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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1490 - (show annotations)
Thu Jun 19 07:51:39 2014 UTC (7 years, 1 month ago) by chpe
File size: 267599 byte(s)
Update to Unicode 7.0.0 release
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.36 xx-xxx-2014
5 ------------------------
6
7 1. Got rid of some compiler warnings in the C++ modules that were shown up by
8 -Wmissing-field-initializers and -Wunused-parameter.
9
10 2. The tests for quantifiers being too big (greater than 65535) were being
11 applied after reading the number, and stupidly assuming that integer
12 overflow would give a negative number. The tests are now applied as the
13 numbers are read.
14
15 3. Tidy code in pcre_exec.c where two branches that used to be different are
16 now the same.
17
18 4. The JIT compiler did not generate match limit checks for certain
19 bracketed expressions with quantifiers. This may lead to exponential
20 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
21 issue should be resolved now.
22
23 5. Fixed an issue, which occures when nested alternatives are optimized
24 with table jumps.
25
26 6. Inserted two casts and changed some ints to size_t in the light of some
27 reported 64-bit compiler warnings (Bugzilla 1477).
28
29 7. Fixed a bug concerned with zero-minimum possessive groups that could match
30 an empty string, which sometimes were behaving incorrectly in the
31 interpreter (though correctly in the JIT matcher). This pcretest input is
32 an example:
33
34 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
35 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
36
37 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
38 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
39 for an empty string was breaking the inner loop and carrying on at a lower
40 level, when possessive repeated groups should always return to a higher
41 level as they have no backtrack points in them. The empty string test now
42 occurs at the outer level.
43
44 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
45 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
46
47 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
48 doesn't).
49
50 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
51 to hold a pattern to be incorrectly computed (too small) when there were
52 named back references to duplicated names. This could cause "internal
53 error: code overflow" or "double free or corruption" or other memory
54 handling errors.
55
56 11. When named subpatterns had the same prefixes, back references could be
57 confused. For example, in this pattern:
58
59 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
60
61 the reference to 'Name' was incorrectly treated as a reference to a
62 duplicate name.
63
64 12. A pattern such as /^s?c/mi8 where the optional character has more than
65 one "other case" was incorrectly compiled such that it would only try to
66 match starting at "c".
67
68 13. When a pattern starting with \s was studied, VT was not included in the
69 list of possible starting characters; this should have been part of the
70 8.34/18 patch.
71
72 14. If a character class started [\Qx]... where x is any character, the class
73 was incorrectly terminated at the ].
74
75 15. If a pattern that started with a caseless match for a character with more
76 than one "other case" was studied, PCRE did not set up the starting code
77 unit bit map for the list of possible characters. Now it does. This is an
78 optimization improvement, not a bug fix.
79
80 16. The Unicode data tables have been updated to Unicode 7.0.0.
81
82
83 Version 8.35 04-April-2014
84 --------------------------
85
86 1. A new flag is set, when property checks are present in an XCLASS.
87 When this flag is not set, PCRE can perform certain optimizations
88 such as studying these XCLASS-es.
89
90 2. The auto-possessification of character sets were improved: a normal
91 and an extended character set can be compared now. Furthermore
92 the JIT compiler optimizes more character set checks.
93
94 3. Got rid of some compiler warnings for potentially uninitialized variables
95 that show up only when compiled with -O2.
96
97 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
98 of a match later then the end of the match. The pcretest program was not
99 handling the case sensibly - it was outputting from the start to the next
100 binary zero. It now reports this situation in a message, and outputs the
101 text from the end to the start.
102
103 5. Fast forward search is improved in JIT. Instead of the first three
104 characters, any three characters with fixed position can be searched.
105 Search order: first, last, middle.
106
107 6. Improve character range checks in JIT. Characters are read by an inprecise
108 function now, which returns with an unknown value if the character code is
109 above a certain treshold (e.g: 256). The only limitation is that the value
110 must be bigger than the treshold as well. This function is useful, when
111 the characters above the treshold are handled in the same way.
112
113 7. The macros whose names start with RAWUCHAR are placeholders for a future
114 mode in which only the bottom 21 bits of 32-bit data items are used. To
115 make this more memorable for those maintaining the code, the names have
116 been changed to start with UCHAR21, and an extensive comment has been added
117 to their definition.
118
119 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
120 to the export list in Makefile.am (they were accidentally omitted from the
121 8.34 tarball).
122
123 9. The informational output from pcretest used the phrase "starting byte set"
124 which is inappropriate for the 16-bit and 32-bit libraries. As the output
125 for "first char" and "need char" really means "non-UTF-char", I've changed
126 "byte" to "char", and slightly reworded the output. The documentation about
127 these values has also been (I hope) clarified.
128
129 10. Another JIT related optimization: use table jumps for selecting the correct
130 backtracking path, when more than four alternatives are present inside a
131 bracket.
132
133 11. Empty match is not possible, when the minimum length is greater than zero,
134 and there is no \K in the pattern. JIT should avoid empty match checks in
135 such cases.
136
137 12. In a caseless character class with UCP support, when a character with more
138 than one alternative case was not the first character of a range, not all
139 the alternative cases were added to the class. For example, s and \x{17f}
140 are both alternative cases for S: the class [RST] was handled correctly,
141 but [R-T] was not.
142
143 13. The configure.ac file always checked for pthread support when JIT was
144 enabled. This is not used in Windows, so I have put this test inside a
145 check for the presence of windows.h (which was already tested for).
146
147 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
148 The algorithm provides a way to skip certain starting offsets, and usually
149 faster than linear prefix searches.
150
151 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
152 as for 'fr_FR' and 'french'. For some reason, however, it then used the
153 Windows-specific input and output files, which have 'french' screwed in.
154 So this could never have worked. One of the problems with locales is that
155 they aren't always the same. I have now updated RunTest so that it checks
156 the output of the locale test (test 3) against three different output
157 files, and it allows the test to pass if any one of them matches. With luck
158 this should make the test pass on some versions of Solaris where it was
159 failing. Because of the uncertainty, the script did not used to stop if
160 test 3 failed; it now does. If further versions of a French locale ever
161 come to light, they can now easily be added.
162
163 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
164 there was a message during ./configure, but it did not stop. This now
165 provokes an error. The invalid example in README has been corrected.
166 If a value less than the minimum is given, the minimum value has always
167 been used, but now a warning is given.
168
169 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
170 was a bug in the test system, which is now fixed. Also, the list of various
171 configurations that are tested for each release did not have one with both
172 16/32 bits and --enable-bar-anycrlf. It now does.
173
174 18. pcretest was missing "-C bsr" for displaying the \R default setting.
175
176 19. Little endian PowerPC systems are supported now by the JIT compiler.
177
178 20. The fast forward newline mechanism could enter to an infinite loop on
179 certain invalid UTF-8 input. Although we don't support these cases
180 this issue can be fixed by a performance optimization.
181
182 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
183 not take account if existing stack usage. There is now a new global
184 variable called pcre_stack_guard that can be set to point to an external
185 function to check stack availability. It is called at the start of
186 processing every parenthesized group.
187
188 22. A typo in the code meant that in ungreedy mode the max/min qualifier
189 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
190 not match "ab".
191
192 23. When UTF was disabled, the JIT program reported some incorrect compile
193 errors. These messages are silenced now.
194
195 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
196 compiler.
197
198 25. Change all the temporary files used in RunGrepTest to be different to those
199 used by RunTest so that the tests can be run simultaneously, for example by
200 "make -j check".
201
202
203 Version 8.34 15-December-2013
204 -----------------------------
205
206 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
207 executable memory. Patch inspired by Carsten Klein.
208
209 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
210 this macro is never tested and has no effect, because the work to support
211 coverage involves only compiling and linking options and special targets in
212 the Makefile. The comment in config.h implied that defining the macro would
213 enable coverage support, which is totally false. There was also support for
214 setting this macro in the CMake files (my fault, I just copied it from
215 configure). SUPPORT_GCOV has now been removed.
216
217 3. Make a small performance improvement in strlen16() and strlen32() in
218 pcretest.
