds a lower bound
2014 to the length of subject needed. It is not necessarily the greatest lower
2015 bound, but using it to avoid searching strings that are too short does give
2016 some useful speed-ups. The value is available to calling programs via
2017 pcre_fullinfo().
2018
2019 26. While implementing 25, I discovered to my embarrassment that pcretest had
2020 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2021 study optimizations had never been tested with that matching function.
2022 Oops. What is worse, even when it was passed study data, there was a bug in
2023 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2024 were also very few tests of studied patterns with pcre_dfa_exec().
2025
2026 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2027 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2028 on the other side of the coin, they are no longer allowed to have different
2029 names, because these cannot be distinguished in PCRE, and this has caused
2030 confusion. (This is a difference from Perl.)
2031
2032 28. When duplicate subpattern names are present (necessarily with different
2033 numbers, as required by 27 above), and a test is made by name in a
2034 conditional pattern, either for a subpattern having been matched, or for
2035 recursion in such a pattern, all the associated numbered subpatterns are
2036 tested, and the overall condition is true if the condition is true for any
2037 one of them. This is the way Perl works, and is also more like the way
2038 testing by number works.
2039
2040
2041 Version 7.9 11-Apr-09
2042 ---------------------
2043
2044 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2045 (pcretest), all targets were linked against these libraries. This included
2046 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2047 libraries. This caused unwanted dependencies to be created. This problem
2048 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2049 pcretest is linked with readline.
2050
2051 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2052 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2053 moved outside it again, because FALSE and TRUE are already defined in AIX,
2054 but BOOL is not.
2055
2056 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2057 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2058
2059 4. The pcregrep documentation said spaces were inserted as well as colons (or
2060 hyphens) following file names and line numbers when outputting matching
2061 lines. This is not true; no spaces are inserted. I have also clarified the
2062 wording for the --colour (or --color) option.
2063
2064 5. In pcregrep, when --colour was used with -o, the list of matching strings
2065 was not coloured; this is different to GNU grep, so I have changed it to be
2066 the same.
2067
2068 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2069 each matching line was coloured. Now it goes on to look for further matches
2070 of any of the test patterns, which is the same behaviour as GNU grep.
2071
2072 7. A pattern that could match an empty string could cause pcregrep to loop; it
2073 doesn't make sense to accept an empty string match in pcregrep, so I have
2074 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2075 seems to be how GNU grep behaves. [But see later change 40 for release
2076 8.33.]
2077
2078 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2079 start or after a newline", because the conditional assertion was not being
2080 correctly handled. The rule now is that both the assertion and what follows
2081 in the first alternative must satisfy the test.
2082
2083 9. If auto-callout was enabled in a pattern with a conditional group whose
2084 condition was an assertion, PCRE could crash during matching, both with
2085 pcre_exec() and pcre_dfa_exec().
2086
2087 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2088 used for matching.
2089
2090 11. Unicode property support in character classes was not working for
2091 characters (bytes) greater than 127 when not in UTF-8 mode.
2092
2093 12. Added the -M command line option to pcretest.
2094
2095 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2096
2097 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2098
2099 16. Added comments and documentation about mis-use of no_arg in the C++
2100 wrapper.
2101
2102 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2103 from Martin Jerabek that uses macro names for all relevant character and
2104 string constants.
2105
2106 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2107 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2108 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2109 these, but not everybody uses configure.
2110
2111 19. A conditional group that had only one branch was not being correctly
2112 recognized as an item that could match an empty string. This meant that an
2113 enclosing group might also not be so recognized, causing infinite looping
2114 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2115 with the subject "ab", where knowledge that the repeated group can match
2116 nothing is needed in order to break the loop.
2117
2118 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2119 exec(), but without supplying a callout function, matching went wrong.
2120
2121 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2122 leak if the size of the offset vector was greater than 30. When the vector
2123 is smaller, the saved offsets during recursion go onto a local stack
2124 vector, but for larger vectors malloc() is used. It was failing to free
2125 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2126 error, in fact).
2127
2128 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2129 heapframe that is used only when UTF-8 support is enabled. This caused no
2130 problem, but was untidy.
2131
2132 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2133 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2134 included within another project.
2135
2136 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2137 slightly modified by me:
2138
2139 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2140 not building pcregrep.
2141
2142 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2143 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2144
2145 25. Forward references, both numeric and by name, in patterns that made use of
2146 duplicate group numbers, could behave incorrectly or give incorrect errors,
2147 because when scanning forward to find the reference group, PCRE was not
2148 taking into account the duplicate group numbers. A pattern such as
2149 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2150
2151 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2152 the feature of a custom pointer more persuasive (as requested by a user).
2153
2154 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2155 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2156 pre-defined.
2157
2158 28. Added support for (*UTF8) at the start of a pattern.
2159
2160 29. Arrange for flags added by the "release type" setting in CMake to be shown
2161 in the configuration summary.
2162
2163
2164 Version 7.8 05-Sep-08
2165 ---------------------
2166
2167 1. Replaced UCP searching code with optimized version as implemented for Ad
2168 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2169 stage table and inline lookup instead of a function, giving speed ups of 2
2170 to 5 times on some simple patterns that I tested. Permission was given to
2171 distribute the MultiStage2.py script that generates the tables (it's not in
2172 the tarball, but is in the Subversion repository).
2173
2174 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2175 scripts.
2176
2177 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2178 a group with a zero qualifier. The result of the study could be incorrect,
2179 or the function might crash, depending on the pattern.
2180
2181 4. Caseless matching was not working for non-ASCII characters in back
2182 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2183 It now works when Unicode Property Support is available.
2184
2185 5. In pcretest, an escape such as \x{de} in the data was always generating
2186 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2187 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2188 truncation.
2189
2190 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2191
2192 7. Added two (int) casts to pcregrep when printing the difference of two
2193 pointers, in case they are 64-bit values.
2194
2195 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2196 test 2 if it fails.
2197
2198 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2199 and a #define of that name to empty if it is not externally set. This is to
2200 allow users of MSVC to set it if necessary.
2201
2202 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2203 the convenience functions in the pcre_get.c source file.
2204
2205 11. An option change at the start of a pattern that had top-level alternatives
2206 could cause overwriting and/or a crash. This command provoked a crash in
2207 some environments:
2208
2209 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2210
2211 This potential security problem was recorded as CVE-2008-2371.
2212
2213 12. For a pattern where the match had to start at the beginning or immediately
2214 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2215 pcre_dfa_exec() could read past the end of the passed subject if there was
2216 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2217 pcretest so that it places the subject at the end of its malloc-ed buffer.
2218
2219 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2220 exec() might read past the end of the data buffer in UTF-8 mode.
2221
2222 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2223 the data contained the byte 0x85 as part of a UTF-8 character within its
2224 first line. This applied both to normal and DFA matching.
2225
2226 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2227 /^[^d]*?$/8 failed to match "abc".
2228
2229 16. Added a missing copyright notice to pcrecpp_internal.h.
2230
2231 17. Make it more clear in the documentation that values returned from
2232 pcre_exec() in ovector are byte offsets, not character counts.
2233
2234 18. Tidied a few places to stop certain compilers from issuing warnings.
2235
2236 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2237 supplied by Stefan Weber. I made a further small update for 7.8 because
2238 there is a change of source arrangements: the pcre_searchfuncs.c module is
2239 replaced by pcre_ucd.c.
2240
2241
2242 Version 7.7 07-May-08
2243 ---------------------
2244
2245 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2246 a string to a long long, pretend we don't even have a long long." This is
2247 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2248
2249 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2250 pre-7.6 versions, which defined a global no_arg variable instead of putting
2251 it in the RE class. (See also #8 below.)
2252
2253 3. Remove a line of dead code, identified by coverity and reported by Nuno
2254 Lopes.
2255
2256 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2257
2258 (1) The include/exclude patterns were being applied to the whole pathnames
2259 of files, instead of just to the final components.
2260
2261 (2) If there was more than one level of directory, the subdirectories were
2262 skipped unless they satisfied the include/exclude conditions. This is
2263 inconsistent with GNU grep (and could even be seen as contrary to the
2264 pcregrep specification - which I improved to make it absolutely clear).
2265 The action now is always to scan all levels of directory, and just
2266 apply the include/exclude patterns to regular files.
2267
2268 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2269 --exclude_dir in the tests to avoid scanning .svn directories.
2270
2271 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2272 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2273 doesn't support NULs in patterns.
2274
2275 7. Added some missing "const"s to declarations of static tables in
2276 pcre_compile.c and pcre_dfa_exec.c.
2277
2278 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2279 caused by fix #2 above. (Subsequently also a second patch to fix the
2280 first patch. And a third patch - this was a messy problem.)
2281
2282 9. Applied Craig's patch to remove the use of push_back().
2283
2284 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2285 matching function regexec().
2286
2287 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2288 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2289 references. PCRE supports relative numbers with this syntax (I don't think
2290 Oniguruma does).
2291
2292 12. Previously, a group with a zero repeat such as (...){0} was completely
2293 omitted from the compiled regex. However, this means that if the group
2294 was called as a subroutine from elsewhere in the pattern, things went wrong
2295 (an internal error was given). Such groups are now left in the compiled
2296 pattern, with a new opcode that causes them to be skipped at execution
2297 time.
2298
2299 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2300 to the way PCRE behaves:
2301
2302 (a) A lone ] character is dis-allowed (Perl treats it as data).
2303
2304 (b) A back reference to an unmatched subpattern matches an empty string
2305 (Perl fails the current match path).
2306
2307 (c) A data ] in a character class must be notated as \] because if the
2308 first data character in a class is ], it defines an empty class. (In
2309 Perl it is not possible to have an empty class.) The empty class []
2310 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2311 The negative empty class [^] matches any one character, independently
2312 of the DOTALL setting.
2313
2314 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2315 non-existent subpattern following a character class starting with ']' and
2316 containing () gave an internal compiling error instead of "reference to
2317 non-existent subpattern". Fortunately, when the pattern did exist, the
2318 compiled code was correct. (When scanning forwards to check for the
2319 existence of the subpattern, it was treating the data ']' as terminating
2320 the class, so got the count wrong. When actually compiling, the reference
2321 was subsequently set up correctly.)
2322
2323 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2324 it was being rejected as not supported by pcre_dfa_exec(), even though
2325 other assertions are supported. I have made pcre_dfa_exec() support
2326 (*FAIL).
2327
2328 16. The implementation of 13c above involved the invention of a new opcode,
2329 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2330 cannot be changed at match time, I realized I could make a small
2331 improvement to matching performance by compiling OP_ALLANY instead of
2332 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2333 on the OP_ANY path.
2334
2335 17. Compiling pcretest on Windows with readline support failed without the
2336 following two fixes: (1) Make the unistd.h include conditional on
2337 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2338
2339 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2340 ncurses library to be included for pcretest when ReadLine support is
2341 requested, but also to allow for it to be overridden. This patch came from
2342 Daniel Bergström.
2343
2344 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2345 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2346 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2347 spotting this.
2348
2349
2350 Version 7.6 28-Jan-08
2351 ---------------------
2352
2353 1. A character class containing a very large number of characters with
2354 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2355 overflow.
2356
2357 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2358 HAVE_LONG_LONG is not defined.
2359
2360 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2361 bring them up to date and include new features. This patch includes:
2362
2363 - Fixed PH's badly added libz and libbz2 support.
2364 - Fixed a problem with static linking.
2365 - Added pcredemo. [But later removed - see 7 below.]
2366 - Fixed dftables problem and added an option.
2367 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2368 HAVE_LONG_LONG.
2369 - Added readline support for pcretest.
2370 - Added an listing of the option settings after cmake has run.
2371
2372 4. A user submitted a patch to Makefile that makes it easy to create
2373 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2374 Makefile.am that cause it to include this special target, without
2375 affecting anything else. Note that the same mingw target plus all
2376 the other distribution libraries and programs are now supported
2377 when configuring with CMake (see 6 below) instead of with
2378 Configure/Make.
2379
2380 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2381 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2382 exported in the Windows port". It has not yet been confirmed that the patch
2383 solves the problem, but it does no harm.
2384
2385 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2386 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2387 with CMake, and also correct the comment about stack recursion.
2388
2389 7. Remove the automatic building of pcredemo from the ./configure system and
2390 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2391 of a program that users should build themselves after PCRE is installed, so
2392 building it automatically is not really right. What is more, it gave
2393 trouble in some build environments.
2394
2395 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2396
2397
2398 Version 7.5 10-Jan-08
2399 ---------------------
2400
2401 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2402 values in parens when parsing an RE using the C++ wrapper."
2403
2404 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2405 Characters greater than 255 were excluded from the class instead of being
2406 included.
2407
2408 3. The same bug as (2) above applied to negated POSIX classes such as
2409 [:^space:].
2410
2411 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2412 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2413 I have changed it.
2414
2415 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2416 first named subpattern) and a construct such as (?&a) would reference the
2417 first named subpattern whose name started with "a" (in other words, the
2418 length check was missing). Both these problems are fixed. "Subpattern name
2419 expected" is now given for (?&) (a zero-length name), and this patch also
2420 makes it give the same error for \k'' (previously it complained that that
2421 was a reference to a non-existent subpattern).
2422
2423 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2424 this is right because (?- can be followed by option settings as well as by
2425 digits. I have, however, made the messages clearer.
2426
2427 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2428 than the number used in the conditional) now cause a compile-time error.
2429 This is actually not compatible with Perl, which accepts such patterns, but
2430 treats the conditional as always being FALSE (as PCRE used to), but it
2431 seems to me that giving a diagnostic is better.
2432
2433 8. Change "alphameric" to the more common word "alphanumeric" in comments
2434 and messages.
2435
2436 9. Fix two occurrences of "backslash" in comments that should have been
2437 "backspace".
2438
2439 10. Remove two redundant lines of code that can never be obeyed (their function
2440 was moved elsewhere).
2441
2442 11. The program that makes PCRE's Unicode character property table had a bug
2443 which caused it to generate incorrect table entries for sequences of
2444 characters that have the same character type, but are in different scripts.
2445 It amalgamated them into a single range, with the script of the first of
2446 them. In other words, some characters were in the wrong script. There were
2447 thirteen such cases, affecting characters in the following ranges:
2448
2449 U+002b0 - U+002c1
2450 U+0060c - U+0060d
2451 U+0061e - U+00612
2452 U+0064b - U+0065e
2453 U+0074d - U+0076d
2454 U+01800 - U+01805
2455 U+01d00 - U+01d77
2456 U+01d9b - U+01dbf
2457 U+0200b - U+0200f
2458 U+030fc - U+030fe
2459 U+03260 - U+0327f
2460 U+0fb46 - U+0fbb1
2461 U+10450 - U+1049d
2462
2463 12. The -o option (show only the matching part of a line) for pcregrep was not
2464 compatible with GNU grep in that, if there was more than one match in a
2465 line, it showed only the first of them. It now behaves in the same way as
2466 GNU grep.
2467
2468 13. If the -o and -v options were combined for pcregrep, it printed a blank
2469 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2470 does the same. The return code can be used to tell if there were any
2471 non-matching lines.