219
220 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
221 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
222 these cases). There was also one in pcretest.c.
223
224 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
225
226 6. In UTF mode, the code for checking whether a group could match an empty
227 string (which is used for indefinitely repeated groups to allow for
228 breaking an infinite loop) was broken when the group contained a repeated
229 negated single-character class with a character that occupied more than one
230 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
231 UTF-8 mode). The effect was undefined: the group might or might not be
232 deemed as matching an empty string, or the program might have crashed.
233
234 7. The code for checking whether a group could match an empty string was not
235 recognizing that \h, \H, \v, \V, and \R must match a character.
236
237 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
238 an empty string. If it can, pcretest shows this in its information output.
239
240 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
241 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
242 matched by pcre_exec() without using JIT:
243
244 (a) If the rest of the pattern did not match after a maximal run of
245 grapheme clusters, the code for backing up to try with fewer of them
246 did not always back up over a full grapheme when characters that do not
247 have the modifier quality were involved, e.g. Hangul syllables.
248
249 (b) If the match point in a subject started with modifier character, and
250 there was no match, the code could incorrectly back up beyond the match
251 point, and potentially beyond the first character in the subject,
252 leading to a segfault or an incorrect match result.
253
254 10. A conditional group with an assertion condition could lead to PCRE
255 recording an incorrect first data item for a match if no other first data
256 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
257 first data item, and therefore matched "ca" after "c" instead of at the
258 start.
259
260 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
261 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
262
263 12. The source of pcregrep now includes z/OS-specific code so that it can be
264 compiled for z/OS as part of the special z/OS distribution.
265
266 13. Added the -T and -TM options to pcretest.
267
268 14. The code in pcre_compile.c for creating the table of named capturing groups
269 has been refactored. Instead of creating the table dynamically during the
270 actual compiling pass, the information is remembered during the pre-compile
271 pass (on the stack unless there are more than 20 named groups, in which
272 case malloc() is used) and the whole table is created before the actual
273 compile happens. This has simplified the code (it is now nearly 150 lines
274 shorter) and prepared the way for better handling of references to groups
275 with duplicate names.
276
277 15. A back reference to a named subpattern when there is more than one of the
278 same name now checks them in the order in which they appear in the pattern.
279 The first one that is set is used for the reference. Previously only the
280 first one was inspected. This change makes PCRE more compatible with Perl.
281
282 16. Unicode character properties were updated from Unicode 6.3.0.
283
284 17. The compile-time code for auto-possessification has been refactored, based
285 on a patch by Zoltan Herczeg. It now happens after instead of during
286 compilation. The code is cleaner, and more cases are handled. The option
287 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
288 options in pcretest are provided to set it. It can also be set by
289 (*NO_AUTO_POSSESS) at the start of a pattern.
290
291 18. The character VT has been added to the default ("C" locale) set of
292 characters that match \s and are generally treated as white space,
293 following this same change in Perl 5.18. There is now no difference between
294 "Perl space" and "POSIX space". Whether VT is treated as white space in
295 other locales depends on the locale.
296
297 19. The code for checking named groups as conditions, either for being set or
298 for being recursed, has been refactored (this is related to 14 and 15
299 above). Processing unduplicated named groups should now be as fast at
300 numerical groups, and processing duplicated groups should be faster than
301 before.
302
303 20. Two patches to the CMake build system, by Alexander Barkov:
304
305 (1) Replace the "source" command by "." in CMakeLists.txt because
306 "source" is a bash-ism.
307
308 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
309 without these the CMake build does not work on Solaris.
310
311 21. Perl has changed its handling of \8 and \9. If there is no previously
312 encountered capturing group of those numbers, they are treated as the
313 literal characters 8 and 9 instead of a binary zero followed by the
314 literals. PCRE now does the same.
315
316 22. Following Perl, added \o{} to specify codepoints in octal, making it
317 possible to specify values greater than 0777 and also making them
318 unambiguous.
319
320 23. Perl now gives an error for missing closing braces after \x{... instead of
321 treating the string as literal. PCRE now does the same.
322
323 24. RunTest used to grumble if an inappropriate test was selected explicitly,
324 but just skip it when running all tests. This make it awkward to run ranges
325 of tests when one of them was inappropriate. Now it just skips any
326 inappropriate tests, as it always did when running all tests.
327
328 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
329 character types such as \d or \w, too many callouts were inserted, and the
330 data that they returned was rubbish.
331
332 26. In UCP mode, \s was not matching two of the characters that Perl matches,
333 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
334 were matched by \h. The code has now been refactored so that the lists of
335 the horizontal and vertical whitespace characters used for \h and \v (which
336 are defined only in one place) are now also used for \s.
337
338 27. Add JIT support for the 64 bit TileGX architecture.
339 Patch by Jiong Wang (Tilera Corporation).
340
341 28. Possessive quantifiers for classes (both explicit and automatically
342 generated) now use special opcodes instead of wrapping in ONCE brackets.
343
344 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
345 (because it's meaningless), this was not happening when PCRE_CASELESS was
346 set. Not wrong, but inefficient.
347
348 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
349 properties for \w, \d, etc) is present in a test regex. Otherwise if the
350 test contains no characters greater than 255, Perl doesn't realise it
351 should be using Unicode semantics.
352
353 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
354 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
355 does in Unicode mode.
356
357 32. Added the "forbid" facility to pcretest so that putting tests into the
358 wrong test files can sometimes be quickly detected.
359
360 33. There is now a limit (default 250) on the depth of nesting of parentheses.
361 This limit is imposed to control the amount of system stack used at compile
362 time. It can be changed at build time by --with-parens-nest-limit=xxx or
363 the equivalent in CMake.
364
365 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
366 errors. Perl warns for these when in warning mode, but PCRE has no facility
367 for giving warnings.
368
369 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
370 However, this was not working for (?!) because it is optimized to (*FAIL),
371 for which PCRE does not allow quantifiers. The optimization is now disabled
372 when a quantifier follows (?!). I can't see any use for this, but it makes
373 things uniform.
374
375 36. Perl no longer allows group names to start with digits, so I have made this
376 change also in PCRE. It simplifies the code a bit.
377
378 37. In extended mode, Perl ignores spaces before a + that indicates a
379 possessive quantifier. PCRE allowed a space before the quantifier, but not
380 before the possessive +. It now does.
381
382 38. The use of \K (reset reported match start) within a repeated possessive
383 group such as (a\Kb)*+ was not working.
384
385 40. Document that the same character tables must be used at compile time and
386 run time, and that the facility to pass tables to pcre_exec() and
387 pcre_dfa_exec() is for use only with saved/restored patterns.
388
389 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
390 features for Builds with MSVC:
391
392 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
393 builds). The .rc files can be used to set FileDescription and many other
394 attributes.
395
396 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
397 This allows higher-level build scripts which want .pdb files to avoid
398 hard-coding the exact files needed."
399
400 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
401 mean "start of word" and "end of word", respectively, as a transition aid.
402
403 43. A minimizing repeat of a class containing codepoints greater than 255 in
404 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
405 compiled to use the heap for recursion.
406
407 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
408 is configured.
409
410
411 Version 8.33 28-May-2013
412 ------------------------
413
414 1. Added 'U' to some constants that are compared to unsigned integers, to
415 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
416 variables that are added to signed variables, to ensure the result is
417 signed and can be negated.
418
419 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
420 CMake config files.
421
422 3. Revise the creation of config.h.generic so that all boolean macros are
423 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
424 overriding via -D on the command line possible.
425
426 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
427 to unsigned int is reported to make a quite noticeable speed difference in
428 a specific Windows environment. Testing on Linux did also appear to show
429 some benefit (and it is clearly not harmful). Also fixed the definition of
430 Xop which should be unsigned.
431
432 5. Related to (4), changing the definition of the intermediate variable cc
433 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
434 improvements.
435
436 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
437 unnecessary spaces.
438
439 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
440 and later.
441
442 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
443
444 9. Optimizing fast_forward_start_bits in JIT.