2472
2473 14. Added --file-offsets and --line-offsets to pcregrep.
2474
2475 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2476 infinitely looping recursion. The bug was that positive lookaheads were not
2477 being skipped when checking for a possible empty match (negative lookaheads
2478 and both kinds of lookbehind were skipped).
2479
2480 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2481 inclusion of <windows.h> to before rather than after the definition of
2482 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2483
2484 17. Specifying a possessive quantifier with a specific limit for a Unicode
2485 character property caused pcre_compile() to compile bad code, which led at
2486 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2487 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2488 caused the error; without that there was no problem.
2489
2490 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2491
2492 19. Added --enable-pcretest-libreadline.
2493
2494 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2495 RE::GlobalReplace(). As a result, the number of replacements returned was
2496 double what it should be. I removed one of the increments, but Craig sent a
2497 later patch that removed the other one (the right fix) and added unit tests
2498 that check the return values (which was not done before).
2499
2500 21. Several CMake things:
2501
2502 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2503 the names libpcre and libpcreposix, not just pcre and pcreposix.
2504
2505 (2) The above change means that pcretest and pcregrep are now correctly
2506 linked with the newly-built libraries, not previously installed ones.
2507
2508 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2509
2510 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2511 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2512 UTF-8 newline character). The key issue is that the pattern starts .*;
2513 this means that the match must be either at the beginning, or after a
2514 newline. The bug was in the code for advancing after a failed match and
2515 checking that the new position followed a newline. It was not taking
2516 account of UTF-8 characters correctly.
2517
2518 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2519 character classes. PCRE was not treating the sequence [:...:] as a
2520 character class unless the ... were all letters. Perl, however, seems to
2521 allow any characters between [: and :], though of course it rejects as
2522 unknown any "names" that contain non-letters, because all the known class
2523 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2524 for example, whereas PCRE did not - it did not recognize a POSIX character
2525 class. This seemed a bit dangerous, so the code has been changed to be
2526 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2527 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2528 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2529 Perl does, and where it didn't before.
2530
2531 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2532 Windows environments %n is disabled by default.
2533
2534
2535 Version 7.4 21-Sep-07
2536 ---------------------
2537
2538 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2539 means that a class such as [\s] counted as "explicit reference to CR or
2540 LF". That isn't really right - the whole point of the change was to try to
2541 help when there was an actual mention of one of the two characters. So now
2542 the change happens only if \r or \n (or a literal CR or LF) character is
2543 encountered.
2544
2545 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2546 of both had grown to the point where there were only 3 bits left.
2547 Fortunately, there was spare space in the data structure, and so I have
2548 moved the internal flags into a new 16-bit field to free up more option
2549 bits.
2550
2551 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2552 but did not set the internal JCHANGED flag - either of these is enough to
2553 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2554 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2555 start sets both bits.
2556
2557 4. Added options (at build time, compile time, exec time) to change \R from
2558 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2559
2560 5. doc/pcresyntax.html was missing from the distribution.
2561
2562 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2563 compatibility, even though it is no longer used.
2564
2565 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2566 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2567 windows.h file is present (where different names are used). [This was
2568 reversed later after testing - see 16 below.]
2569
2570 8. Changed all #include <config.h> to #include "config.h". There were also
2571 some further <pcre.h> cases that I changed to "pcre.h".
2572
2573 9. When pcregrep was used with the --colour option, it missed the line ending
2574 sequence off the lines that it output.
2575
2576 10. It was pointed out to me that arrays of string pointers cause lots of
2577 relocations when a shared library is dynamically loaded. A technique of
2578 using a single long string with a table of offsets can drastically reduce
2579 these. I have refactored PCRE in four places to do this. The result is
2580 dramatic:
2581
2582 Originally: 290
2583 After changing UCP table: 187
2584 After changing error message table: 43
2585 After changing table of "verbs" 36
2586 After changing table of Posix names 22
2587
2588 Thanks to the folks working on Gregex for glib for this insight.
2589
2590 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2591 unicode-properties was also set.
2592
2593 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2594
2595 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2596 checked only for CRLF.
2597
2598 14. Added casts to pcretest.c to avoid compiler warnings.
2599
2600 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2601
2602 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2603 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2604 entirely. This removes changes made in 7 above.
2605
2606 17. The CMake files have been updated, and there is now more information about
2607 building with CMake in the NON-UNIX-USE document.
2608
2609
2610 Version 7.3 28-Aug-07
2611 ---------------------
2612
2613 1. In the rejigging of the build system that eventually resulted in 7.1, the
2614 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2615 brackets there is not right, since it causes compilers to look for an
2616 installed pcre.h, not the version that is in the source that is being
2617 compiled (which of course may be different). I have changed it back to:
2618
2619 #include "pcre.h"
2620
2621 I have a vague recollection that the change was concerned with compiling in
2622 different directories, but in the new build system, that is taken care of
2623 by the VPATH setting the Makefile.
2624
2625 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2626 when the subject happened to end in the byte 0x85 (e.g. if the last
2627 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2628 characters but of course it shouldn't be taken as a newline when it is part
2629 of another character. The bug was that, for an unlimited repeat of . in
2630 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2631 characters when looking for a newline.
2632
2633 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2634
2635 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2636 in debug output.
2637
2638 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2639 long printing in the pcrecpp unittest when running under MinGW.
2640
2641 6. ESC_K was left out of the EBCDIC table.
2642
2643 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2644 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2645 limit also applies to "virtual nesting" when a pattern is recursive, and in
2646 this case 1000 isn't so big. I have been able to remove this limit at the
2647 expense of backing off one optimization in certain circumstances. Normally,
2648 when pcre_exec() would call its internal match() function recursively and
2649 immediately return the result unconditionally, it uses a "tail recursion"
2650 feature to save stack. However, when a subpattern that can match an empty
2651 string has an unlimited repetition quantifier, it no longer makes this
2652 optimization. That gives it a stack frame in which to save the data for
2653 checking that an empty string has been matched. Previously this was taken
2654 from the 1000-entry workspace that had been reserved. So now there is no
2655 explicit limit, but more stack is used.
2656
2657 8. Applied Daniel's patches to solve problems with the import/export magic
2658 syntax that is required for Windows, and which was going wrong for the
2659 pcreposix and pcrecpp parts of the library. These were overlooked when this
2660 problem was solved for the main library.
2661
2662 9. There were some crude static tests to avoid integer overflow when computing
2663 the size of patterns that contain repeated groups with explicit upper
2664 limits. As the maximum quantifier is 65535, the maximum group length was
2665 set at 30,000 so that the product of these two numbers did not overflow a
2666 32-bit integer. However, it turns out that people want to use groups that
2667 are longer than 30,000 bytes (though not repeat them that many times).
2668 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2669 made it possible to implement the integer overflow checks in a much more
2670 dynamic way, which I have now done. The artificial limitation on group
2671 length has been removed - we now have only the limit on the total length of
2672 the compiled pattern, which depends on the LINK_SIZE setting.
2673
2674 10. Fixed a bug in the documentation for get/copy named substring when
2675 duplicate names are permitted. If none of the named substrings are set, the
2676 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2677 empty string.
2678
2679 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2680 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2681 because the ] is interpreted as the first data character and the
2682 terminating ] is not found. PCRE has been made compatible with Perl in this
2683 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2684 cause memory overwriting.
2685
2686 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2687 string has been matched (to stop an infinite loop). It was not recognizing
2688 a conditional subpattern that could match an empty string if that
2689 subpattern was within another subpattern. For example, it looped when
2690 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2691 condition was not nested. This bug has been fixed.
2692
2693 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2694 past the start of the subject in the presence of bytes with the top bit
2695 set, for example "\x8aBCD".
2696
2697 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2698 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2699
2700 14. Optimized (?!) to (*FAIL).
2701
2702 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2703 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2704 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2705 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2706 does: it's just the validity check that is more restrictive.
2707
2708 16. Inserted checks for integer overflows during escape sequence (backslash)
2709 processing, and also fixed erroneous offset values for syntax errors during
2710 backslash processing.
2711
2712 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2713 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2714
2715 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2716 caused an overrun.
2717
2718 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2719 something other than just ASCII characters) inside a group that had an
2720 unlimited repeat caused a loop at compile time (while checking to see
2721 whether the group could match an empty string).
2722
2723 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2724 [\P{Any}] did so. (Error in the code for printing property names.)
2725
2726 21. An orphan \E inside a character class could cause a crash.
2727
2728 22. A repeated capturing bracket such as (A)? could cause a wild memory
2729 reference during compilation.
2730
2731 23. There are several functions in pcre_compile() that scan along a compiled
2732 expression for various reasons (e.g. to see if it's fixed length for look
2733 behind). There were bugs in these functions when a repeated \p or \P was
2734 present in the pattern. These operators have additional parameters compared
2735 with \d, etc, and these were not being taken into account when moving along
2736 the compiled data. Specifically:
2737
2738 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2739 length.
2740
2741 (b) An item such as \pL+ within a repeated group could cause crashes or
2742 loops.
2743
2744 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2745 "reference to non-existent subpattern" error.
2746
2747 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2748
2749 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2750 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2751
2752 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2753
2754 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2755 character were causing crashes (broken optimization).
2756
2757 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2758 \p or \P) caused a compile-time loop.
2759
2760 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2761 break. For example, the unstudied pattern [\r\n]A does not match the string
2762 "\r\nA" because change 7.0/46 below moves the current point on by two
2763 characters after failing to match at the start. However, the pattern \nA
2764 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2765 the same is true. There doesn't seem any very clean way out of this, but
2766 what I have chosen to do makes the common cases work: PCRE now takes note
2767 of whether there can be an explicit match for \r or \n anywhere in the
2768 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2769 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2770 pattern has explicit CR or LF references.
2771
2772 29. Added (*CR) etc for changing newline setting at start of pattern.
2773
2774
2775 Version 7.2 19-Jun-07
2776 ---------------------
2777
2778 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2779 which is apparently normally available under Windows.
2780
2781 2. Re-jig the pcregrep tests with different newline settings in an attempt
2782 to make them independent of the local environment's newline setting.
2783
2784 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2785
2786 4. Some of the "internals" tests were previously cut out when the link size
2787 was not 2, because the output contained actual offsets. The recent new
2788 "Z" feature of pcretest means that these can be cut out, making the tests
2789 usable with all link sizes.
2790
2791 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2792 stack recursion. This gives a massive performance boost under BSD, but just
2793 a small improvement under Linux. However, it saves one field in the frame
2794 in all cases.
2795
2796 6. Added more features from the forthcoming Perl 5.10:
2797
2798 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2799 recursion call. It refers to the nth most recently opened parentheses.
2800
2801 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2802 to be opened parentheses.
2803
2804 (c) Conditions that refer to capturing parentheses can be specified
2805 relatively, for example, (?(-2)... or (?(+3)...
2806
2807 (d) \K resets the start of the current match so that everything before
2808 is not part of it.
2809
2810 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2811
2812 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2813 reference syntax.
2814
2815 (g) (?| introduces a group in which the numbering of parentheses in each
2816 alternative starts with the same number.
2817
2818 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2819
2820 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2821 PCRE_INFO_JCHANGED.
2822
2823 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2824 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2825 for detecting groups that can match an empty string.
2826
2827 9. A pattern with a very large number of alternatives (more than several
2828 hundred) was running out of internal workspace during the pre-compile
2829 phase, where pcre_compile() figures out how much memory will be needed. A
2830 bit of new cunning has reduced the workspace needed for groups with
2831 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2832 workspace instead of running out of the 4096 that are available.
2833
2834 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2835
2836 11. Applied patch from Google to remove an optimization that didn't quite work.
2837 The report of the bug said:
2838
2839 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2840 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2841 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2842
2843 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2844 it matched the wrong number of bytes.
2845
2846
2847 Version 7.1 24-Apr-07
2848 ---------------------
2849
2850 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2851 that is more "standard", making use of automake and other Autotools. There
2852 is some re-arrangement of the files and adjustment of comments consequent
2853 on this.
2854
2855 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2856 for recursive directory scanning broke on some systems because the files
2857 are not scanned in any specific order and on different systems the order
2858 was different. A call to "sort" has been inserted into RunGrepTest for the
2859 approprate test as a short-term fix. In the longer term there may be an
2860 alternative.
2861
2862 3. I had an email from Eric Raymond about problems translating some of PCRE's
2863 man pages to HTML (despite the fact that I distribute HTML pages, some
2864 people do their own conversions for various reasons). The problems
2865 concerned the use of low-level troff macros .br and .in. I have therefore
2866 removed all such uses from the man pages (some were redundant, some could
2867 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2868 HTML has been updated to handle .nf/.fi and to complain if it encounters
2869 .br or .in.
2870
2871 4. Updated comments in configure.ac that get placed in config.h.in and also
2872 arranged for config.h to be included in the distribution, with the name
2873 config.h.generic, for the benefit of those who have to compile without
2874 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2875
2876 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2877 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2878 makevp.bat for the current PCRE, using the additional files
2879 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2880
2881 6. A Windows user reported a minor discrepancy with test 2, which turned out
2882 to be caused by a trailing space on an input line that had got lost in his
2883 copy. The trailing space was an accident, so I've just removed it.
2884
2885 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2886 that is needed.
2887
2888 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2889 as "const" (a) because they are and (b) because it helps the PHP
2890 maintainers who have recently made a script to detect big data structures
2891 in the php code that should be moved to the .rodata section. I remembered
2892 to update Builducptable as well, so it won't revert if ucptable.h is ever
2893 re-created.
2894
2895 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2896 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2897 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2898 support is not required. This saves 1.5-2K of code, which is important in
2899 some applications.
2900
2901 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2902 so as not to refer to the tables, even though these functions will never be
2903 called when UTF-8 support is disabled. Otherwise there are problems with a
2904 shared library.
2905
2906 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2907
2908 (a) It was defining its arguments as char * instead of void *.
2909
2910 (b) It was assuming that all moves were upwards in memory; this was true
2911 a long time ago when I wrote it, but is no longer the case.
2912
2913 The emulated memove() is provided for those environments that have neither
2914 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2915 is clearly not the case, as these two bugs were recently reported.
2916
2917 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2918 and Detrail to create the HTML documentation, the .txt form of the man
2919 pages, and it removes trailing spaces from listed files. It also creates
2920 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2921 case, it wraps all the #defines with #ifndefs. This script should be run
2922 before "make dist".
2923
2924 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2925 with Unicode property support.
2926
2927 (a) For a maximizing quantifier, if the two different cases of the
2928 character were of different lengths in their UTF-8 codings (there are
2929 some cases like this - I found 11), and the matching function had to
2930 back up over a mixture of the two cases, it incorrectly assumed they
2931 were both the same length.