445
446 10. Adding support for callouts in JIT, and fixing some issues revealed
447 during this work. Namely:
448
449 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
450
451 (b) Minimum length was not checked before the matching is started.
452
453 11. The value of capture_last that is passed to callouts was incorrect in some
454 cases when there was a capture on one path that was subsequently abandoned
455 after a backtrack. Also, the capture_last value is now reset after a
456 recursion, since all captures are also reset in this case.
457
458 12. The interpreter no longer returns the "too many substrings" error in the
459 case when an overflowing capture is in a branch that is subsequently
460 abandoned after a backtrack.
461
462 13. In the pathological case when an offset vector of size 2 is used, pcretest
463 now prints out the matched string after a yield of 0 or 1.
464
465 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
466 Only supported by the JIT compiler at the moment.
467
468 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
469
470 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
471 offset of the starting point of the matching process, provided the offsets
472 vector is large enough.
473
474 17. The \A escape now records a lookbehind value of 1, though its execution
475 does not actually inspect the previous character. This is to ensure that,
476 in partial multi-segment matching, at least one character from the old
477 segment is retained when a new segment is processed. Otherwise, if there
478 are no lookbehinds in the pattern, \A might match incorrectly at the start
479 of a new segment.
480
481 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
482
483 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
484 gives some modest performance improvement in 8-bit mode.
485
486 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
487 be expressed in certain programming languages using Universal Character
488 Names.
489
490 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
491 which points out that "non characters" are not "characters that may not
492 appear in Unicode strings" but rather "characters that are reserved for
493 internal use and have only local meaning".
494
495 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
496 there was a conditional group that depended on an assertion, if the
497 assertion was false, the callout that immediately followed the alternation
498 in the condition was skipped when pcre_exec() was used for matching.
499
500 23. Allow an explicit callout to be inserted before an assertion that is the
501 condition for a conditional group, for compatibility with automatic
502 callouts, which always insert a callout at this point.
503
504 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
505 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
506
507 25. (*PRUNE) is now supported by the JIT compiler.
508
509 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
510
511 27. Fix the case where there are two or more SKIPs with arguments that may be
512 ignored.
513
514 28. (*SKIP) is now supported by the JIT compiler.
515
516 29. (*THEN) is now supported by the JIT compiler.
517
518 30. Update RunTest with additional test selector options.
519
520 31. The way PCRE handles backtracking verbs has been changed in two ways.
521
522 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
523 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
524 backtracking. In some cases this makes it more Perl-compatible, but Perl's
525 rather obscure rules do not always do the same thing.
526
527 (2) Previously, backtracking verbs were confined within assertions. This is
528 no longer the case for positive assertions, except for (*ACCEPT). Again,
529 this sometimes improves Perl compatibility, and sometimes does not.
530
531 32. A number of tests that were in test 2 because Perl did things differently
532 have been moved to test 1, because either Perl or PCRE has changed, and
533 these tests are now compatible.
534
535 32. Backtracking control verbs are now handled in the same way in JIT and
536 interpreter.
537
538 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
539 contained a forward subroutine reference caused a compile error.
540
541 34. Auto-detect and optimize limited repetitions in JIT.
542
543 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
544 blocking (*UTF) etc.
545
546 36. In the interpreter, maximizing pattern repetitions for characters and
547 character types now use tail recursion, which reduces stack usage.
548
549 37. The value of the max lookbehind was not correctly preserved if a compiled
550 and saved regex was reloaded on a host of different endianness.
551
552 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
553 of the compiled pattern block, expand the flags field from 16 to 32 bits
554 because it was almost full.
555
556 39. Try madvise first before posix_madvise.
557
558 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
559 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
560 original change locked out all matches of empty strings. This has been
561 changed so that one match of an empty string per line is recognized.
562 Subsequent searches on the same line (for colouring or for --only-matching,
563 for example) do not recognize empty strings.
564
565 41. Applied a user patch to fix a number of spelling mistakes in comments.
566
567 42. Data lines longer than 65536 caused pcretest to crash.
568
569 43. Clarified the data type for length and startoffset arguments for pcre_exec
570 and pcre_dfa_exec in the function-specific man pages, where they were
571 explicitly stated to be in bytes, never having been updated. I also added
572 some clarification to the pcreapi man page.
573
574 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
575 a segmentation fault.
576
577
578 Version 8.32 30-November-2012
579 -----------------------------
580
581 1. Improved JIT compiler optimizations for first character search and single
582 character iterators.
583
584 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
585 Patch by Daniel Richard G.
586
587 3. Single character iterator optimizations in the JIT compiler.
588
589 4. Improved JIT compiler optimizations for character ranges.
590
591 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
592 Reported by Giuseppe D'Angelo.
593
594 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
595 of a line, was being set incorrectly in cases where .* appeared inside
596 atomic brackets at the start of a pattern, or where there was a subsequent
597 *PRUNE or *SKIP.
598
599 7. Improved instruction cache flush for POWER/PowerPC.
600 Patch by Daniel Richard G.
601
602 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
603 grep:
604
605 (a) There is now no limit to the number of patterns to be matched.
606
607 (b) An error is given if a pattern is too long.
608
609 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
610 are now supported.
611
612 (d) --exclude-from and --include-from (multiple use) have been added.
613
614 (e) Exclusions and inclusions now apply to all files and directories, not
615 just to those obtained from scanning a directory recursively.
616
617 (f) Multiple uses of -f and --file-list are now supported.
618
619 (g) In a Windows environment, the default for -d has been changed from
620 "read" (the GNU grep default) to "skip", because otherwise the presence
621 of a directory in the file list provokes an error.
622
623 (h) The documentation has been revised and clarified in places.
624
625 9. Improve the matching speed of capturing brackets.
626
627 10. Changed the meaning of \X so that it now matches a Unicode extended
628 grapheme cluster.
629
630 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
631 out POSIX threads when JIT support is configured.
632
633 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
634
635 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
636 structure could go wrong in environments where size_t is not the same size
637 as int.
638
639 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
640 set.
641
642 15. The EBCDIC support had decayed; later updates to the code had included
643 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
644 general tidy up of EBCDIC-related issues, and the documentation was also
645 not quite right. There is now a test that can be run on ASCII systems to
646 check some of the EBCDIC-related things (but is it not a full test).
647
648 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
649 in a small tidy to the code.
650
651 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
652
653 18. If the --only-matching (-o) option in pcregrep is specified multiple
654 times, each one causes appropriate output. For example, -o1 -o2 outputs the
655 substrings matched by the 1st and 2nd capturing parentheses. A separating
656 string can be specified by --om-separator (default empty).
657
658 19. Improving the first n character searches.
659
660 20. Turn case lists for horizontal and vertical white space into macros so that
661 they are defined only once.
662
663 21. This set of changes together give more compatible Unicode case-folding
664 behaviour for characters that have more than one other case when UCP
665 support is available.
666
667 (a) The Unicode property table now has offsets into a new table of sets of
668 three or more characters that are case-equivalent. The MultiStage2.py
669 script that generates these tables (the pcre_ucd.c file) now scans
670 CaseFolding.txt instead of UnicodeData.txt for character case
671 information.
672
673 (b) The code for adding characters or ranges of characters to a character
674 class has been abstracted into a generalized function that also handles
675 case-independence. In UTF-mode with UCP support, this uses the new data
676 to handle characters with more than one other case.
677
678 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
679 whose other case is greater than 256 are now correctly matched
680 caselessly. Previously, the high codepoint matched the low one, but not
681 vice versa.
682
683 (d) The processing of \h, \H, \v, and \ in character classes now makes use
684 of the new class addition function, using character lists defined as
685 macros alongside the case definitions of 20 above.
686
687 (e) Caseless back references now work with characters that have more than
688 one other case.
689
690 (f) General caseless matching of characters with more than one other case
691 is supported.
692
693 22. Unicode character properties were updated from Unicode 6.2.0
694
695 23. Improved CMake support under Windows. Patch by Daniel Richard G.
696
697 24. Add support for 32-bit character strings, and UTF-32
698
699 25. Major JIT compiler update (code refactoring and bugfixing).