2932
2933 (b) When PCRE was configured to use the heap rather than the stack for
2934 recursion during matching, it was not correctly preserving the data for
2935 the other case of a UTF-8 character when checking ahead for a match
2936 while processing a minimizing repeat. If the check also involved
2937 matching a wide character, but failed, corruption could cause an
2938 erroneous result when trying to check for a repeat of the original
2939 character.
2940
2941 13. Some tidying changes to the testing mechanism:
2942
2943 (a) The RunTest script now detects the internal link size and whether there
2944 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2945 values substituted by "configure". (The RunGrepTest script already did
2946 this for UTF-8.) The configure.ac script no longer substitutes the
2947 relevant variables.
2948
2949 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2950 with length and offset values. This means that the output is different
2951 for different internal link sizes. Test 2 is skipped for link sizes
2952 other than 2 because of this, bypassing the problem. Unfortunately,
2953 there was also a test in test 3 (the locale tests) that used /B and
2954 failed for link sizes other than 2. Rather than cut the whole test out,
2955 I have added a new /Z option to pcretest that replaces the length and
2956 offset values with spaces. This is now used to make test 3 independent
2957 of link size. (Test 2 will be tidied up later.)
2958
2959 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2960 segmentation fault instead of returning the appropriate error message.
2961
2962 15. In multiline mode when the newline sequence was set to "any", the pattern
2963 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2964 This doesn't seem right; it now treats the CRLF combination as the line
2965 ending, and so does not match in that case. It's only a pattern such as ^$
2966 that would hit this one: something like ^ABC$ would have failed after \r
2967 and then tried again after \r\n.
2968
2969 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2970 in an attempt to make files that differ only in their line terminators
2971 compare equal. This works on Linux.
2972
2973 17. Under certain error circumstances pcregrep might try to free random memory
2974 as it exited. This is now fixed, thanks to valgrind.
2975
2976 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2977 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2978 was because its rules for how to advance for /g after matching an empty
2979 string at the end of a line did not allow for this case. They now check for
2980 it specially.
2981
2982 20. pcretest is supposed to handle patterns and data of any length, by
2983 extending its buffers when necessary. It was getting this wrong when the
2984 buffer for a data line had to be extended.
2985
2986 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2987 CRLF as a newline sequence.
2988
2989 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2990 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2991 I have nevertheless tidied it up.
2992
2993 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2994
2995 24. Added a man page for pcre-config.
2996
2997
2998 Version 7.0 19-Dec-06
2999 ---------------------
3000
3001 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3002 moving to gcc 4.1.1.
3003
3004 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3005 sys/time.h, which is documented as needed for this function. It doesn't
3006 seem to matter on Linux, but it showed up on some releases of OS X.
3007
3008 3. It seems that there are systems where bytes whose values are greater than
3009 127 match isprint() in the "C" locale. The "C" locale should be the
3010 default when a C program starts up. In most systems, only ASCII printing
3011 characters match isprint(). This difference caused the output from pcretest
3012 to vary, making some of the tests fail. I have changed pcretest so that:
3013
3014 (a) When it is outputting text in the compiled version of a pattern, bytes
3015 other than 32-126 are always shown as hex escapes.
3016
3017 (b) When it is outputting text that is a matched part of a subject string,
3018 it does the same, unless a different locale has been set for the match
3019 (using the /L modifier). In this case, it uses isprint() to decide.
3020
3021 4. Fixed a major bug that caused incorrect computation of the amount of memory
3022 required for a compiled pattern when options that changed within the
3023 pattern affected the logic of the preliminary scan that determines the
3024 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3025 that the computed length was too small. The symptoms of this bug were
3026 either the PCRE error "internal error: code overflow" from pcre_compile(),
3027 or a glibc crash with a message such as "pcretest: free(): invalid next
3028 size (fast)". Examples of patterns that provoked this bug (shown in
3029 pcretest format) are:
3030
3031 /(?-x: )/x
3032 /(?x)(?-x: \s*#\s*)/
3033 /((?i)[\x{c0}])/8
3034 /(?i:[\x{c0}])/8
3035
3036 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3037 is now done differently.
3038
3039 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3040 wrapper classes; (b) implement a new function in the C++ scanner that is
3041 more efficient than the old way of doing things because it avoids levels of
3042 recursion in the regex matching; (c) add a paragraph to the documentation
3043 for the FullMatch() function.
3044
3045 6. The escape sequence \n was being treated as whatever was defined as
3046 "newline". Not only was this contrary to the documentation, which states
3047 that \n is character 10 (hex 0A), but it also went horribly wrong when
3048 "newline" was defined as CRLF. This has been fixed.
3049
3050 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3051 was being set to -1 for the "end of line" case (supposedly a value that no
3052 character can have). Though this value is never used (the check for end of
3053 line is "zero bytes in current character"), it caused compiler complaints.
3054 I've changed it to 0xffffffff.
3055
3056 8. In pcre_version.c, the version string was being built by a sequence of
3057 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3058 string (as it is for production releases) called a macro with an empty
3059 argument. The C standard says the result of this is undefined. The gcc
3060 compiler treats it as an empty string (which was what was wanted) but it is
3061 reported that Visual C gives an error. The source has been hacked around to
3062 avoid this problem.
3063
3064 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3065 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3066 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3067 of them did).
3068
3069 10. Originally, pcretest opened its input and output without "b"; then I was
3070 told that "b" was needed in some environments, so it was added for release
3071 5.0 to both the input and output. (It makes no difference on Unix-like
3072 systems.) Later I was told that it is wrong for the input on Windows. I've
3073 now abstracted the modes into two macros, to make it easier to fiddle with
3074 them, and removed "b" from the input mode under Windows.
3075
3076 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3077
3078 12. Added -help and --help to pcretest as an official way of being reminded
3079 of the options.
3080
3081 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3082 and pcrecpp.cc because they annoy compilers at high warning levels.
3083
3084 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3085
3086 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3087 scripts are not C programs :-) and which was not noticed because it works
3088 on Linux.
3089
3090 16. pcretest is supposed to handle any length of pattern and data line (as one
3091 line or as a continued sequence of lines) by extending its input buffer if
3092 necessary. This feature was broken for very long pattern lines, leading to
3093 a string of junk being passed to pcre_compile() if the pattern was longer
3094 than about 50K.
3095
3096 17. I have done a major re-factoring of the way pcre_compile() computes the
3097 amount of memory needed for a compiled pattern. Previously, there was code
3098 that made a preliminary scan of the pattern in order to do this. That was
3099 OK when PCRE was new, but as the facilities have expanded, it has become
3100 harder and harder to keep it in step with the real compile phase, and there
3101 have been a number of bugs (see for example, 4 above). I have now found a
3102 cunning way of running the real compile function in a "fake" mode that
3103 enables it to compute how much memory it would need, while actually only
3104 ever using a few hundred bytes of working memory and without too many
3105 tests of the mode. This should make future maintenance and development
3106 easier. A side effect of this work is that the limit of 200 on the nesting
3107 depth of parentheses has been removed (though this was never a serious
3108 limitation, I suspect). However, there is a downside: pcre_compile() now
3109 runs more slowly than before (30% or more, depending on the pattern). I
3110 hope this isn't a big issue. There is no effect on runtime performance.
3111
3112 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3113 newline (only possible for the last line of a file) and it was a
3114 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3115
3116 19. Added additional timing features to pcretest. (1) The -tm option now times
3117 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3118 separate command line item, by a number that specifies the number of
3119 repeats to use when timing. The default is 50000; this gives better
3120 precision, but takes uncomfortably long for very large patterns.
3121
3122 20. Extended pcre_study() to be more clever in cases where a branch of a
3123 subpattern has no definite first character. For example, (a*|b*)[cd] would
3124 previously give no result from pcre_study(). Now it recognizes that the
3125 first character must be a, b, c, or d.
3126
3127 21. There was an incorrect error "recursive call could loop indefinitely" if
3128 a subpattern (or the entire pattern) that was being tested for matching an
3129 empty string contained only one non-empty item after a nested subpattern.
3130 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3131 incorrectly, because the \d was being skipped in the check.
3132
3133 22. The pcretest program now has a new pattern option /B and a command line
3134 option -b, which is equivalent to adding /B to every pattern. This causes
3135 it to show the compiled bytecode, without the additional information that
3136 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3137 is the same as /B/I).
3138
3139 23. A new optimization is now able automatically to treat some sequences such
3140 as a*b as a*+b. More specifically, if something simple (such as a character
3141 or a simple class like \d) has an unlimited quantifier, and is followed by
3142 something that cannot possibly match the quantified thing, the quantifier
3143 is automatically "possessified".
3144
3145 24. A recursive reference to a subpattern whose number was greater than 39
3146 went wrong under certain circumstances in UTF-8 mode. This bug could also
3147 have affected the operation of pcre_study().
3148
3149 25. Realized that a little bit of performance could be had by replacing
3150 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3151
3152 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3153
3154 27. Possessive quantifiers such as a++ were previously implemented by turning
3155 them into atomic groups such as ($>a+). Now they have their own opcodes,
3156 which improves performance. This includes the automatically created ones
3157 from 23 above.
3158
3159 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3160 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3161 the first matched character to be a colon. This applied both to named and
3162 numbered groups.
3163
3164 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3165
3166 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3167 building PCRE on Windows easier, so I have included it in the distribution.
3168
3169 31. There is now a check in pcretest against a ridiculously large number being
3170 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3171 loop, the loop is abandoned.
3172
3173 32. Forward references to subpatterns in conditions such as (?(2)...) where
3174 subpattern 2 is defined later cause pcre_compile() to search forwards in
3175 the pattern for the relevant set of parentheses. This search went wrong
3176 when there were unescaped parentheses in a character class, parentheses
3177 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3178
3179 33. "Subroutine" calls and backreferences were previously restricted to
3180 referencing subpatterns earlier in the regex. This restriction has now
3181 been removed.
3182
3183 34. Added a number of extra features that are going to be in Perl 5.10. On the
3184 whole, these are just syntactic alternatives for features that PCRE had
3185 previously implemented using the Python syntax or my own invention. The
3186 other formats are all retained for compatibility.
3187
3188 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3189 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3190 also .NET compatible.
3191
3192 (b) A recursion or subroutine call to a named group can now be defined as
3193 (?&name) as well as (?P>name).
3194
3195 (c) A backreference to a named group can now be defined as \k<name> or
3196 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3197 5.10, are also .NET compatible.
3198
3199 (d) A conditional reference to a named group can now use the syntax
3200 (?(<name>) or (?('name') as well as (?(name).
3201
3202 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3203 groups (named and numbered) that are never evaluated inline, but can be
3204 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3205 is always false. There may be only one alternative in such a group.
3206
3207 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3208 as the simple (?(R). The condition is true only if the most recent
3209 recursion is that of the given number or name. It does not search out
3210 through the entire recursion stack.
3211
3212 (g) The escape \gN or \g{N} has been added, where N is a positive or
3213 negative number, specifying an absolute or relative reference.
3214
3215 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3216 some "unreachable code" warnings.
3217
3218 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3219 things, this adds five new scripts.
3220
3221 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3222 There were also incompatibilities regarding the handling of \Q..\E inside
3223 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3224 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3225
3226 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3227 matches an empty string, and forcibly breaks the loop. There were bugs in
3228 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3229 against aaaa the result was just "a" rather than "aaaa", for example. Two
3230 separate and independent bugs (that affected different cases) have been
3231 fixed.
3232
3233 39. Refactored the code to abolish the use of different opcodes for small
3234 capturing bracket numbers. This is a tidy that I avoided doing when I
3235 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3236 The new approach is not only tidier, it makes it possible to reduce the
3237 memory needed to fix the previous bug (38).
3238
3239 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3240 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3241 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3242 mode.
3243
3244 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3245 report.
3246
3247 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3248 copy construction and assignment in the C++ wrapper.
3249
3250 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3251 couple of bugs that could have given wrong results in the "--newline=crlf"
3252 case.
3253
3254 44. Added a number of casts and did some reorganization of signed/unsigned int
3255 variables following suggestions from Dair Grant. Also renamed the variable
3256 "this" as "item" because it is a C++ keyword.
3257
3258 45. Arranged for dftables to add
3259
3260 #include "pcre_internal.h"
3261
3262 to pcre_chartables.c because without it, gcc 4.x may remove the array
3263 definition from the final binary if PCRE is built into a static library and
3264 dead code stripping is activated.
3265
3266 46. For an unanchored pattern, if a match attempt fails at the start of a
3267 newline sequence, and the newline setting is CRLF or ANY, and the next two
3268 characters are CRLF, advance by two characters instead of one.
3269
3270
3271 Version 6.7 04-Jul-06
3272 ---------------------
3273
3274 1. In order to handle tests when input lines are enormously long, pcretest has
3275 been re-factored so that it automatically extends its buffers when
3276 necessary. The code is crude, but this _is_ just a test program. The
3277 default size has been increased from 32K to 50K.
3278
3279 2. The code in pcre_study() was using the value of the re argument before
3280 testing it for NULL. (Of course, in any sensible call of the function, it
3281 won't be NULL.)
3282
3283 3. The memmove() emulation function in pcre_internal.h, which is used on
3284 systems that lack both memmove() and bcopy() - that is, hardly ever -
3285 was missing a "static" storage class specifier.
3286
3287 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3288 containing an extended class (one that cannot be represented by a bitmap
3289 because it contains high-valued characters or Unicode property items, e.g.
3290 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3291 pattern, but PCRE should not loop if you do not (it no longer does).
3292 [Detail: two cases were found: (a) a repeated subpattern containing an
3293 extended class; (b) a recursive reference to a subpattern that followed a
3294 previous extended class. It wasn't skipping over the extended class
3295 correctly when UTF-8 mode was not set.]
3296
3297 5. A negated single-character class was not being recognized as fixed-length
3298 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3299 compile error "lookbehind assertion is not fixed length".
3300
3301 6. The RunPerlTest auxiliary script was showing an unexpected difference
3302 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3303 write a Perl script that can interpret lines of an input file either as
3304 byte characters or as UTF-8, which is what "perltest" was being required to
3305 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3306 can't do is switch easily at run time between having the "use utf8;" pragma
3307 or not. In the end, I fudged it by using the RunPerlTest script to insert
3308 "use utf8;" explicitly for the UTF-8 tests.
3309
3310 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3311 the end of the subject string, contrary to the documentation and to what
3312 Perl does. This was true of both matching functions. Now it matches only at
3313 the start of the subject and immediately after *internal* newlines.
3314
3315 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3316 a pointer to an int instead of a pointer to an unsigned long int. This
3317 caused problems on 64-bit systems.
3318
3319 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3320 instance of the 'standard' template library not being so standard".
3321
3322 10. There was no check on the number of named subpatterns nor the maximum
3323 length of a subpattern name. The product of these values is used to compute
3324 the size of the memory block for a compiled pattern. By supplying a very
3325 long subpattern name and a large number of named subpatterns, the size
3326 computation could be caused to overflow. This is now prevented by limiting
3327 the length of names to 32 characters, and the number of named subpatterns
3328 to 10,000.
3329
3330 11. Subpatterns that are repeated with specific counts have to be replicated in
3331 the compiled pattern. The size of memory for this was computed from the
3332 length of the subpattern and the repeat count. The latter is limited to
3333 65535, but there was no limit on the former, meaning that integer overflow
3334 could in principle occur. The compiled length of a repeated subpattern is
3335 now limited to 30,000 bytes in order to prevent this.
3336
3337 12. Added the optional facility to have named substrings with the same name.
3338
3339 13. Added the ability to use a named substring as a condition, using the
3340 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3341 are numbers (not recommended). Forward references are permitted.
3342
3343 14. Added forward references in named backreferences (if you see what I mean).
3344
3345 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3346 pattern could run off the end of the subject. For example, the pattern
3347 "(?s)(.{1,5})"8 did this with the subject "ab".
3348
3349 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3350 PCRE_CASELESS was set when matching characters that were quantified with ?
3351 or *.
3352
3353 17. A character class other than a single negated character that had a minimum
3354 but no maximum quantifier - for example [ab]{6,} - was not handled
3355 correctly by pce_dfa_exec(). It would match only one character.
3356
3357 18. A valid (though odd) pattern that looked like a POSIX character
3358 class but used an invalid character after [ (for example [[,abc,]]) caused
3359 pcre_compile() to give the error "Failed: internal error: code overflow" or
3360 in some cases to crash with a glibc free() error. This could even happen if
3361 the pattern terminated after [[ but there just happened to be a sequence of
3362 letters, a binary zero, and a closing ] in the memory that followed.
3363
3364 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3365 over the years. Originally (before any Unicode support), just the bottom 8
3366 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3367 output from "man perlunicode" includes this:
3368
3369 The regular expression compiler produces polymorphic opcodes. That
3370 is, the pattern adapts to the data and automatically switches to
3371 the Unicode character scheme when presented with Unicode data--or
3372 instead uses a traditional byte scheme when presented with byte
3373 data.
3374
3375 Sadly, a wide octal escape does not cause a switch, and in a string with
3376 no other multibyte characters, these octal escapes are treated as before.
3377 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3378 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3379 Unicode string.
3380
3381 I have not perpetrated such confusion in PCRE. Up till now, it took just
3382 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3383 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3384 translate to the appropriate multibyte character.
3385
3386 29. Applied some refactoring to reduce the number of warnings from Microsoft
3387 and Borland compilers. This has included removing the fudge introduced
3388 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3389 a warning about an unused variable.
3390
3391 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3392 characters since release 4.0, because Perl (from release 5.004) does not.
3393 [Or at least, is documented not to: some releases seem to be in conflict
3394 with the documentation.] However, when a pattern was studied with
3395 pcre_study() and all its branches started with \s, PCRE still included VT
3396 as a possible starting character. Of course, this did no harm; it just
3397 caused an unnecessary match attempt.
3398
3399 22. Removed a now-redundant internal flag bit that recorded the fact that case
3400 dependency changed within the pattern. This was once needed for "required
3401 byte" processing, but is no longer used. This recovers a now-scarce options
3402 bit. Also moved the least significant internal flag bit to the most-
3403 significant bit of the word, which was not previously used (hangover from
3404 the days when it was an int rather than a uint) to free up another bit for
3405 the future.
3406
3407 23. Added support for CRLF line endings as well as CR and LF. As well as the
3408 default being selectable at build time, it can now be changed at runtime
3409 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3410 specify that it is scanning data with non-default line endings.
3411
3412 24. Changed the definition of CXXLINK to make it agree with the definition of
3413 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3414
3415 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3416 recursions. This makes a big different to stack usage for some patterns.
3417
3418 26. If a subpattern containing a named recursion or subroutine reference such
3419 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3420 the space required for the compiled pattern went wrong and gave too small a
3421 value. Depending on the environment, this could lead to "Failed: internal
3422 error: code overflow at offset 49" or "glibc detected double free or
3423 corruption" errors.
3424
3425 27. Applied patches from Google (a) to support the new newline modes and (b) to
3426 advance over multibyte UTF-8 characters in GlobalReplace.
3427
3428 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3429 difference for some implementation of PCRE in some Windows version.
3430
3431 29. Added some extra testing facilities to pcretest:
3432
3433 \q<number> in a data line sets the "match limit" value
3434 \Q<number> in a data line sets the "match recursion limt" value
3435 -S <number> sets the stack size, where <number> is in megabytes
3436
3437 The -S option isn't available for Windows.
3438
3439
3440 Version 6.6 06-Feb-06
3441 ---------------------
3442
3443 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3444 in pcreposix.h. I have copied the definition from pcre.h.
3445
3446 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3447 because pcre.h is no longer a built file.
3448
3449 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3450 not normally included in the compiled code.
3451
3452
3453 Version 6.5 01-Feb-06
3454 ---------------------
3455
3456 1. When using the partial match feature with pcre_dfa_exec(), it was not
3457 anchoring the second and subsequent partial matches at the new starting
3458 point. This could lead to incorrect results. For example, with the pattern
3459 /1234/, partially matching against "123" and then "a4" gave a match.
3460
3461 2. Changes to pcregrep:
3462
3463 (a) All non-match returns from pcre_exec() were being treated as failures
3464 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3465 error message is output. Some extra information is given for the
3466 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3467 probably the only errors that are likely to be caused by users (by
3468 specifying a regex that has nested indefinite repeats, for instance).
3469 If there are more than 20 of these errors, pcregrep is abandoned.
3470
3471 (b) A binary zero was treated as data while matching, but terminated the
3472 output line if it was written out. This has been fixed: binary zeroes
3473 are now no different to any other data bytes.
3474
3475 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3476 used to set a locale for matching. The --locale=xxxx long option has
3477 been added (no short equivalent) to specify a locale explicitly on the
3478 pcregrep command, overriding the environment variables.
3479
3480 (d) When -B was used with -n, some line numbers in the output were one less
3481 than they should have been.
3482
3483 (e) Added the -o (--only-matching) option.
3484
3485 (f) If -A or -C was used with -c (count only), some lines of context were
3486 accidentally printed for the final match.
3487
3488 (g) Added the -H (--with-filename) option.
3489
3490 (h) The combination of options -rh failed to suppress file names for files
3491 that were found from directory arguments.
3492
3493 (i) Added the -D (--devices) and -d (--directories) options.
3494
3495 (j) Added the -F (--fixed-strings) option.
3496
3497 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3498
3499 (l) Added the --colo(u)r option.
3500
3501 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3502 is not present by default.
3503
3504 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3505 items such as (?R) or (?1), when the recursion could match a number of
3506 alternatives. If it matched one of the alternatives, but subsequently,
3507 outside the recursion, there was a failure, the code tried to back up into
3508 the recursion. However, because of the way PCRE is implemented, this is not
3509 possible, and the result was an incorrect result from the match.
3510
3511 In order to prevent this happening, the specification of recursion has
3512 been changed so that all such subpatterns are automatically treated as
3513 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3514
3515 4. I had overlooked the fact that, in some locales, there are characters for
3516 which isalpha() is true but neither isupper() nor islower() are true. In
3517 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3518 and ordfeminine) are like this. This affected the treatment of \w and \W
3519 when they appeared in character classes, but not when they appeared outside
3520 a character class. The bit map for "word" characters is now created
3521 separately from the results of isalnum() instead of just taking it from the
3522 upper, lower, and digit maps. (Plus the underscore character, of course.)
3523
3524 5. The above bug also affected the handling of POSIX character classes such as
3525 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3526 permanent tables. Instead, the bit maps for such a class were previously
3527 created as the appropriate unions of the upper, lower, and digit bitmaps.
3528 Now they are created by subtraction from the [[:word:]] class, which has
3529 its own bitmap.
3530
3531 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3532 It is created by subtracting the vertical space characters (\x09, \x0a,
3533 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3534 subtraction was done in the overall bitmap for a character class, meaning
3535 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3536 be recognized. This bug has been fixed.
3537
3538 7. Patches from the folks at Google:
3539
3540 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3541 real life, but is still worth protecting against".
3542
3543 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3544 regular expressions".
3545
3546 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3547 have it.
3548
3549 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3550 "configure" and the latter not, in order to fix a problem somebody had
3551 with compiling the Arg class on HP-UX.
3552
3553 (e) Improve the error-handling of the C++ wrapper a little bit.
3554
3555 (f) New tests for checking recursion limiting.
3556
3557 8. The pcre_memmove() function, which is used only if the environment does not
3558 have a standard memmove() function (and is therefore rarely compiled),
3559 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3560 returning a result (though PCRE never actually uses the result).
3561
3562 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3563 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3564 returned instead of calling malloc() with an overflowing number that would
3565 most likely cause subsequent chaos.
3566
3567 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3568
3569 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3570 with this option is matched, the nmatch and pmatch options of regexec() are
3571 ignored.
3572
3573 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3574 provided in case anyone wants to the the POSIX interface with UTF-8
3575 strings.
3576
3577 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3578 C++ linking (needed for some HP-UX environments).
3579
3580 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3581 (unused parameter) and in the pcre_printint() function (omitted "default"
3582 switch label when the default is to do nothing).
3583
3584 15. Added some code to make it possible, when PCRE is compiled as a C++
3585 library, to replace subject pointers for pcre_exec() with a smart pointer
3586 class, thus making it possible to process discontinuous strings.
3587
3588 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3589 much the same function. They were added by different people who were trying
3590 to make PCRE easy to compile on non-Unix systems. It has been suggested
3591 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3592 for compiling on Windows systems. I have therefore replaced it with
3593 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3594 defaults to "extern" for C or "extern C" for C++, which works fine on
3595 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3596 SCOPE with something explicit in config.h. In addition:
3597
3598 (a) pcreposix.h still had just "extern" instead of either of these macros;
3599 I have replaced it with PCRE_DATA_SCOPE.
3600
3601 (b) Functions such as _pcre_xclass(), which are internal to the library,
3602 but external in the C sense, all had PCRE_EXPORT in their definitions.
3603 This is apparently wrong for the Windows case, so I have removed it.
3604 (It makes no difference on Unix-like systems.)
3605
3606 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3607 of recursive calls to match(). This is different to MATCH_LIMIT because
3608 that limits the total number of calls to match(), not all of which increase
3609 the depth of recursion. Limiting the recursion depth limits the amount of
3610 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3611 when PCRE is compiled, and changed at run time. A patch from Google adds
3612 this functionality to the C++ interface.
3613
3614 18. Changes to the handling of Unicode character properties:
3615
3616 (a) Updated the table to Unicode 4.1.0.
3617
3618 (b) Recognize characters that are not in the table as "Cn" (undefined).
3619
3620 (c) I revised the way the table is implemented to a much improved format
3621 which includes recognition of ranges. It now supports the ranges that
3622 are defined in UnicodeData.txt, and it also amalgamates other
3623 characters into ranges. This has reduced the number of entries in the
3624 table from around 16,000 to around 3,000, thus reducing its size
3625 considerably. I realized I did not need to use a tree structure after
3626 all - a binary chop search is just as efficient. Having reduced the
3627 number of entries, I extended their size from 6 bytes to 8 bytes to
3628 allow for more data.
3629
3630 (d) Added support for Unicode script names via properties such as \p{Han}.
3631
3632 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3633 matching that character.
3634
3635 20. When matching a repeated Unicode property with a minimum greater than zero,
3636 (for example \pL{2,}), PCRE could look past the end of the subject if it
3637 reached it while seeking the minimum number of characters. This could
3638 happen only if some of the characters were more than one byte long, because
3639 there is a check for at least the minimum number of bytes.
3640
3641 21. Refactored the implementation of \p and \P so as to be more general, to
3642 allow for more different types of property in future. This has changed the
3643 compiled form incompatibly. Anybody with saved compiled patterns that use
3644 \p or \P will have to recompile them.
3645
3646 22. Added "Any" and "L&" to the supported property types.
3647
3648 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3649 but give a compile time error if the value is greater than 0xff.
3650
3651 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3652 accidentally not being installed or uninstalled.
3653
3654 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3655 made were to insert the current release number. This seemed silly, because
3656 it made things harder for people building PCRE on systems that don't run
3657 "configure". I have turned pcre.h into a distributed file, no longer built
3658 by "configure", with the version identification directly included. There is
3659 no longer a pcre.h.in file.
3660
3661 However, this change necessitated a change to the pcre-config script as
3662 well. It is built from pcre-config.in, and one of the substitutions was the
3663 release number. I have updated configure.ac so that ./configure now finds
3664 the release number by grepping pcre.h.
3665
3666 26. Added the ability to run the tests under valgrind.
3667
3668
3669 Version 6.4 05-Sep-05
3670 ---------------------
3671
3672 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3673 "--" to be printed when multiple files were scanned, even when none of the
3674 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3675 consider it to be a bug, and have restored the previous behaviour.
3676
3677 2. A couple of code tidies to get rid of compiler warnings.
3678
3679 3. The pcretest program used to cheat by referring to symbols in the library
3680 whose names begin with _pcre_. These are internal symbols that are not
3681 really supposed to be visible externally, and in some environments it is
3682 possible to suppress them. The cheating is now confined to including
3683 certain files from the library's source, which is a bit cleaner.
3684
3685 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3686 file's purpose clearer.
3687
3688 5. Reorganized pcre_ucp_findchar().
3689
3690
3691 Version 6.3 15-Aug-05
3692 ---------------------
3693
3694 1. The file libpcre.pc.in did not have general read permission in the tarball.
3695
3696 2. There were some problems when building without C++ support:
3697
3698 (a) If C++ support was not built, "make install" and "make test" still
3699 tried to test it.
3700
3701 (b) There were problems when the value of CXX was explicitly set. Some
3702 changes have been made to try to fix these, and ...
3703
3704 (c) --disable-cpp can now be used to explicitly disable C++ support.
3705
3706 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3707 backslash in a target when C++ was disabled. This confuses some
3708 versions of "make", apparently. Using an intermediate variable solves
3709 this. (Same for CPP_LOBJ.)
3710
3711 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3712 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3713 necessary on certain architectures.
3714
3715 4. Added a setting of -export-symbols-regex to the link command to remove
3716 those symbols that are exported in the C sense, but actually are local
3717 within the library, and not documented. Their names all begin with
3718 "_pcre_". This is not a perfect job, because (a) we have to except some
3719 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3720 available (and never for static libraries). I have made a note to try to
3721 find a way round (a) in the future.
3722
3723
3724 Version 6.2 01-Aug-05
3725 ---------------------
3726
3727 1. There was no test for integer overflow of quantifier values. A construction
3728 such as {1111111111111111} would give undefined results. What is worse, if
3729 a minimum quantifier for a parenthesized subpattern overflowed and became
3730 negative, the calculation of the memory size went wrong. This could have
3731 led to memory overwriting.
3732
3733 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3734
3735 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3736 operating environments where this matters.
3737
3738 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3739 PCRE options from within the C++ wrapper.