700 Experimental Sparc 32 support is added.
701
702 26. Applied a modified version of Daniel Richard G's patch to create
703 pcre.h.generic and config.h.generic by "make" instead of in the
704 PrepareRelease script.
705
706 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
707 pcre_compile.c when checking for a zero character.
708
709 28. Introducing a native interface for JIT. Through this interface, the compiled
710 machine code can be directly executed. The purpose of this interface is to
711 provide fast pattern matching, so several sanity checks are not performed.
712 However, feature tests are still performed. The new interface provides
713 1.4x speedup compared to the old one.
714
715 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
716 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
717 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
718
719 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
720 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
721 the "old" RFC 2279). Instead, it ended up passing a negative length to
722 pcre_exec().
723
724 31. Add support for GCC's visibility feature to hide internal functions.
725
726 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
727 "unknown -C option" after outputting 0 or 1.
728
729 33. There is now support for generating a code coverage report for the test
730 suite in environments where gcc is the compiler and lcov is installed. This
731 is mainly for the benefit of the developers.
732
733 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
734 unaddressable using valgrind annotations, allowing valgrind to detect
735 invalid memory accesses. This is mainly for the benefit of the developers.
736
737 25. (*UTF) can now be used to start a pattern in any of the three libraries.
738
739 26. Give configure error if --enable-cpp but no C++ compiler found.
740
741
742 Version 8.31 06-July-2012
743 -------------------------
744
745 1. Fixing a wrong JIT test case and some compiler warnings.
746
747 2. Removed a bashism from the RunTest script.
748
749 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
750 to unsigned type, result still unsigned" that was given by an MS compiler
751 on encountering the code "-sizeof(xxx)".
752
753 4. Partial matching support is added to the JIT compiler.
754
755 5. Fixed several bugs concerned with partial matching of items that consist
756 of more than one character:
757
758 (a) /^(..)\1/ did not partially match "aba" because checking references was
759 done on an "all or nothing" basis. This also applied to repeated
760 references.
761
762 (b) \R did not give a hard partial match if \r was found at the end of the
763 subject.
764
765 (c) \X did not give a hard partial match after matching one or more
766 characters at the end of the subject.
767
768 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
769 a partial match for the string "\r".
770
771 (e) When newline was set to CRLF, the metacharacter "." did not recognize
772 a partial match for a CR character at the end of the subject string.
773
774 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
775 running pcretest, the text "(JIT)" added to the output whenever JIT is
776 actually used to run the match.
777
778 7. Individual JIT compile options can be set in pcretest by following -s+[+]
779 or /S+[+] with a digit between 1 and 7.
780
781 8. OP_NOT now supports any UTF character not just single-byte ones.
782
783 9. (*MARK) control verb is now supported by the JIT compiler.
784
785 10. The command "./RunTest list" lists the available tests without actually
786 running any of them. (Because I keep forgetting what they all are.)
787
788 11. Add PCRE_INFO_MAXLOOKBEHIND.
789
790 12. Applied a (slightly modified) user-supplied patch that improves performance
791 when the heap is used for recursion (compiled with --disable-stack-for-
792 recursion). Instead of malloc and free for each heap frame each time a
793 logical recursion happens, frames are retained on a chain and re-used where
794 possible. This sometimes gives as much as 30% improvement.
795
796 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
797 call.
798
799 14. As documented, (*COMMIT) is now confined to within a positive assertion.
800
801 15. It is now possible to link pcretest with libedit as an alternative to
802 libreadline.
803
804 16. (*COMMIT) control verb is now supported by the JIT compiler.
805
806 17. The Unicode data tables have been updated to Unicode 6.1.0.
807
808 18. Added --file-list option to pcregrep.
809
810 19. Added binary file support to pcregrep, including the -a, --binary-files,
811 -I, and --text options.
812
813 20. The madvise function is renamed for posix_madvise for QNX compatibility
814 reasons. Fixed by Giuseppe D'Angelo.
815
816 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
817
818 22. Changed the option for creating symbolic links for 16-bit man pages from
819 -s to -sf so that re-installing does not cause issues.
820
821 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
822
823 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
824 matches in certain environments (the workspace was not being correctly
825 retained). Also added to pcre_dfa_exec() a simple plausibility check on
826 some of the workspace data at the beginning of a restart.
827
828 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
829 was not doing so when it should - probably a typo introduced by SVN 528
830 (change 8.10/14).
831
832 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
833 \w+ when the character tables indicated that \x{c4} was a word character.
834 There were several related cases, all because the tests for doing a table
835 lookup were testing for characters less than 127 instead of 255.
836
837 27. If a pattern contains capturing parentheses that are not used in a match,
838 their slots in the ovector are set to -1. For those that are higher than
839 any matched groups, this happens at the end of processing. In the case when
840 there were back references that the ovector was too small to contain
841 (causing temporary malloc'd memory to be used during matching), and the
842 highest capturing number was not used, memory off the end of the ovector
843 was incorrectly being set to -1. (It was using the size of the temporary
844 memory instead of the true size.)
845
846 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
847 ovector size, it uses memory at the end of the block that it has got.
848
849 29. Check for an overlong MARK name and give an error at compile time. The
850 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
851
852 30. JIT compiler update.
853
854 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
855 Rill for the patch.
856
857 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
858
859 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
860
861 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
862 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
863 was enabled.
864
865 35. Improve JIT code generation for greedy plus quantifier.
866
867 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
868 1 to "aa" instead of to an empty string. The bug affected repeated groups
869 that could potentially match an empty string.
870
871 37. Optimizing single character iterators in JIT.
872
873 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
874 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
875 codepoints that are too big for the mode are faulted, and in a UTF mode,
876 disallowed codepoints are also faulted.
877
878 39. If PCRE was compiled with UTF support, in three places in the DFA
879 matcher there was code that should only have been obeyed in UTF mode, but
880 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
881 processing when bytes with values greater than 127 were present. In 16-bit
882 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
883 both cases the values are those that cannot be the first data item in a UTF
884 character. The three items that might have provoked this were recursions,
885 possessively repeated groups, and atomic groups.
886
887 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
888 and pcregrep, because some OS require shared objects to be explicitly
889 passed to ld, causing the link step to fail if they are not.
890
891 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
892 mode, patterns that started with \h* or \R* might be incorrectly matched.
893
894
895 Version 8.30 04-February-2012
896 -----------------------------
897
898 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
899 name is defined in ctype.h.
900
901 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
902 only in quite long subpatterns.
903
904 3. Removed the function pcre_info(), which has been obsolete and deprecated
905 since it was replaced by pcre_fullinfo() in February 2000.
906
907 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
908 match a (*MARK), and the match failed at the start of the subject, a
909 reference to memory before the start of the subject could occur. This bug
910 was introduced by fix 17 of release 8.21.
911
912 5. A reference to an unset group with zero minimum repetition was giving
913 totally wrong answers (in non-JavaScript-compatibility mode). For example,
914 /(another)?(\1?)test/ matched against "hello world test". This bug was
915 introduced in release 8.13.
916
917 6. Add support for 16-bit character strings (a large amount of work involving
918 many changes and refactorings).
919
920 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
921 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
922 from a file.
923
924 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
925 rounding is not applied in this particular case).
926
927 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
928 if they appear, or are escaped, in patterns.
929
930 10. Get rid of a number of -Wunused-but-set-variable warnings.
931
932 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
933 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
934 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
935 also returns the mark "x". This bug applied to capturing parentheses,
936 non-capturing parentheses, and atomic parentheses. It also applied to some
937 assertions.
938
939 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
940 information out of configure.ac instead of relying on pcre.h.generic, which
941 is not stored in the repository.
942
943 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
944 -lreadline.
945
946 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
947
948 15. Applied Graycode's patch to put the top-level frame on the stack rather
949 than the heap when not using the stack for recursion. This gives a
950 performance improvement in many cases when recursion is not deep.
951
952 16. Experimental code added to "pcretest -C" to output the stack frame size.
953
954
955 Version 8.21 12-Dec-2011
956 ------------------------
957
958 1. Updating the JIT compiler.