3740
3741 5. Named capturing subpatterns were not being correctly counted when a pattern
3742 was compiled. This caused two problems: (a) If there were more than 100
3743 such subpatterns, the calculation of the memory needed for the whole
3744 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3745 back references of the form \12, where the number was greater than 9, were
3746 not recognized as back references, even though there were sufficient
3747 previous subpatterns.
3748
3749 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3750 versions of gcc, e.g. 2.95.4.
3751
3752
3753 Version 6.1 21-Jun-05
3754 ---------------------
3755
3756 1. There was one reference to the variable "posix" in pcretest.c that was not
3757 surrounded by "#if !defined NOPOSIX".
3758
3759 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3760 the cross-check on the old pcre_info() function, for the benefit of the
3761 cut-down version of PCRE that is currently imported into Exim.
3762
3763 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3764 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3765 patterns that start (?i) but I don't think that matters. The use of (?i) is
3766 just an example; this all applies to the other options as well.
3767
3768 4. Since libtool seems to echo the compile commands it is issuing, the output
3769 from "make" can be reduced a bit by putting "@" in front of each libtool
3770 compile command.
3771
3772 5. Patch from the folks at Google for configure.in to be a bit more thorough
3773 in checking for a suitable C++ installation before trying to compile the
3774 C++ stuff. This should fix a reported problem when a compiler was present,
3775 but no suitable headers.
3776
3777 6. The man pages all had just "PCRE" as their title. I have changed them to
3778 be the relevant file name. I have also arranged that these names are
3779 retained in the file doc/pcre.txt, which is a concatenation in text format
3780 of all the man pages except the little individual ones for each function.
3781
3782 7. The NON-UNIX-USE file had not been updated for the different set of source
3783 files that come with release 6. I also added a few comments about the C++
3784 wrapper.
3785
3786
3787 Version 6.0 07-Jun-05
3788 ---------------------
3789
3790 1. Some minor internal re-organization to help with my DFA experiments.
3791
3792 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3793 didn't matter for the library itself when fully configured, but did matter
3794 when compiling without UCP support, or within Exim, where the ucp files are
3795 not imported.
3796
3797 3. Refactoring of the library code to split up the various functions into
3798 different source modules. The addition of the new DFA matching code (see
3799 below) to a single monolithic source would have made it really too
3800 unwieldy, quite apart from causing all the code to be include in a
3801 statically linked application, when only some functions are used. This is
3802 relevant even without the DFA addition now that patterns can be compiled in
3803 one application and matched in another.
3804
3805 The downside of splitting up is that there have to be some external
3806 functions and data tables that are used internally in different modules of
3807 the library but which are not part of the API. These have all had their
3808 names changed to start with "_pcre_" so that they are unlikely to clash
3809 with other external names.
3810
3811 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3812 a different (DFA) algorithm. Although it is slower than the original
3813 function, it does have some advantages for certain types of matching
3814 problem.
3815
3816 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3817 including restarting after a partial match.
3818
3819 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3820 defined when compiling for Windows was sent to me. I have put it into the
3821 code, though I have no means of testing or verifying it.
3822
3823 7. Added the pcre_refcount() auxiliary function.
3824
3825 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3826 match before or at the first newline in the subject string. In pcretest,
3827 the /f option on a pattern can be used to set this.
3828
3829 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3830 would behave wrongly. This has been present in PCRE since release 4.0.
3831
3832 10. A number of changes to the pcregrep command:
3833
3834 (a) Refactored how -x works; insert ^(...)$ instead of setting
3835 PCRE_ANCHORED and checking the length, in preparation for adding
3836 something similar for -w.
3837
3838 (b) Added the -w (match as a word) option.
3839
3840 (c) Refactored the way lines are read and buffered so as to have more
3841 than one at a time available.
3842
3843 (d) Implemented a pcregrep test script.
3844
3845 (e) Added the -M (multiline match) option. This allows patterns to match
3846 over several lines of the subject. The buffering ensures that at least
3847 8K, or the rest of the document (whichever is the shorter) is available
3848 for matching (and similarly the previous 8K for lookbehind assertions).
3849
3850 (f) Changed the --help output so that it now says
3851
3852 -w, --word-regex(p)
3853
3854 instead of two lines, one with "regex" and the other with "regexp"
3855 because that confused at least one person since the short forms are the
3856 same. (This required a bit of code, as the output is generated
3857 automatically from a table. It wasn't just a text change.)
3858
3859 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3860 option but starts with a hyphen. Could be a pattern or a path name
3861 starting with a hyphen, for instance.
3862
3863 (h) "-" can be given as a file name to represent stdin.
3864
3865 (i) When file names are being printed, "(standard input)" is used for
3866 the standard input, for compatibility with GNU grep. Previously
3867 "<stdin>" was used.
3868
3869 (j) The option --label=xxx can be used to supply a name to be used for
3870 stdin when file names are being printed. There is no short form.
3871
3872 (k) Re-factored the options decoding logic because we are going to add
3873 two more options that take data. Such options can now be given in four
3874 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3875
3876 (l) Added the -A, -B, and -C options for requesting that lines of context
3877 around matches be printed.
3878
3879 (m) Added the -L option to print the names of files that do not contain
3880 any matching lines, that is, the complement of -l.
3881
3882 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3883 continue to scan other files.
3884
3885 (o) The -s option was incorrectly implemented. For compatibility with other
3886 greps, it now suppresses the error message for a non-existent or non-
3887 accessible file (but not the return code). There is a new option called
3888 -q that suppresses the output of matching lines, which was what -s was
3889 previously doing.
3890
3891 (p) Added --include and --exclude options to specify files for inclusion
3892 and exclusion when recursing.
3893
3894 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3895 Hopefully, it now does.
3896
3897 12. Missing cast in pcre_study().
3898
3899 13. Added an "uninstall" target to the makefile.
3900
3901 14. Replaced "extern" in the function prototypes in Makefile.in with
3902 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3903 world, but is set differently for Windows.
3904
3905 15. Added a second compiling function called pcre_compile2(). The only
3906 difference is that it has an extra argument, which is a pointer to an
3907 integer error code. When there is a compile-time failure, this is set
3908 non-zero, in addition to the error test pointer being set to point to an
3909 error message. The new argument may be NULL if no error number is required
3910 (but then you may as well call pcre_compile(), which is now just a
3911 wrapper). This facility is provided because some applications need a
3912 numeric error indication, but it has also enabled me to tidy up the way
3913 compile-time errors are handled in the POSIX wrapper.
3914
3915 16. Added VPATH=.libs to the makefile; this should help when building with one
3916 prefix path and installing with another. (Or so I'm told by someone who
3917 knows more about this stuff than I do.)
3918
3919 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3920 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3921 match everything, including newlines. This is not POSIX-compatible, but
3922 somebody wanted the feature. From pcretest it can be activated by using
3923 both the P and the s flags.
3924
3925 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3926
3927 19. libpcre.pc was being incorrectly installed as executable.
3928
3929 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3930 it now also looks for '\r' so that it will work unmodified on Windows.
3931
3932 21. Added Google's contributed C++ wrapper to the distribution.
3933
3934 22. Added some untidy missing memory free() calls in pcretest, to keep
3935 Electric Fence happy when testing.
3936
3937
3938
3939 Version 5.0 13-Sep-04
3940 ---------------------
3941
3942 1. Internal change: literal characters are no longer packed up into items
3943 containing multiple characters in a single byte-string. Each character
3944 is now matched using a separate opcode. However, there may be more than one
3945 byte in the character in UTF-8 mode.
3946
3947 2. The pcre_callout_block structure has two new fields: pattern_position and
3948 next_item_length. These contain the offset in the pattern to the next match
3949 item, and its length, respectively.
3950
3951 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3952 insertion of callouts before each pattern item. Added the /C option to
3953 pcretest to make use of this.
3954
3955 4. On the advice of a Windows user, the lines
3956
3957 #if defined(_WIN32) || defined(WIN32)
3958 _setmode( _fileno( stdout ), 0x8000 );
3959 #endif /* defined(_WIN32) || defined(WIN32) */
3960
3961 have been added to the source of pcretest. This apparently does useful
3962 magic in relation to line terminators.
3963
3964 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3965 for the benefit of those environments where the "b" makes a difference.
3966
3967 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3968 to know about it. I have put a hack into configure.in that adds in code
3969 to set GCC=yes if CC=icc. This seems to end up at a point in the
3970 generated configure script that is early enough to affect the setting of
3971 compiler options, which is what is needed, but I have no means of testing
3972 whether it really works. (The user who reported this had patched the
3973 generated configure script, which of course I cannot do.)
3974
3975 LATER: After change 22 below (new libtool files), the configure script
3976 seems to know about icc (and also ecc). Therefore, I have commented out
3977 this hack in configure.in.
3978
3979 7. Added support for pkg-config (2 patches were sent in).
3980
3981 8. Negated POSIX character classes that used a combination of internal tables
3982 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3983 [[:^ascii]]. Typically, they would match almost any characters. The other
3984 POSIX classes were not broken in this way.
3985
3986 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3987 to find the match, as PCRE was deluded into thinking that the match had to
3988 start at the start point or following a newline. The same bug applied to
3989 patterns with negative forward assertions or any backward assertions
3990 preceding ".*" at the start, unless the pattern required a fixed first
3991 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3992
3993 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3994 starting at the last subject character, bytes beyond the end of the subject
3995 string were read.
3996
3997 11. Renamed the variable "class" as "classbits" to make life easier for C++
3998 users. (Previously there was a macro definition, but it apparently wasn't
3999 enough.)
4000
4001 12. Added the new field "tables" to the extra data so that tables can be passed
4002 in at exec time, or the internal tables can be re-selected. This allows
4003 a compiled regex to be saved and re-used at a later time by a different
4004 program that might have everything at different addresses.
4005
4006 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4007 -R library as well as a -L library.
4008
4009 14. The debugging options of pcretest (-d on the command line or D on a
4010 pattern) showed incorrect output for anything following an extended class
4011 that contained multibyte characters and which was followed by a quantifier.
4012
4013 15. Added optional support for general category Unicode character properties
4014 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4015 support. It adds about 90K to the size of the library. The meanings of the
4016 inbuilt class escapes such as \d and \s have NOT been changed.
4017
4018 16. Updated pcredemo.c to include calls to free() to release the memory for the
4019 compiled pattern.
4020
4021 17. The generated file chartables.c was being created in the source directory
4022 instead of in the building directory. This caused the build to fail if the
4023 source directory was different from the building directory, and was
4024 read-only.
4025
4026 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4027 file. No doubt somebody will tell me if they don't make sense... Also added
4028 Dan Mooney's comments about building on OpenVMS.
4029
4030 19. Added support for partial matching via the PCRE_PARTIAL option for
4031 pcre_exec() and the \P data escape in pcretest.
4032
4033 20. Extended pcretest with 3 new pattern features:
4034
4035 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4036 write the compiled pattern to the file whose name is "rest-of-line".
4037 This is a straight binary dump of the data, with the saved pointer to
4038 the character tables forced to be NULL. The study data, if any, is
4039 written too. After writing, pcretest reads a new pattern.
4040
4041 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4042 compiled pattern from the given file. There must not be any
4043 occurrences of "<" in the file name (pretty unlikely); if there are,
4044 pcretest will instead treat the initial "<" as a pattern delimiter.
4045 After reading in the pattern, pcretest goes on to read data lines as
4046 usual.
4047
4048 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4049 and 16-bit fields in a compiled pattern, to simulate a pattern that
4050 was compiled on a host of opposite endianness.
4051
4052 21. The pcre-exec() function can now cope with patterns that were compiled on
4053 hosts of opposite endianness, with this restriction:
4054
4055 As for any compiled expression that is saved and used later, the tables
4056 pointer field cannot be preserved; the extra_data field in the arguments
4057 to pcre_exec() should be used to pass in a tables address if a value
4058 other than the default internal tables were used at compile time.
4059
4060 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4061 now diagnosed as an error. Previously, most of the time, a negative number
4062 would have been treated as zero, but if in addition "ovector" was passed as
4063 NULL, a crash could occur.
4064
4065 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4066 new versions from the libtool 1.5 distribution (the last one is a copy of
4067 a file called libtool.m4). This seems to have fixed the need to patch
4068 "configure" to support Darwin 1.3 (which I used to do). However, I still
4069 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4070 workstation).
4071
4072 24. Changed the PCRE licence to be the more standard "BSD" licence.
4073
4074
4075 Version 4.5 01-Dec-03
4076 ---------------------
4077
4078 1. There has been some re-arrangement of the code for the match() function so
4079 that it can be compiled in a version that does not call itself recursively.
4080 Instead, it keeps those local variables that need separate instances for
4081 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4082 needs to "recurse". Keeping track of where control must go is done by means
4083 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4084 hide most of the details from the main code, and operates only if
4085 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4086 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4087 operating.
4088
4089 To make it easier for callers to provide specially tailored get/free
4090 functions for this usage, two new functions, pcre_stack_malloc, and
4091 pcre_stack_free, are used. They are always called in strict stacking order,
4092 and the size of block requested is always the same.
4093
4094 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4095 PCRE has been compiled to use the stack or the heap for recursion. The
4096 -C option of pcretest uses this to show which version is compiled.
4097
4098 A new data escape \S, is added to pcretest; it causes the amounts of store
4099 obtained and freed by both kinds of malloc/free at match time to be added
4100 to the output.
4101
4102 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4103 what's available on my current Linux desktop machine.
4104
4105 3. When matching a UTF-8 string, the test for a valid string at the start has
4106 been extended. If start_offset is not zero, PCRE now checks that it points
4107 to a byte that is the start of a UTF-8 character. If not, it returns
4108 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4109 this is necessary because there may be backward assertions in the pattern.
4110 When matching the same subject several times, it may save resources to use
4111 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4112
4113 4. The code for checking the validity of UTF-8 strings has been tightened so
4114 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4115 containing "overlong sequences".
4116
4117 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4118 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4119 should have been "&", but it just so happened that all the cases this let
4120 through by mistake were picked up later in the function.
4121
4122 6. I had used a variable called "isblank" - this is a C99 function, causing
4123 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4124
4125 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4126 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4127 so the version is shown; (c) stop "make test" echoing "./RunTest".
4128
4129 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4130
4131 9. The prototype for memmove() for systems that don't have it was using
4132 size_t, but the inclusion of the header that defines size_t was later. I've
4133 moved the #includes for the C headers earlier to avoid this.
4134
4135 10. Added some adjustments to the code to make it easier to compiler on certain
4136 special systems:
4137
4138 (a) Some "const" qualifiers were missing.
4139 (b) Added the macro EXPORT before all exported functions; by default this
4140 is defined to be empty.
4141 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4142 that it reads its output file name as an argument instead of writing
4143 to the standard output and assuming this can be redirected.
4144
4145 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4146 class containing characters with values greater than 255, PCRE compilation
4147 went into a loop.