959
960 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
961 are added as well.
962
963 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
964 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
965 calling _pcre_jit_exec. Some extra comments are added.
966
967 4. (*MARK) settings inside atomic groups that do not contain any capturing
968 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
969 was introduced by change 18 for 8.20.
970
971 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
972 ECMA-262 standard.
973
974 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
975 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
976 This bug was probably introduced by change 9 of 8.13.
977
978 7. While fixing 6 above, I noticed that a number of other items were being
979 incorrectly rejected as "not fixed length". This arose partly because newer
980 opcodes had not been added to the fixed-length checking code. I have (a)
981 corrected the bug and added tests for these items, and (b) arranged for an
982 error to occur if an unknown opcode is encountered while checking for fixed
983 length instead of just assuming "not fixed length". The items that were
984 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
985 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
986 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
987
988 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
989 being incorrectly compiled and would have given unpredicatble results.
990
991 9. A possessively repeated subpattern with minimum repeat count greater than
992 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
993 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
994 the first (A) could occur when it should not.
995
996 10. Add a cast and remove a redundant test from the code.
997
998 11. JIT should use pcre_malloc/pcre_free for allocation.
999
1000 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1001 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1002 is anything other than /usr, -L is still shown).
1003
1004 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1005
1006 14. Perl does not support \N without a following name in a [] class; PCRE now
1007 also gives an error.
1008
1009 15. If a forward reference was repeated with an upper limit of around 2000,
1010 it caused the error "internal error: overran compiling workspace". The
1011 maximum number of forward references (including repeats) was limited by the
1012 internal workspace, and dependent on the LINK_SIZE. The code has been
1013 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1014 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1015 of around 200,000 forward references. While doing this, I also speeded up
1016 the filling in of repeated forward references.
1017
1018 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1019 incorrectly expecting the subject to contain another "a" after the start.
1020
1021 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1022 in the match, the SKIP should be ignored. This was not happening; instead
1023 the SKIP was being treated as NOMATCH. For patterns such as
1024 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1025 tested.
1026
1027 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1028 now much more compatible with Perl, in particular in cases where the result
1029 is a non-match for a non-anchored pattern. For example, if
1030 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1031 "m", where previously it did not return a name. A side effect of this
1032 change is that for partial matches, the last encountered mark name is
1033 returned, as for non matches. A number of tests that were previously not
1034 Perl-compatible have been moved into the Perl-compatible test files. The
1035 refactoring has had the pleasing side effect of removing one argument from
1036 the match() function, thus reducing its stack requirements.
1037
1038 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1039 subsequent uses of /S (without +) incorrectly behaved like /S+.
1040
1041 21. Retrieve executable code size support for the JIT compiler and fixing
1042 some warnings.
1043
1044 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1045 not work when the shorter character appeared right at the end of the
1046 subject string.
1047
1048 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1049 systems.
1050
1051 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1052 output it when the /M option is used in pcretest.
1053
1054 25. The CheckMan script was not being included in the distribution. Also, added
1055 an explicit "perl" to run Perl scripts from the PrepareRelease script
1056 because this is reportedly needed in Windows.
1057
1058 26. If study data was being save in a file and studying had not found a set of
1059 "starts with" bytes for the pattern, the data written to the file (though
1060 never used) was taken from uninitialized memory and so caused valgrind to
1061 complain.
1062
1063 27. Updated RunTest.bat as provided by Sheri Pierce.
1064
1065 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1066
1067 29. Computation of memory usage for the table of capturing group names was
1068 giving an unnecessarily large value.
1069
1070
1071 Version 8.20 21-Oct-2011
1072 ------------------------
1073
1074 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1075 a POSIX class. After further experiments with Perl, which convinced me that
1076 Perl has bugs and confusions, a closing square bracket is no longer allowed
1077 in a POSIX name. This bug also affected patterns with classes that started
1078 with full stops.
1079
1080 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1081 captured substring, but while checking the failing first alternative,
1082 substring 1 is temporarily captured. If the output vector supplied to
1083 pcre_exec() was not big enough for this capture, the yield of the function
1084 was still zero ("insufficient space for captured substrings"). This cannot
1085 be totally fixed without adding another stack variable, which seems a lot
1086 of expense for a edge case. However, I have improved the situation in cases
1087 such as /(a)(b)x|abc/ matched against "abc", where the return code
1088 indicates that fewer than the maximum number of slots in the ovector have
1089 been set.
1090
1091 3. Related to (2) above: when there are more back references in a pattern than
1092 slots in the output vector, pcre_exec() uses temporary memory during
1093 matching, and copies in the captures as far as possible afterwards. It was
1094 using the entire output vector, but this conflicts with the specification
1095 that only 2/3 is used for passing back captured substrings. Now it uses
1096 only the first 2/3, for compatibility. This is, of course, another edge
1097 case.
1098
1099 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1100 main code base, and can be used by building with --enable-jit. When this is
1101 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1102 runtime --no-jit option is given.
1103
1104 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1105 ovector, the return from the function was zero, implying that there were
1106 other matches that did not fit. The correct "exactly full" value is now
1107 returned.
1108
1109 6. If a subpattern that was called recursively or as a subroutine contained
1110 (*PRUNE) or any other control that caused it to give a non-standard return,
1111 invalid errors such as "Error -26 (nested recursion at the same subject
1112 position)" or even infinite loops could occur.
1113
1114 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1115 computing the minimum length on reaching *ACCEPT, and so ended up with the
1116 wrong value of 1 rather than 0. Further investigation indicates that
1117 computing a minimum subject length in the presence of *ACCEPT is difficult
1118 (think back references, subroutine calls), and so I have changed the code
1119 so that no minimum is registered for a pattern that contains *ACCEPT.
1120
1121 8. If (*THEN) was present in the first (true) branch of a conditional group,
1122 it was not handled as intended. [But see 16 below.]
1123
1124 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1125 Sheri Pierce.
1126
1127 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1128 the first byte in a match must be "a".
1129
1130 11. Change 17 for 8.13 increased the recursion depth for patterns like
1131 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1132 pattern contains any instances of (*THEN). If it does not, the old
1133 optimizations are restored. It would be nice to do this on a per-group
1134 basis, but at the moment that is not feasible.
1135
1136 12. In some environments, the output of pcretest -C is CRLF terminated. This
1137 broke RunTest's code that checks for the link size. A single white space
1138 character after the value is now allowed for.
1139
1140 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1141 For "fr", it uses the Windows-specific input and output files.
1142
1143 14. If (*THEN) appeared in a group that was called recursively or as a
1144 subroutine, it did not work as intended. [But see next item.]
1145
1146 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1147 pattern fragments (but not containing any | characters). If A and B are
1148 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1149 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1150 D. In other words, Perl considers parentheses that do not contain any |
1151 characters to be part of a surrounding alternative, whereas PCRE was
1152 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1153 differently. PCRE now behaves in the same way as Perl, except in the case
1154 of subroutine/recursion calls such as (?1) which have in any case always
1155 been different (but PCRE had them first :-).
1156
1157 16. Related to 15 above: Perl does not treat the | in a conditional group as
1158 creating alternatives. Such a group is treated in the same way as an
1159 ordinary group without any | characters when processing (*THEN). PCRE has
1160 been changed to match Perl's behaviour.
1161
1162 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1163 RunGrepTest script failed.
1164
1165 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1166 inevitable for groups that contain captures, but it can lead to a lot of
1167 stack use in large patterns. The old behaviour has been restored for atomic
1168 groups that do not contain any capturing parentheses.
1169
1170 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1171 suppress the check for a minimum subject length at run time. (If it was
1172 given to pcre_exec() or pcre_dfa_exec() it did work.)
1173
1174 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1175 fail to work when decoding hex characters in data strings in EBCDIC
1176 environments.
1177
1178 21. It appears that in at least one Mac OS environment, the isxdigit() function
1179 is implemented as a macro that evaluates to its argument more than once,
1180 contravening the C 90 Standard (I haven't checked a later standard). There
1181 was an instance in pcretest which caused it to go wrong when processing
1182 \x{...} escapes in subject strings. The has been rewritten to avoid using
1183 things like p++ in the argument of isxdigit().
1184
1185
1186 Version 8.13 16-Aug-2011
1187 ------------------------
1188
1189 1. The Unicode data tables have been updated to Unicode 6.0.0.
1190
1191 2. Two minor typos in pcre_internal.h have been fixed.
1192
1193 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1194 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1195 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1196
1197 4. There were a number of related bugs in the code for matching backrefences
1198 caselessly in UTF-8 mode when codes for the characters concerned were
1199 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1200 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1201 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1202 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1203 2-byte code at the end of the subject (it thought there wasn't enough data
1204 left).
1205
1206 5. Comprehensive information about what went wrong is now returned by
1207 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1208 as the output vector has at least 2 elements. The offset of the start of
1209 the failing character and a reason code are placed in the vector.
1210
1211 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1212 now returned is for the first byte of the failing character, instead of the
1213 last byte inspected. This is an incompatible change, but I hope it is small
1214 enough not to be a problem. It makes the returned offset consistent with
1215 pcre_exec() and pcre_dfa_exec().
1216
1217 7. pcretest now gives a text phrase as well as the error number when
1218 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1219 failure, the offset and reason code are output.
1220
1221 8. When \R was used with a maximizing quantifier it failed to skip backwards
1222 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1223 back over a single character (\n). This seems wrong (because it treated the
1224 two characters as a single entity when going forwards), conflicts with the
1225 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1226 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1227 has been changed.
1228
1229 9. Some internal refactoring has changed the processing so that the handling
1230 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1231 time (the PCRE_DOTALL option was changed this way some time ago: version
1232 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1233 which was always a bit of a fudge. It also means that there is one less
1234 argument for the match() function, which reduces its stack requirements
1235 slightly. This change also fixes an incompatibility with Perl: the pattern
1236 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1237
1238 10. More internal refactoring has drastically reduced the number of recursive
1239 calls to match() for possessively repeated groups such as (abc)++ when
1240 using pcre_exec().
1241
1242 11. While implementing 10, a number of bugs in the handling of groups were
1243 discovered and fixed:
1244
1245 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1246 (a|)*(?1) gave a compile-time internal error.
1247 ((a|)+)+ did not notice that the outer group could match an empty string.
1248 (^a|^)+ was not marked as anchored.
1249 (.*a|.*)+ was not marked as matching at start or after a newline.
1250
1251 12. Yet more internal refactoring has removed another argument from the match()
1252 function. Special calls to this function are now indicated by setting a
1253 value in a variable in the "match data" data block.
1254
1255 13. Be more explicit in pcre_study() instead of relying on "default" for
1256 opcodes that mean there is no starting character; this means that when new
1257 ones are added and accidentally left out of pcre_study(), testing should
1258 pick them up.
1259
1260 14. The -s option of pcretest has been documented for ages as being an old
1261 synonym of -m (show memory usage). I have changed it to mean "force study
1262 for every regex", that is, assume /S for every regex. This is similar to -i
1263 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1264 using it. It makes it easier to run collections of tests with and without
1265 study enabled, and thereby test pcre_study() more easily. All the standard
1266 tests are now run with and without -s (but some patterns can be marked as
1267 "never study" - see 20 below).
1268
1269 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1270 restoration of the capturing data to the outer values was not happening
1271 correctly.
1272
1273 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1274 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1275 pattern had matched an empty string, and so incorrectly returned a no
1276 match.
1277
1278 17. There was optimizing code for the last branch of non-capturing parentheses,
1279 and also for the obeyed branch of a conditional subexpression, which used
1280 tail recursion to cut down on stack usage. Unfortunately, now that there is
1281 the possibility of (*THEN) occurring in these branches, tail recursion is
1282 no longer possible because the return has to be checked for (*THEN). These
1283 two optimizations have therefore been removed. [But see 8.20/11 above.]
1284
1285 18. If a pattern containing \R was studied, it was assumed that \R always
1286 matched two bytes, thus causing the minimum subject length to be
1287 incorrectly computed because \R can also match just one byte.
1288
1289 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1290 was incorrectly computed.
1291
1292 20. If /S is present twice on a test pattern in pcretest input, it now
1293 *disables* studying, thereby overriding the use of -s on the command line
1294 (see 14 above). This is necessary for one or two tests to keep the output
1295 identical in both cases.
1296
1297 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1298 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1299
1300 22. When an atomic group that contained a capturing parenthesis was
1301 successfully matched, but the branch in which it appeared failed, the
1302 capturing was not being forgotten if a higher numbered group was later
1303 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1304 group 1 to "a", when in fact it should be unset. This applied to multi-
1305 branched capturing and non-capturing groups, repeated or not, and also to
1306 positive assertions (capturing in negative assertions does not happen
1307 in PCRE) and also to nested atomic groups.
1308
1309 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1310 subject after a captured substring, to make it easier to tell which of a
1311 number of identical substrings has been captured.
1312
1313 24. The way atomic groups are processed by pcre_exec() has been changed so that
1314 if they are repeated, backtracking one repetition now resets captured
1315 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1316 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1317 "aaa". Previously, it would have been "a". As part of this code
1318 refactoring, the way recursive calls are handled has also been changed.
1319
1320 25. If an assertion condition captured any substrings, they were not passed
1321 back unless some other capturing happened later. For example, if
1322 (?(?=(a))a) was matched against "a", no capturing was returned.
1323
1324 26. When studying a pattern that contained subroutine calls or assertions,
1325 the code for finding the minimum length of a possible match was handling
1326 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1327 group 1 called group 2 while simultaneously a separate group 2 called group
1328 1). A stack overflow occurred in this case. I have fixed this by limiting
1329 the recursion depth to 10.
1330
1331 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1332 Fortmann. This supports explicit test numbers on the command line, and has
1333 argument validation and error reporting.
1334
1335 28. An instance of \X with an unlimited repeat could fail if at any point the
1336 first character it looked at was a mark character.
1337
1338 29. Some minor code refactoring concerning Unicode properties and scripts
1339 should reduce the stack requirement of match() slightly.
1340
1341 30. Added the '=' option to pcretest to check the setting of unused capturing
1342 slots at the end of the pattern, which are documented as being -1, but are
1343 not included in the return count.
1344
1345 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1346 compiled something random. Now it gives a compile-time error (as does
1347 Perl).
1348
1349 32. A *MARK encountered during the processing of a positive assertion is now
1350 recorded and passed back (compatible with Perl).
1351
1352 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1353 had alternative anchored branches, the search for a second match in a line
1354 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1355 matched the line "0102" twice. The same bug affected patterns that started
1356 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1357 twice.
1358
1359 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1360 does, and because of capturing effects, quantifying parenthesized
1361 assertions may at times be useful. Quantifiers are now allowed for
1362 parenthesized assertions.
1363
1364 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1365
1366 36. \g was being checked for fancy things in a character class, when it should
1367 just be a literal "g".
1368
1369 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1370 appearance of a nested POSIX class supersedes an apparent external class.
1371 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1372 unescaped square brackets may also appear as part of class names. For
1373 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1374 more like Perl. (But see 8.20/1 above.)
1375
1376 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1377 was because it thought it was \N{name}, which is not supported).
1378
1379 39. Add minix to OS list not supporting the -S option in pcretest.
1380
1381 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1382 cannot analyze patterns in sufficient detail to catch mutual recursions
1383 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1384 subgroup is called recursively as a subpattern for a second time at the
1385 same position in the subject string. In previous releases this might have
1386 been caught by the recursion limit, or it might have run out of stack.
1387
1388 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1389 happen only once. PCRE was, however incorrectly giving a compile time error
1390 "recursive call could loop indefinitely" because it cannot analyze the
1391 pattern in sufficient detail. The compile time test no longer happens when
1392 PCRE is compiling a conditional subpattern, but actual runaway loops are
1393 now caught at runtime (see 40 above).
1394
1395 42. It seems that Perl allows any characters other than a closing parenthesis
1396 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1397 has been changed to be the same.