4148
4149 12. A recursive reference to a subpattern that was within another subpattern
4150 that had a minimum quantifier of zero caused PCRE to crash. For example,
4151 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4152 recursion. If the recursively-called subpattern itself had a zero repeat,
4153 that was OK.
4154
4155 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4156 buffer into which it was copied (for escape processing) was still set at
4157 1024, so long lines caused crashes.
4158
4159 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4160 "internal error: code overflow...". This applied to any character class
4161 that was followed by a possessive quantifier.
4162
4163 15. Modified the Makefile to add libpcre.la as a prerequisite for
4164 libpcreposix.la because I was told this is needed for a parallel build to
4165 work.
4166
4167 16. If a pattern that contained .* following optional items at the start was
4168 studied, the wrong optimizing data was generated, leading to matching
4169 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4170 matching string must start with a or b or c. The correct conclusion for
4171 this pattern is that a match can start with any character.
4172
4173
4174 Version 4.4 13-Aug-03
4175 ---------------------
4176
4177 1. In UTF-8 mode, a character class containing characters with values between
4178 127 and 255 was not handled correctly if the compiled pattern was studied.
4179 In fixing this, I have also improved the studying algorithm for such
4180 classes (slightly).
4181
4182 2. Three internal functions had redundant arguments passed to them. Removal
4183 might give a very teeny performance improvement.
4184
4185 3. Documentation bug: the value of the capture_top field in a callout is *one
4186 more than* the number of the hightest numbered captured substring.
4187
4188 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4189 in incorrectly linking with a previously installed version. They now link
4190 explicitly with libpcre.la.
4191
4192 5. configure.in no longer needs to recognize Cygwin specially.
4193
4194 6. A problem in pcre.in for Windows platforms is fixed.
4195
4196 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4197 pcretest, it used to include the size of the study block as part of its
4198 output. Unfortunately, the structure contains a field that has a different
4199 size on different hardware architectures. This meant that the tests that
4200 showed this size failed. As the block is currently always of a fixed size,
4201 this information isn't actually particularly useful in pcretest output, so
4202 I have just removed it.
4203
4204 8. Three pre-processor statements accidentally did not start in column 1.
4205 Sadly, there are *still* compilers around that complain, even though
4206 standard C has not required this for well over a decade. Sigh.
4207
4208 9. In pcretest, the code for checking callouts passed small integers in the
4209 callout_data field, which is a void * field. However, some picky compilers
4210 complained about the casts involved for this on 64-bit systems. Now
4211 pcretest passes the address of the small integer instead, which should get
4212 rid of the warnings.
4213
4214 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4215 both compile and run time, and gives an error if an invalid UTF-8 sequence
4216 is found. There is a option for disabling this check in cases where the
4217 string is known to be correct and/or the maximum performance is wanted.
4218
4219 11. In response to a bug report, I changed one line in Makefile.in from
4220
4221 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4222 to
4223 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4224
4225 to look similar to other lines, but I have no way of telling whether this
4226 is the right thing to do, as I do not use Windows. No doubt I'll get told
4227 if it's wrong...
4228
4229
4230 Version 4.3 21-May-03
4231 ---------------------
4232
4233 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4234 Makefile.
4235
4236 2. Some refactoring to improve the quality of the code:
4237
4238 (i) The utf8_table... variables are now declared "const".
4239
4240 (ii) The code for \cx, which used the "case flipping" table to upper case
4241 lower case letters, now just substracts 32. This is ASCII-specific,
4242 but the whole concept of \cx is ASCII-specific, so it seems
4243 reasonable.
4244
4245 (iii) PCRE was using its character types table to recognize decimal and
4246 hexadecimal digits in the pattern. This is silly, because it handles
4247 only 0-9, a-f, and A-F, but the character types table is locale-
4248 specific, which means strange things might happen. A private
4249 table is now used for this - though it costs 256 bytes, a table is
4250 much faster than multiple explicit tests. Of course, the standard
4251 character types table is still used for matching digits in subject
4252 strings against \d.
4253
4254 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4255 ending in _t are). So I've renamed it as ESC_tee.
4256
4257 3. The first argument for regexec() in the POSIX wrapper should have been
4258 defined as "const".
4259
4260 4. Changed pcretest to use malloc() for its buffers so that they can be
4261 Electric Fenced for debugging.
4262
4263 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4264 to read one or more bytes before the start of the subject string. Often this
4265 had no effect on PCRE's behaviour, but in some circumstances it could
4266 provoke a segmentation fault.
4267
4268 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4269 to try to read one or more bytes before the start of the subject string.
4270
4271 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4272 UTF-8 support could misbehave in various ways if the subject string
4273 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4274 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4275 back over UTF-8 characters.)
4276
4277
4278 Version 4.2 14-Apr-03
4279 ---------------------
4280
4281 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4282
4283 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4284 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4285 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4286 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4287 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4288 and BUILD_EXEEXT
4289 Note: automatic setting of the BUILD variables is not yet working
4290 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4291 compile-time but not at link-time
4292 [LINK]: use for linking executables only
4293 make different versions for Windows and non-Windows
4294 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4295 libraries
4296 [LINK_FOR_BUILD]: new variable
4297 [OBJEXT]: use throughout
4298 [EXEEXT]: use throughout
4299 <winshared>: new target
4300 <wininstall>: new target
4301 <dftables.o>: use native compiler
4302 <dftables>: use native linker
4303 <install>: handle Windows platform correctly
4304 <clean>: ditto
4305 <check>: ditto
4306 copy DLL to top builddir before testing
4307
4308 As part of these changes, -no-undefined was removed again. This was reported
4309 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4310 in any case.
4311
4312 3. Some tidies to get rid of compiler warnings:
4313
4314 . In the match_data structure, match_limit was an unsigned long int, whereas
4315 match_call_count was an int. I've made them both unsigned long ints.
4316
4317 . In pcretest the fact that a const uschar * doesn't automatically cast to
4318 a void * provoked a warning.
4319
4320 . Turning on some more compiler warnings threw up some "shadow" variables
4321 and a few more missing casts.
4322
4323 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4324 option, a class that contained a single character with a value between 128
4325 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4326
4327 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4328 option, a class that contained several characters, but with at least one
4329 whose value was between 128 and 255 caused PCRE to crash.
4330
4331
4332 Version 4.1 12-Mar-03
4333 ---------------------
4334
4335 1. Compiling with gcc -pedantic found a couple of places where casts were
4336 needed, and a string in dftables.c that was longer than standard compilers are
4337 required to support.
4338
4339 2. Compiling with Sun's compiler found a few more places where the code could
4340 be tidied up in order to avoid warnings.
4341
4342 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4343 first of these names is deprecated in the latest Autoconf in favour of the name
4344 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4345 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4346 analogy I have changed it to CFLAGS_FOR_BUILD.
4347
4348 4. Added -no-undefined to the linking command in the Makefile, because this is
4349 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4350 linking step for the pcreposix library.
4351
4352 5. PCRE was failing to diagnose the case of two named groups with the same
4353 name.
4354
4355 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4356 literal character that is needed in the subject for a match, and scans along to
4357 ensure that it is present before embarking on the full matching process. This
4358 saves time in cases of nested unlimited repeats that are never going to match.
4359 Problem: the scan can take a lot of time if the subject is very long (e.g.
4360 megabytes), thus penalizing straightforward matches. It is now done only if the
4361 amount of subject to be scanned is less than 1000 bytes.
4362
4363 7. A lesser problem with the same optimization is that it was recording the
4364 first character of an anchored pattern as "needed", thus provoking a search
4365 right along the subject, even when the first match of the pattern was going to
4366 fail. The "needed" character is now not set for anchored patterns, unless it
4367 follows something in the pattern that is of non-fixed length. Thus, it still
4368 fulfils its original purpose of finding quick non-matches in cases of nested
4369 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4370
4371
4372 Version 4.0 17-Feb-03
4373 ---------------------
4374
4375 1. If a comment in an extended regex that started immediately after a meta-item
4376 extended to the end of string, PCRE compiled incorrect data. This could lead to
4377 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4378
4379 2. Moved to autoconf 2.53 and libtool 1.4.2.
4380
4381 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4382 the special perltest8 script is no longer needed - all the tests can be run
4383 from a single perltest script.
4384
4385 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4386 by \s. It has now been removed in PCRE. This means it isn't recognized as
4387 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4388 class [:space:] *does* include VT, thereby creating a mess.
4389
4390 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4391 space and tab.
4392
4393 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4394 its new features into the main test script, reducing the number of scripts.
4395
4396 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4397 were backward compatible, and made the (?i) apply to the whole pattern, as if
4398 /i were given. Now it behaves more logically, and applies the option setting
4399 only to what follows. PCRE has been changed to follow suit. However, if it
4400 finds options settings right at the start of the pattern, it extracts them into
4401 the global options, as before. Thus, they show up in the info data.
4402
4403 8. Added support for the \Q...\E escape sequence. Characters in between are
4404 treated as literals. This is slightly different from Perl in that $ and @ are
4405 also handled as literals inside the quotes. In Perl, they will cause variable
4406 interpolation. Note the following examples:
4407
4408 Pattern PCRE matches Perl matches
4409
4410 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4411 \Qabc\$xyz\E abc\$xyz abc\$xyz
4412 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4413
4414 For compatibility with Perl, \Q...\E sequences are recognized inside character
4415 classes as well as outside them.
4416
4417 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4418 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4419 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4420 signed/unsigned warnings.
4421
4422 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4423 option for pcretest, so I've replaced it by a simple function that does just
4424 that job.
4425
4426 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4427 "pcregrep -".
4428
4429 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4430 Java package. This provides some syntactic sugar for simple cases of what my
4431 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4432 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4433 item, you can use this simplified notation. Note that only makes sense with
4434 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4435 greediness, whatever the setting of the PCRE_UNGREEDY option.
4436
4437 13. A change of greediness default within a pattern was not taking effect at
4438 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4439 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4440 was abstracted outside.
4441
4442 14. PCRE now supports the \G assertion. It is true when the current matching
4443 position is at the start point of the match. This differs from \A when the
4444 starting offset is non-zero. Used with the /g option of pcretest (or similar
4445 code), it works in the same way as it does for Perl's /g option. If all
4446 alternatives of a regex begin with \G, the expression is anchored to the start
4447 match position, and the "anchored" flag is set in the compiled expression.
4448
4449 15. Some bugs concerning the handling of certain option changes within patterns
4450 have been fixed. These applied to options other than (?ims). For example,
4451 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4452 been the other way round. Some of this was related to change 7 above.
4453
4454 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4455 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4456 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4457 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4458
4459 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4460 mode. Unlike ".", it always matches newline, whatever the setting of
4461 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4462 assertions. Perl allows it, but it doesn't (in general) work because it can't
4463 calculate the length of the lookbehind. At least, that's the case for Perl
4464 5.8.0 - I've been told they are going to document that it doesn't work in
4465 future.
4466
4467 18. Added an error diagnosis for escapes that PCRE does not support: these are
4468 \L, \l, \N, \P, \p, \U, \u, and \X.
4469
4470 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4471 reading past the end of the pattern in cases such as /[abcd/.
4472
4473 20. PCRE was getting more memory than necessary for patterns with classes that
4474 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4475
4476 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4477 compiling PCRE for use with Virtual Pascal.
4478
4479 22. Small fix to the Makefile to make it work properly if the build is done
4480 outside the source tree.
4481
4482 23. Added a new extension: a condition to go with recursion. If a conditional
4483 subpattern starts with (?(R) the "true" branch is used if recursion has
4484 happened, whereas the "false" branch is used only at the top level.
4485
4486 24. When there was a very long string of literal characters (over 255 bytes
4487 without UTF support, over 250 bytes with UTF support), the computation of how
4488 much memory was required could be incorrect, leading to segfaults or other
4489 strange effects.
4490
4491 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4492 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4493 there was a subsequent back reference to those brackets. This meant that, for
4494 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4495 possible to check for precisely this case. All we can do is abandon the
4496 optimization if .* occurs inside capturing brackets when there are any back
4497 references whatsoever. (See below for a better fix that came later.)
4498
4499 26. The handling of the optimization for finding the first character of a
4500 non-anchored pattern, and for finding a character that is required later in the
4501 match were failing in some cases. This didn't break the matching; it just
4502 failed to optimize when it could. The way this is done has been re-implemented.
4503
4504 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4505
4506 28. Added a new feature that provides some of the functionality that Perl
4507 provides with (?{...}). The facility is termed a "callout". The way it is done
4508 in PCRE is for the caller to provide an optional function, by setting
4509 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4510 global variable. By default it is unset, which disables all calling out. To get
4511 the function called, the regex must include (?C) at appropriate points. This
4512 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4513 This provides a means of identifying different callout points. When PCRE
4514 reaches such a point in the regex, if pcre_callout has been set, the external
4515 function is called. It is provided with data in a structure called
4516 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4517 matching continues; if it returns a non-zero value, the match at the current
4518 point fails. However, backtracking will occur if possible. [This was changed
4519 later and other features added - see item 49 below.]
4520
4521 29. pcretest is upgraded to test the callout functionality. It provides a
4522 callout function that displays information. By default, it shows the start of
4523 the match and the current position in the text. There are some new data escapes
4524 to vary what happens:
4525
4526 \C+ in addition, show current contents of captured substrings
4527 \C- do not supply a callout function
4528 \C!n return 1 when callout number n is reached
4529 \C!n!m return 1 when callout number n is reached for the mth time
4530
4531 30. If pcregrep was called with the -l option and just a single file name, it
4532 output "<stdin>" if a match was found, instead of the file name.
4533
4534 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4535 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4536 pcre_exec(). This saves a malloc/free per call. The default value of
4537 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4538 when configuring.
4539
4540 32. The default maximum size of a compiled pattern is 64K. There have been a
4541 few cases of people hitting this limit. The code now uses macros to handle the
4542 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4543 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4544 configuring. Tests 2 and 5 work only with 2-byte links because they output
4545 debugging information about compiled patterns.
4546
4547 33. Internal code re-arrangements:
4548
4549 (a) Moved the debugging function for printing out a compiled regex into
4550 its own source file (printint.c) and used #include to pull it into
4551 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4552 separate copies.
4553
4554 (b) Defined the list of op-code names for debugging as a macro in
4555 internal.h so that it is next to the definition of the opcodes.
4556
4557 (c) Defined a table of op-code lengths for simpler skipping along compiled
4558 code. This is again a macro in internal.h so that it is next to the
4559 definition of the opcodes.
4560
4561 34. Added support for recursive calls to individual subpatterns, along the
4562 lines of Robin Houston's patch (but implemented somewhat differently).
4563
4564 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4565 allow it to read and process whole directories in Win32. This code was
4566 contributed by Lionel Fourquaux; it has not been tested by me.
4567
4568 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4569 used to name a group. Names consist of alphanumerics and underscores, and must
4570 be unique. Back references use the syntax (?P=name) and recursive calls use
4571 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4572 numbers. The function pcre_fullinfo() can be used after compilation to extract
4573 a name/number map. There are three relevant calls:
4574
4575 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4576 PCRE_INFO_NAMECOUNT yields the number of entries
4577 PCRE_INFO_NAMETABLE yields a pointer to the map.
4578
4579 The map is a vector of fixed-size entries. The size of each entry depends on
4580 the length of the longest name used. The first two bytes of each entry are the
4581 group number, most significant byte first. There follows the corresponding
4582 name, zero terminated. The names are in alphabetical order.
4583
4584 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4585 case instead of 255. Making it the same both with and without UTF-8 support
4586 means that the same test output works with both.
4587
4588 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4589 calling malloc() with a zero argument.
4590
4591 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4592 optimization. I've improved things by keeping a bitmap of backreferences with
4593 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4594 fact referenced, the optimization can be applied. It is unlikely that a
4595 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4596 the match to follow \n) will appear inside brackets with a number greater than
4597 31, but if it does, any back reference > 31 suppresses the optimization.
4598
4599 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4600 of disabling numbered capturing parentheses. Any opening parenthesis that is
4601 not followed by ? behaves as if it were followed by ?: but named parentheses
4602 can still be used for capturing (and they will acquire numbers in the usual
4603 way).
4604
4605 41. Redesigned the return codes from the match() function into yes/no/error so
4606 that errors can be passed back from deep inside the nested calls. A malloc
4607 failure while inside a recursive subpattern call now causes the
4608 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4609
4610 42. It is now possible to set a limit on the number of times the match()
4611 function is called in a call to pcre_exec(). This facility makes it possible to
4612 limit the amount of recursion and backtracking, though not in a directly
4613 obvious way, because the match() function is used in a number of different
4614 circumstances. The count starts from zero for each position in the subject
4615 string (for non-anchored patterns). The default limit is, for compatibility, a
4616 large number, namely 10 000 000. You can change this in two ways:
4617
4618 (a) When configuring PCRE before making, you can use --with-match-limit=n
4619 to set a default value for the compiled library.
4620
4621 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4622 a different value is set. See 45 below.
4623
4624 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4625
4626 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4627 of things that can be changed at compile time. The first argument specifies
4628 what is wanted and the second points to where the information is to be placed.
4629 The current list of available information is:
4630
4631 PCRE_CONFIG_UTF8
4632
4633 The output is an integer that is set to one if UTF-8 support is available;
4634 otherwise it is set to zero.
4635
4636 PCRE_CONFIG_NEWLINE
4637
4638 The output is an integer that it set to the value of the code that is used for
4639 newline. It is either LF (10) or CR (13).
4640
4641 PCRE_CONFIG_LINK_SIZE
4642
4643 The output is an integer that contains the number of bytes used for internal
4644 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4645
4646 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4647
4648 The output is an integer that contains the threshold above which the POSIX
4649 interface uses malloc() for output vectors. See item 31 above.
4650
4651 PCRE_CONFIG_MATCH_LIMIT
4652
4653 The output is an unsigned integer that contains the default limit of the number
4654 of match() calls in a pcre_exec() execution. See 42 above.
4655
4656 44. pcretest has been upgraded by the addition of the -C option. This causes it
4657 to extract all the available output from the new pcre_config() function, and to
4658 output it. The program then exits immediately.
4659
4660 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4661 order to support additional features. One way would have been to define
4662 pcre_exec2() (for example) with extra arguments, but this would not have been
4663 extensible, and would also have required all calls to the original function to
4664 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4665 is used for passing in "extra" data from pcre_study().
4666
4667 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4668 contains the following fields:
4669
4670 flags a bitmap indicating which of the following fields are set
4671 study_data opaque data from pcre_study()
4672 match_limit a way of specifying a limit on match() calls for a specific
4673 call to pcre_exec()
4674 callout_data data for callouts (see 49 below)
4675
4676 The flag bits are also defined in pcre.h, and are
4677
4678 PCRE_EXTRA_STUDY_DATA
4679 PCRE_EXTRA_MATCH_LIMIT
4680 PCRE_EXTRA_CALLOUT_DATA
4681
4682 The pcre_study() function now returns one of these new pcre_extra blocks, with
4683 the actual study data pointed to by the study_data field, and the
4684 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4685 before. That is, this change is entirely upwards-compatible and requires no
4686 change to existing code.
4687
4688 If you want to pass in additional data to pcre_exec(), you can either place it
4689 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4690 block.
4691
4692 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4693 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4694 times with different match limits, until it finds the minimum value needed for
4695 pcre_exec() to complete. The value is then output. This can be instructive; for
4696 most simple matches the number is quite small, but for pathological cases it
4697 gets very large very quickly.
4698
4699 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4700 returns the size of the data block pointed to by the study_data field in a
4701 pcre_extra block, that is, the value that was passed as the argument to
4702 pcre_malloc() when PCRE was getting memory in which to place the information
4703 created by pcre_study(). The fourth argument should point to a size_t variable.
4704 pcretest has been extended so that this information is shown after a successful
4705 pcre_study() call when information about the compiled regex is being displayed.
4706
4707 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4708 because what follows is always an absolute path. (Later: it turns out that this
4709 is more than cosmetic for MinGW, because it doesn't like empty path
4710 components.)
4711
4712 49. Some changes have been made to the callout feature (see 28 above):
4713
4714 (i) A callout function now has three choices for what it returns:
4715
4716 0 => success, carry on matching
4717 > 0 => failure at this point, but backtrack if possible
4718 < 0 => serious error, return this value from pcre_exec()
4719
4720 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4721 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4722 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4723 use by callout functions. It will never be used by PCRE itself.
4724
4725 (ii) The pcre_extra structure (see 45 above) has a void * field called
4726 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4727 pcre_callout_block structure has a field of the same name. The contents of
4728 the field passed in the pcre_extra structure are passed to the callout
4729 function in the corresponding field in the callout block. This makes it
4730 easier to use the same callout-containing regex from multiple threads. For
4731 testing, the pcretest program has a new data escape
4732
4733 \C*n pass the number n (may be negative) as callout_data
4734
4735 If the callout function in pcretest receives a non-zero value as
4736 callout_data, it returns that value.
4737
4738 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4739 there were some redundant $(CFLAGS) in commands that are now specified as
4740 $(LINK), which already includes $(CFLAGS).
4741
4742 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4743 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4744 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4745 one-byte characters throughout. Note that case-insensitive matching applies
4746 only to characters whose values are less than 256. PCRE doesn't support the
4747 notion of cases for higher-valued characters.
4748
4749 (i) A character class whose characters are all within 0-255 is handled as
4750 a bit map, and the map is inverted for negative classes. Previously, a
4751 character > 255 always failed to match such a class; however it should
4752 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4753
4754 (ii) A negated character class with a single character < 255 is coded as
4755 "not this character" (OP_NOT). This wasn't working properly when the test
4756 character was multibyte, either singly or repeated.
4757
4758 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4759 mode, for example: \x{100}{2,3}.
4760
4761 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4762 singly or repeated) now correctly test multibyte characters. However,
4763 PCRE doesn't recognize any characters with values greater than 255 as
4764 digits, spaces, or word characters. Such characters always match \D, \S,
4765 and \W, and never match \d, \s, or \w.
4766
4767 (v) Classes may now contain characters and character ranges with values
4768 greater than 255. For example: [ab\x{100}-\x{400}].
4769
4770 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4771 PCRE in UTF-8 mode.
4772
4773 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4774 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4775 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4776 value.)
4777
4778 53. The single man page has become too large. I have therefore split it up into
4779 a number of separate man pages. These also give rise to individual HTML pages;
4780 these are now put in a separate directory, and there is an index.html page that
4781 lists them all. Some hyperlinking between the pages has been installed.
4782
4783 54. Added convenience functions for handling named capturing parentheses.
4784
4785 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4786 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4787 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4788 are faulted.
4789
4790 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4791 calling configure. These values are used when compiling the dftables.c program
4792 which is run to generate the source of the default character tables. They
4793 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4794 you will need to set these values.
4795
4796 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4797
4798
4799 Version 3.9 02-Jan-02
4800 ---------------------
4801
4802 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4803
4804 2. If --disable-static was given, the building process failed when trying to
4805 build pcretest and pcregrep. (For some reason it was using libtool to compile
4806 them, which is not right, as they aren't part of the library.)
4807
4808
4809 Version 3.8 18-Dec-01
4810 ---------------------
4811
4812 1. The experimental UTF-8 code was completely screwed up. It was packing the
4813 bytes in the wrong order. How dumb can you get?
4814
4815
4816 Version 3.7 29-Oct-01
4817 ---------------------
4818
4819 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4820 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4821 this didn't happen under Solaris 8, where I normally test things.
4822
4823 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4824 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4825 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4826 this fix an hour or so after the initial 3.7 release.)
4827
4828
4829 Version 3.6 23-Oct-01
4830 ---------------------
4831
4832 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4833 offsets passed as NULL with zero offset count.
4834
4835 2. The config.guess and config.sub files had not been updated when I moved to
4836 the latest autoconf.
4837
4838
4839 Version 3.5 15-Aug-01
4840 ---------------------
4841
4842 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4843 had been forgotten.
4844
4845 2. By using declared but undefined structures, we can avoid using "void"
4846 definitions in pcre.h while keeping the internal definitions of the structures
4847 private.
4848
4849 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4850 user point of view, this means that both static and shared libraries are built
4851 by default, but this can be individually controlled. More of the work of
4852 handling this static/shared cases is now inside libtool instead of PCRE's make
4853 file.
4854
4855 4. The pcretest utility is now installed along with pcregrep because it is
4856 useful for users (to test regexs) and by doing this, it automatically gets
4857 relinked by libtool. The documentation has been turned into a man page, so
4858 there are now .1, .txt, and .html versions in /doc.
4859
4860 5. Upgrades to pcregrep:
4861 (i) Added long-form option names like gnu grep.
4862 (ii) Added --help to list all options with an explanatory phrase.
4863 (iii) Added -r, --recursive to recurse into sub-directories.
4864 (iv) Added -f, --file to read patterns from a file.
4865
4866 6. pcre_exec() was referring to its "code" argument before testing that
4867 argument for NULL (and giving an error if it was NULL).
4868
4869 7. Upgraded Makefile.in to allow for compiling in a different directory from
4870 the source directory.
4871
4872 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4873 options bits, the pointer it was passed was to an int instead of to an unsigned
4874 long int. This mattered only on 64-bit systems.
4875
4876 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4877 generated) instead of pcre.in, which it its source. Also made the same change
4878 in several of the .c files.
4879
4880 10. A new release of gcc defines printf() as a macro, which broke pcretest
4881 because it had an ifdef in the middle of a string argument for printf(). Fixed
4882 by using separate calls to printf().
4883
4884 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4885 script, to force use of CR or LF instead of \n in the source. On non-Unix
4886 systems, the value can be set in config.h.
4887
4888 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4889 absolute limit. Changed the text of the error message to make this clear, and
4890 likewise updated the man page.
4891
4892 13. The limit of 99 on the number of capturing subpatterns has been removed.
4893 The new limit is 65535, which I hope will not be a "real" limit.
4894
4895
4896 Version 3.4 22-Aug-00
4897 ---------------------
4898
4899 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4900
4901 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4902
4903
4904 Version 3.3 01-Aug-00
4905 ---------------------
4906
4907 1. If an octal character was given, but the value was greater than \377, it
4908 was not getting masked to the least significant bits, as documented. This could
4909 lead to crashes in some systems.
4910
4911 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4912 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4913
4914 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4915 These just pass their arguments on to (pcre_free)(), but they are provided
4916 because some uses of PCRE bind it to non-C systems that can call its functions,
4917 but cannot call free() or pcre_free() directly.
4918
4919 4. Add "make test" as a synonym for "make check". Corrected some comments in
4920 the Makefile.
4921
4922 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4923 Makefile.
4924
4925 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4926 command called pgrep for grepping around the active processes.
4927
4928 7. Added the beginnings of support for UTF-8 character strings.
4929
4930 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4931 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4932 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4933 out for the ar command.)
4934
4935
4936 Version 3.2 12-May-00
4937 ---------------------
4938
4939 This is purely a bug fixing release.
4940
4941 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4942 of ZA. This was just one example of several cases that could provoke this bug,
4943 which was introduced by change 9 of version 2.00. The code for breaking
4944 infinite loops after an iteration that matches an empty string was't working
4945 correctly.
4946
4947 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4948 when matched against abbab (for example). After matching an empty string, it
4949 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4950 caused it to match further down the string than it should.
4951
4952 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4953 was there because it doesn't seem to be needed, and it causes trouble on some
4954 systems, as it is not a Standard C header. It has been removed.
4955
4956 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4957 were reported on the Macintosh. The changes were from
4958
4959 while ((c = *(++ptr)) != 0 && c != '\n');
4960 to
4961 while ((c = *(++ptr)) != 0 && c != '\n') ;
4962
4963 Totally extraordinary, but if that's what it takes...
4964
4965 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4966 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4967 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4968 assumes the way PCRE uses memmove() (always moving upwards).
4969
4970 6. PCRE is being used in one environment where strchr() is not available. There
4971 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4972 faster code anyway.
4973
4974
4975 Version 3.1 09-Feb-00
4976 ---------------------
4977
4978 The only change in this release is the fixing of some bugs in Makefile.in for
4979 the "install" target:
4980
4981 (1) It was failing to install pcreposix.h.
4982
4983 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4984
4985
4986 Version 3.0 01-Feb-00
4987 ---------------------
4988
4989 1. Add support for the /+ modifier to perltest (to output $` like it does in
4990 pcretest).
4991
4992 2. Add support for the /g modifier to perltest.
4993
4994 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4995 matches null strings.
4996
4997 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4998 pattern. Perl treats empty patterns specially - it reuses the most recent
4999 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5000 effect.
5001
5002 5. The POSIX interface was broken in that it was just handing over the POSIX
5003 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5004 required a bigger vector, with some working space on the end. This means that
5005 the POSIX wrapper now has to get and free some memory, and copy the results.
5006
5007 6. Added some simple autoconf support, placing the test data and the
5008 documentation in separate directories, re-organizing some of the
5009 information files, and making it build pcre-config (a GNU standard). Also added
5010 libtool support for building PCRE as a shared library, which is now the
5011 default.
5012
5013 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5014 09 are not valid octal constants. Single digits will be used for minor values
5015 less than 10.
5016
5017 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5018 existing programs that set these in the POSIX interface can use PCRE without
5019 modification.
5020
5021 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5022 return all that pcre_info() returns, plus additional data. The pcre_info()
5023 function is retained for compatibility, but is considered to be obsolete.
5024
5025 10. Added experimental recursion feature (?R) to handle one common case that
5026 Perl 5.6 will be able to do with (?p{...}).
5027
5028 11. Added support for POSIX character classes like [:alpha:], which Perl is
5029 adopting.
5030
5031
5032 Version 2.08 31-Aug-99
5033 ----------------------
5034
5035 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5036 trying to match at the startoffset position, but instead was moving forward to
5037 the next newline as if a previous match had failed.