1398
1399 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1400 as not to get warnings when autogen.sh is called. Also changed
1401 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1402
1403 44. To help people who use pcregrep to scan files containing exceedingly long
1404 lines, the following changes have been made:
1405
1406 (a) The default value of the buffer size parameter has been increased from
1407 8K to 20K. (The actual buffer used is three times this size.)
1408
1409 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1410 PCRE is built.
1411
1412 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1413 to be set at run time.
1414
1415 (d) Numerical values in pcregrep options can be followed by K or M, for
1416 example --buffer-size=50K.
1417
1418 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1419 given and the return code is set to 2.
1420
1421 45. Add a pointer to the latest mark to the callout data block.
1422
1423 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1424 partial match of an empty string instead of no match. This was specific to
1425 the use of ".".
1426
1427 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1428 complete match instead of a partial match. This bug was dependent on both
1429 the PCRE_UTF8 and PCRE_DOTALL options being set.
1430
1431 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1432 starting byte set, because \b was not being ignored.
1433
1434
1435 Version 8.12 15-Jan-2011
1436 ------------------------
1437
1438 1. Fixed some typos in the markup of the man pages, and wrote a script that
1439 checks for such things as part of the documentation building process.
1440
1441 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1442 --match-limit and --recursion-limit options (added for 8.11). In
1443 particular, this made one of the standard tests fail. (The integer value
1444 went into the wrong half of a long int.)
1445
1446 3. If the --colour option was given to pcregrep with -v (invert match), it
1447 did strange things, either producing crazy output, or crashing. It should,
1448 of course, ignore a request for colour when reporting lines that do not
1449 match.
1450
1451 4. Another pcregrep bug caused similar problems if --colour was specified with
1452 -M (multiline) and the pattern match finished with a line ending.
1453
1454 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1455 matched in multiline mode, the following line was shown as part of the
1456 match. This seems wrong, so I have changed it.
1457
1458 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1459 the check for further matches in the same line (so they could be coloured)
1460 to overrun the end of the current line. If another match was found, it was
1461 incorrectly shown (and then shown again when found in the next line).
1462
1463 7. If pcregrep was compiled under Windows, there was a reference to the
1464 function pcregrep_exit() before it was defined. I am assuming this was
1465 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1466 reported by a user. I've moved the definition above the reference.
1467
1468
1469 Version 8.11 10-Dec-2010
1470 ------------------------
1471
1472 1. (*THEN) was not working properly if there were untried alternatives prior
1473 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1474 backtracked to try for "b" instead of moving to the next alternative branch
1475 at the same level (in this case, to look for "c"). The Perl documentation
1476 is clear that when (*THEN) is backtracked onto, it goes to the "next
1477 alternative in the innermost enclosing group".
1478
1479 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1480 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1481 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1482 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1483 (*THEN).
1484
1485 3. If \s appeared in a character class, it removed the VT character from
1486 the class, even if it had been included by some previous item, for example
1487 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1488 of \s, but is part of the POSIX "space" class.)
1489
1490 4. A partial match never returns an empty string (because you can always
1491 match an empty string at the end of the subject); however the checking for
1492 an empty string was starting at the "start of match" point. This has been
1493 changed to the "earliest inspected character" point, because the returned
1494 data for a partial match starts at this character. This means that, for
1495 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1496 (previously it gave "no match").
1497
1498 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1499 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1500 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1501 has an implication that the given string is incomplete (because a partial
1502 match is preferred over a full match). For this reason, these items now
1503 give a partial match in this situation. [Aside: previously, the one case
1504 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1505 match rather than a full match, which was wrong by the old rules, but is
1506 now correct.]
1507
1508 6. There was a bug in the handling of #-introduced comments, recognized when
1509 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1510 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1511 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1512 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1513 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1514 places in pcre_compile().
1515
1516 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1517 comments when looking ahead for named forward references to subpatterns,
1518 the only newline sequence it recognized was NL. It now handles newlines
1519 according to the set newline convention.
1520
1521 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1522 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1523 cater for a lack of strerror(). These oversights have been fixed.
1524
1525 9. Added --match-limit and --recursion-limit to pcregrep.
1526
1527 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1528
1529 11. When the -o option was used, pcregrep was setting a return code of 1, even
1530 when matches were found, and --line-buffered was not being honoured.
1531
1532 12. Added an optional parentheses number to the -o and --only-matching options
1533 of pcregrep.
1534
1535 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1536 can match an empty string. The code to do it in pcretest and pcredemo
1537 needed fixing:
1538
1539 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1540 only one byte after an empty string match just before CRLF (this case
1541 just got forgotten; "any" and "anycrlf" were OK).
1542
1543 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1544 mode when an empty string match preceded an ASCII character followed by
1545 a non-ASCII character. (The code for advancing by one character rather
1546 than one byte was nonsense.)
1547
1548 (c) The pcredemo.c sample program did not have any code at all to handle
1549 the cases when CRLF is a valid newline sequence.
1550
1551 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1552 as a starting offset was within the subject string. There is now a new
1553 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1554 negative or greater than the length of the string. In order to test this,
1555 pcretest is extended to allow the setting of negative starting offsets.
1556
1557 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1558 starting offset points to the beginning of a UTF-8 character was
1559 unnecessarily clumsy. I tidied it up.
1560
1561 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1562 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1563
1564 17. Nobody had reported that the --include_dir option, which was added in
1565 release 7.7 should have been called --include-dir (hyphen, not underscore)
1566 for compatibility with GNU grep. I have changed it to --include-dir, but
1567 left --include_dir as an undocumented synonym, and the same for
1568 --exclude-dir, though that is not available in GNU grep, at least as of
1569 release 2.5.4.
1570
1571 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1572 characters from a string of bytes) have been redefined so as not to use
1573 loops, in order to improve performance in some environments. At the same
1574 time, I abstracted some of the common code into auxiliary macros to save
1575 repetition (this should not affect the compiled code).
1576
1577 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1578 compile-time error is now given if \c is not followed by an ASCII
1579 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1580 different, and any byte value is allowed.)
1581
1582 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1583 START_OPTIMIZE option, which is now allowed at compile time - but just
1584 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1585 to pcregrep and other applications that have no direct access to PCRE
1586 options. The new /Y option in pcretest sets this option when calling
1587 pcre_compile().
1588
1589 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1590 back references. Groups containing recursive back references were forced to
1591 be atomic by that change, but in the case of named groups, the amount of
1592 memory required was incorrectly computed, leading to "Failed: internal
1593 error: code overflow". This has been fixed.
1594
1595 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1596 pcretest.c, to avoid build problems in some Borland environments.
1597
1598
1599 Version 8.10 25-Jun-2010
1600 ------------------------
1601
1602 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1603 THEN.
1604
1605 2. (*ACCEPT) was not working when inside an atomic group.
1606
1607 3. Inside a character class, \B is treated as a literal by default, but
1608 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1609 causes the error). The code is unchanged, but I tidied the documentation.
1610
1611 4. Inside a character class, PCRE always treated \R and \X as literals,
1612 whereas Perl faults them if its -w option is set. I have changed PCRE so
1613 that it faults them when PCRE_EXTRA is set.
1614
1615 5. Added support for \N, which always matches any character other than
1616 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1617
1618 6. When compiling pcregrep with newer versions of gcc which may have
1619 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1620 declared with attribute warn_unused_result" were given. Just casting the
1621 result to (void) does not stop the warnings; a more elaborate fudge is
1622 needed. I've used a macro to implement this.
1623
1624 7. Minor change to pcretest.c to avoid a compiler warning.
1625
1626 8. Added four artifical Unicode properties to help with an option to make
1627 \s etc use properties (see next item). The new properties are: Xan
1628 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1629
1630 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1631 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1632 this option. Modified pcretest to add /W to test this facility. Added
1633 REG_UCP to make it available via the POSIX interface.
1634
1635 10. Added --line-buffered to pcregrep.
1636
1637 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1638 studied, and the match started with a letter with a code point greater than
1639 127 whose first byte was different to the first byte of the other case of
1640 the letter, the other case of this starting letter was not recognized
1641 (#976).