5038
5039 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5040 and could get into a loop if a null string was matched other than at the start
5041 of the subject.
5042
5043 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5044 be distinguished at compile time, and for completeness also added PCRE_DATE.
5045
5046 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5047 in GnuWin32 environments.
5048
5049
5050 Version 2.07 29-Jul-99
5051 ----------------------
5052
5053 1. The documentation is now supplied in plain text form and HTML as well as in
5054 the form of man page sources.
5055
5056 2. C++ compilers don't like assigning (void *) values to other pointer types.
5057 In particular this affects malloc(). Although there is no problem in Standard
5058 C, I've put in casts to keep C++ compilers happy.
5059
5060 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5061 should be (const char *).
5062
5063 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5064 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5065 However, I haven't made this a standard facility. The documentation doesn't
5066 mention it, and the Makefile doesn't support it.
5067
5068 5. The Makefile now contains an "install" target, with editable destinations at
5069 the top of the file. The pcretest program is not installed.
5070
5071 6. pgrep -V now gives the PCRE version number and date.
5072
5073 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5074 causing the entire string to be ignored, instead of just the last character.
5075
5076 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5077 non-matching string, it can take a very, very long time, even for strings of
5078 quite modest length, because of the nested recursion. PCRE now does better in
5079 some of these cases. It does this by remembering the last required literal
5080 character in the pattern, and pre-searching the subject to ensure it is present
5081 before running the real match. In other words, it applies a heuristic to detect
5082 some types of certain failure quickly, and in the above example, if presented
5083 with a string that has no trailing " it gives "no match" very quickly.
5084
5085 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5086 other alternatives are tried instead.
5087
5088
5089 Version 2.06 09-Jun-99
5090 ----------------------
5091
5092 1. Change pcretest's output for amount of store used to show just the code
5093 space, because the remainder (the data block) varies in size between 32-bit and
5094 64-bit systems.
5095
5096 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5097 start matching at. This allows lookbehinds to work when searching for multiple
5098 occurrences in a string.
5099
5100 3. Added additional options to pcretest for testing multiple occurrences:
5101
5102 /+ outputs the rest of the string that follows a match
5103 /g loops for multiple occurrences, using the new startoffset argument
5104 /G loops for multiple occurrences by passing an incremented pointer
5105
5106 4. PCRE wasn't doing the "first character" optimization for patterns starting
5107 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5108 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5109 the letter 'x'. On long subject strings, this gives a significant speed-up.
5110
5111
5112 Version 2.05 21-Apr-99
5113 ----------------------
5114
5115 1. Changed the type of magic_number from int to long int so that it works
5116 properly on 16-bit systems.
5117
5118 2. Fixed a bug which caused patterns starting with .* not to work correctly
5119 when the subject string contained newline characters. PCRE was assuming
5120 anchoring for such patterns in all cases, which is not correct because .* will
5121 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5122 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5123 must be retried after every newline in the subject.
5124
5125
5126 Version 2.04 18-Feb-99
5127 ----------------------
5128
5129 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5130 computation of the store needed to hold the pattern was incorrect (too large).
5131 If such patterns were nested a few deep, this could multiply and become a real
5132 problem.
5133
5134 2. Added /M option to pcretest to show the memory requirement of a specific
5135 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5136
5137 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5138 compiled in such a way that the backtracking after subsequent failure was
5139 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5140 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5141
5142
5143 Version 2.03 02-Feb-99
5144 ----------------------
5145
5146 1. Fixed typo and small mistake in man page.
5147
5148 2. Added 4th condition (GPL supersedes if conflict) and created separate
5149 LICENCE file containing the conditions.
5150
5151 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5152 Perl, that is the internal \ allows the delimiter to be included in the
5153 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5154 the final delimiter, add \ to the end of the pattern (to test the error).
5155
5156 4. Added the convenience functions for extracting substrings after a successful
5157 match. Updated pcretest to make it able to test these functions.
5158
5159
5160 Version 2.02 14-Jan-99
5161 ----------------------
5162
5163 1. Initialized the working variables associated with each extraction so that
5164 their saving and restoring doesn't refer to uninitialized store.
5165
5166 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5167 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5168 fix the problem.
5169
5170 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5171 calls, and wasn't printing the correct value for compiling calls. Increased the
5172 default value of LOOPREPEAT, and the number of significant figures in the
5173 times.
5174
5175 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5176
5177 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5178 a building problem on Windows NT with a FAT file system.
5179
5180
5181 Version 2.01 21-Oct-98
5182 ----------------------
5183
5184 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5185 to character tables built by pcre_maketables() in the current locale. If NULL
5186 is passed, the default tables are used.
5187
5188
5189 Version 2.00 24-Sep-98
5190 ----------------------
5191
5192 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5193 it any more.
5194
5195 2. Allow quantification of (?>) groups, and make it work correctly.
5196
5197 3. The first character computation wasn't working for (?>) groups.
5198
5199 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5200 end of the subject) and add 5.005's \z, which really does match only at the
5201 very end of the subject.
5202
5203 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5204
5205 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5206 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5207 localized options. All options to pcre_study() were also removed.
5208
5209 7. Add other new features from 5.005:
5210
5211 $(?<= positive lookbehind
5212 $(?<! negative lookbehind
5213 (?imsx-imsx) added the unsetting capability
5214 such a setting is global if at outer level; local otherwise
5215 (?imsx-imsx:) non-capturing groups with option setting
5216 (?(cond)re|re) conditional pattern matching
5217
5218 A backreference to itself in a repeated group matches the previous
5219 captured string.
5220
5221 8. General tidying up of studying (both automatic and via "study")
5222 consequential on the addition of new assertions.
5223
5224 9. As in 5.005, unlimited repeated groups that could match an empty substring
5225 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5226 runtime if any iteration does actually match an empty substring.
5227
5228 10. Include the RunTest script in the distribution.
5229
5230 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5231 discrepancies, some of which were old and were also with respect to 5.004. They
5232 have now been fixed.
5233
5234
5235 Version 1.09 28-Apr-98
5236 ----------------------
5237
5238 1. A negated single character class followed by a quantifier with a minimum
5239 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5240 program crashes, or just wrong answers. This did not apply to negated classes
5241 containing more than one character, or to minima other than one.
5242
5243
5244 Version 1.08 27-Mar-98
5245 ----------------------
5246
5247 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5248
5249 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5250 latter must appear before anything that relies on it in the pattern.
5251
5252
5253 Version 1.07 16-Feb-98
5254 ----------------------
5255
5256 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5257 repeat of a potentially empty string).
5258
5259
5260 Version 1.06 23-Jan-98
5261 ----------------------
5262
5263 1. Added Markus Oberhumer's little patches for C++.
5264
5265 2. Literal strings longer than 255 characters were broken.
5266
5267
5268 Version 1.05 23-Dec-97
5269 ----------------------
5270
5271 1. Negated character classes containing more than one character were failing if
5272 PCRE_CASELESS was set at run time.
5273
5274
5275 Version 1.04 19-Dec-97
5276 ----------------------
5277
5278 1. Corrected the man page, where some "const" qualifiers had been omitted.
5279
5280 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5281 input syntax.
5282
5283 3. Fixed memory leak which occurred when a regex with back references was
5284 matched with an offsets vector that wasn't big enough. The temporary memory
5285 that is used in this case wasn't being freed if the match failed.
5286
5287 4. Tidied pcretest to ensure it frees memory that it gets.
5288
5289 5. Temporary memory was being obtained in the case where the passed offsets
5290 vector was exactly big enough.
5291
5292 6. Corrected definition of offsetof() from change 5 below.
5293
5294 7. I had screwed up change 6 below and broken the rules for the use of
5295 setjmp(). Now fixed.
5296
5297
5298 Version 1.03 18-Dec-97
5299 ----------------------
5300
5301 1. A erroneous regex with a missing opening parenthesis was correctly
5302 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5303 on some systems.
5304
5305 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5306 it was reported that one broken compiler failed on the former because "code" is
5307 also an independent variable.
5308
5309 3. The erroneous regex a[]b caused an array overrun reference.
5310
5311 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5312 fail on data ending with that character. (It was going on too far, and checking
5313 the next character, typically a binary zero.) This was specific to the
5314 optimized code for single-character negative classes.
5315
5316 5. Added a contributed patch from the TIN world which does the following:
5317
5318 + Add an undef for memmove, in case the the system defines a macro for it.
5319
5320 + Add a definition of offsetof(), in case there isn't one. (I don't know
5321 the reason behind this - offsetof() is part of the ANSI standard - but
5322 it does no harm).
5323
5324 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5325 most of the places where whitespace preceded '#'. I have given up and
5326 allowed the remaining 2 cases to be at the margin.
5327
5328 + Rename some variables in pcre to eliminate shadowing. This seems very
5329 pedantic, but does no harm, of course.
5330
5331 6. Moved the call to setjmp() into its own function, to get rid of warnings
5332 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5333
5334 7. Constructs such as \d{8,} were compiling into the equivalent of
5335 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5336 outcome, but in this particular case used more store than had been allocated,
5337 which caused the bug to be discovered because it threw up an internal error.
5338
5339 8. The debugging code in both pcre and pcretest for outputting the compiled
5340 form of a regex was going wrong in the case of back references followed by
5341 curly-bracketed repeats.
5342
5343
5344 Version 1.02 12-Dec-97
5345 ----------------------
5346
5347 1. Typos in pcre.3 and comments in the source fixed.
5348
5349 2. Applied a contributed patch to get rid of places where it used to remove
5350 'const' from variables, and fixed some signed/unsigned and uninitialized
5351 variable warnings.
5352
5353 3. Added the "runtest" target to Makefile.
5354
5355 4. Set default compiler flag to -O2 rather than just -O.
5356
5357
5358 Version 1.01 19-Nov-97
5359 ----------------------
5360
5361 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5362 like /([ab]*)*/, that is, for classes with more than one character in them.
5363
5364 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5365 as /((?>a*))*/ (a PCRE_EXTRA facility).
5366
5367
5368 Version 1.00 18-Nov-97
5369 ----------------------
5370
5371 1. Added compile-time macros to support systems such as SunOS4 which don't have
5372 memmove() or strerror() but have other things that can be used instead.
5373
5374 2. Arranged that "make clean" removes the executables.
5375
5376
5377 Version 0.99 27-Oct-97
5378 ----------------------
5379
5380 1. Fixed bug in code for optimizing classes with only one character. It was
5381 initializing a 32-byte map regardless, which could cause it to run off the end
5382 of the memory it had got.
5383
5384 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5385
5386
5387 Version 0.98 22-Oct-97
5388 ----------------------
5389
5390 1. Fixed bug in code for handling temporary memory usage when there are more
5391 back references than supplied space in the ovector. This could cause segfaults.
5392
5393
5394 Version 0.97 21-Oct-97
5395 ----------------------
5396
5397 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5398
5399 2. Optimized negated single characters not to use a bit map.
5400
5401 3. Brought error texts together as macro definitions; clarified some of them;
5402 fixed one that was wrong - it said "range out of order" when it meant "invalid
5403 escape sequence".
5404
5405 4. Changed some char * arguments to const char *.
5406
5407 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5408
5409 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5410 pcretest.
5411
5412
5413 Version 0.96 16-Oct-97
5414 ----------------------
5415
5416 1. Added a simple "pgrep" utility to the distribution.
5417
5418 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5419 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5420 where "ddd" means "one or more decimal digits".
5421
5422 3. Fixed serious bug. If a pattern had a back reference, but the call to
5423 pcre_exec() didn't supply a large enough ovector to record the related
5424 identifying subpattern, the match always failed. PCRE now remembers the number
5425 of the largest back reference, and gets some temporary memory in which to save
5426 the offsets during matching if necessary, in order to ensure that
5427 backreferences always work.
5428
5429 4. Increased the compatibility with Perl in a number of ways:
5430
5431 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5432 to request this handling. The option can be set at compile or exec time.
5433
5434 (b) $ matches before a terminating newline by default; an option
5435 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5436 mode). The option can be set at compile or exec time.
5437
5438 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5439 the same as Perl's. If the decimal number it represents is less than 10
5440 or there aren't that many previous left capturing parentheses, an octal
5441 escape is read. Inside a character class, it's always an octal escape,
5442 even if it is a single digit.
5443
5444 (d) An escaped but undefined alphabetic character is taken as a literal,
5445 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5446 escapes.
5447
5448 (e) {0} is now permitted. (The previous item is removed from the compiled
5449 pattern).
5450
5451 5. Changed all the names of code files so that the basic parts are no longer
5452 than 10 characters, and abolished the teeny "globals.c" file.
5453
5454 6. Changed the handling of character classes; they are now done with a 32-byte
5455 bit map always.
5456
5457 7. Added the -d and /D options to pcretest to make it possible to look at the
5458 internals of compilation without having to recompile pcre.
5459
5460
5461 Version 0.95 23-Sep-97
5462 ----------------------
5463
5464 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5465 \x20 at the start of a run of normal characters. These were being treated as
5466 real characters, instead of the source characters being re-checked.
5467
5468
5469 Version 0.94 18-Sep-97
5470 ----------------------
5471
5472 1. The functions are now thread-safe, with the caveat that the global variables
5473 containing pointers to malloc() and free() or alternative functions are the
5474 same for all threads.
5475
5476 2. Get pcre_study() to generate a bitmap of initial characters for non-
5477 anchored patterns when this is possible, and use it if passed to pcre_exec().
5478
5479
5480 Version 0.93 15-Sep-97
5481 ----------------------
5482
5483 1. /(b)|(:+)/ was computing an incorrect first character.
5484
5485 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5486 but not actually doing anything yet.
5487
5488 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5489 as Perl does (e.g. [-az] or [az-]).
5490
5491 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5492 all possible positions.
5493
5494 5. Split up into different modules to avoid including unneeded functions in a
5495 compiled binary. However, compile and exec are still in one module. The "study"
5496 function is split off.
5497
5498 6. The character tables are now in a separate module whose source is generated
5499 by an auxiliary program - but can then be edited by hand if required. There are
5500 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5501 toupper() in the code.
5502
5503 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5504 make them global. Abolish the function for setting them, as the caller can now
5505 set them directly.
5506
5507
5508 Version 0.92 11-Sep-97
5509 ----------------------
5510
5511 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5512 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5513
5514 2. Caseless matching was not working in character classes if the characters in
5515 the pattern were in upper case.
5516
5517 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5518
5519 4. Make PCRE_ANCHORED public and accept as a compile option.
5520
5521 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5522 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5523 pass them.
5524
5525 6. Give an error if bad option bits passed at compile or run time.
5526
5527 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5528 pcretest to cause it to pass that flag.
5529
5530 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5531 options, and the first character, if set.
5532
5533 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5534
5535
5536 Version 0.91 10-Sep-97
5537 ----------------------
5538
5539 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5540 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5541
5542 2. PCRE was looping on encountering an indefinitely repeated back reference to
5543 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5544 Perl does - treats the match as successful.
5545
5546 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5