1642
1643 12. If a pattern that was studied started with a repeated Unicode property
1644 test, for example, \p{Nd}+, there was the theoretical possibility of
1645 setting up an incorrect bitmap of starting bytes, but fortunately it could
1646 not have actually happened in practice until change 8 above was made (it
1647 added property types that matched character-matching opcodes).
1648
1649 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1650 possible starting bytes for non-anchored patterns.
1651
1652 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1653 \R, and also a number of cases that involve Unicode properties, both
1654 explicit and implicit when PCRE_UCP is set.
1655
1656 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1657 input, it could crash or give wrong results if characters with values
1658 greater than 0xc0 were present in the subject string. (Detail: it assumed
1659 UTF-8 input when processing these items.)
1660
1661 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1662 size_t is 64-bit (#991).
1663
1664 17. Added a check for running out of memory when PCRE is compiled with
1665 --disable-stack-for-recursion (#990).
1666
1667 18. If the last data line in a file for pcretest does not have a newline on
1668 the end, a newline was missing in the output.
1669
1670 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1671 less than 128) in its various bitmaps. However, there is a facility for
1672 generating tables according to the current locale when PCRE is compiled. It
1673 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1674 characters, are recognized by isspace() and therefore were getting set in
1675 these tables, and indeed these tables seem to approximate to ISO 8859. This
1676 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1677 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1678 which of course cannot start UTF-8 characters. I have changed the code so
1679 that only real ASCII characters (less than 128) and the correct starting
1680 bytes for UTF-8 encodings are set for characters greater than 127 when in
1681 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1682 altogether.)
1683
1684 20. Added the /T option to pcretest so as to be able to run tests with non-
1685 standard character tables, thus making it possible to include the tests
1686 used for 19 above in the standard set of tests.
1687
1688 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1689 reference to a subpattern the other side of a comment that contains an
1690 opening parenthesis caused either an internal compiling error, or a
1691 reference to the wrong subpattern.
1692
1693
1694 Version 8.02 19-Mar-2010
1695 ------------------------
1696
1697 1. The Unicode data tables have been updated to Unicode 5.2.0.
1698
1699 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1700 configured.
1701
1702 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1703 original author of that file, following a query about its status.
1704
1705 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1706 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1707
1708 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1709 quantifier applied to a forward-referencing subroutine call, could compile
1710 incorrect code or give the error "internal error: previously-checked
1711 referenced subpattern not found".
1712
1713 6. Both MS Visual Studio and Symbian OS have problems with initializing
1714 variables to point to external functions. For these systems, therefore,
1715 pcre_malloc etc. are now initialized to local functions that call the
1716 relevant global functions.
1717
1718 7. There were two entries missing in the vectors called coptable and poptable
1719 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1720 I've fixed the data, and added a kludgy way of testing at compile time that
1721 the lengths are correct (equal to the number of opcodes).
1722
1723 8. Following on from 7, I added a similar kludge to check the length of the
1724 eint vector in pcreposix.c.
1725
1726 9. Error texts for pcre_compile() are held as one long string to avoid too
1727 much relocation at load time. To find a text, the string is searched,
1728 counting zeros. There was no check for running off the end of the string,
1729 which could happen if a new error number was added without updating the
1730 string.
1731
1732 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1733
1734 11. \K was not working if it appeared in an atomic group or in a group that
1735 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1736 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1737 the assertion is positive, but not if it is negative.
1738
1739 12. Change 11 fortuitously reduced the size of the stack frame used in the
1740 "match()" function of pcre_exec.c by one pointer. Forthcoming
1741 implementation of support for (*MARK) will need an extra pointer on the
1742 stack; I have reserved it now, so that the stack frame size does not
1743 decrease.
1744
1745 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1746 item in branch that calls a recursion is a subroutine call - as in the
1747 second branch in the above example - was incorrectly given the compile-
1748 time error "recursive call could loop indefinitely" because pcre_compile()
1749 was not correctly checking the subroutine for matching a non-empty string.
1750
1751 14. The checks for overrunning compiling workspace could trigger after an
1752 overrun had occurred. This is a "should never occur" error, but it can be
1753 triggered by pathological patterns such as hundreds of nested parentheses.
1754 The checks now trigger 100 bytes before the end of the workspace.
1755
1756 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1757
1758
1759 Version 8.01 19-Jan-2010
1760 ------------------------
1761
1762 1. If a pattern contained a conditional subpattern with only one branch (in
1763 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1764 computed the wrong minimum data length (which is of course zero for such
1765 subpatterns). This could cause incorrect "no match" results.
1766
1767 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1768 the pattern is reset in the first branch, pcre_compile() failed with
1769 "internal error: code overflow at offset...". This happened only when
1770 the reset was to the original external option setting. (An optimization
1771 abstracts leading options settings into an external setting, which was the
1772 cause of this.)
1773
1774 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1775 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1776 assertion pattern did not match (meaning that the assertion was true), it
1777 was incorrectly treated as false if the SKIP had been reached during the
1778 matching. This also applied to assertions used as conditions.
1779
1780 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1781 assertion subpattern, including such a pattern used as a condition,
1782 unpredictable results occurred, instead of the error return
1783 PCRE_ERROR_DFA_UITEM.
1784
1785 5. The C++ GlobalReplace function was not working like Perl for the special
1786 situation when an empty string is matched. It now does the fancy magic
1787 stuff that is necessary.
1788
1789 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1790 removed. (These were left over from very, very early versions of PCRE.)
1791
1792 7. Some cosmetic changes to the code to make life easier when compiling it
1793 as part of something else:
1794
1795 (a) Change DEBUG to PCRE_DEBUG.
1796
1797 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1798 called "current" as "current_branch", to prevent a collision with the
1799 Linux macro when compiled as a kernel module.
1800
1801 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1802 prevent a collision with the Linux macro when compiled as a kernel
1803 module.
1804
1805 8. In pcre_compile() there are some checks for integer overflows that used to
1806 cast potentially large values to (double). This has been changed to that
1807 when building, a check for int64_t is made, and if it is found, it is used
1808 instead, thus avoiding the use of floating point arithmetic. (There is no
1809 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1810 double.
1811
1812 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1813 2005 (difference between two addresses compared to an unsigned value).
1814
1815 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1816 custom one, because of the following reported problem in Windows:
1817
1818 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1819 under Win32.
1820 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1821 therefore missing the function definition.
1822 - The compiler thus generates a "C" signature for the test function.
1823 - The linker fails to find the "C" function.
1824 - PCRE fails to configure if asked to do so against libbz2.
1825
1826 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1827 messages were output:
1828
1829 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1830 rerunning libtoolize, to keep the correct libtool macros in-tree.
1831 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1832
1833 I have done both of these things.
1834
1835 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1836 most of the time, it *can* run out if it is given a pattern that contains a
1837 runaway infinite recursion. I updated the discussion in the pcrestack man
1838 page.
1839
1840 13. Now that we have gone to the x.xx style of version numbers, the minor
1841 version may start with zero. Using 08 or 09 is a bad idea because users
1842 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1843 interpreted as invalid octal numbers. I've updated the previous comment in
1844 configure.ac, and also added a check that gives an error if 08 or 09 are
1845 used.
1846
1847 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1848 causing partial matching to fail when the end of the subject matched \W
1849 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1850
1851 15. There were some discrepancies between the declarations in pcre_internal.h
1852 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1853 their definitions. The declarations used "const uschar *" and the
1854 definitions used USPTR. Even though USPTR is normally defined as "const
1855 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1856 reported that: "This difference in casting confuses some C++ compilers, for
1857 example, SunCC recognizes above declarations as different functions and
1858 generates broken code for hbpcre." I have changed the declarations to use
1859 USPTR.
1860
1861 16. GNU libtool is named differently on some systems. The autogen.sh script now
1862 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1863 (FreeBSD).
1864
1865 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1866 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1867 comment: "Figure out how to create a longlong from a string: strtoll and
1868 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1869 instance, but it only takes 2 args instead of 3!"
1870
1871 18. A subtle bug concerned with back references has been fixed by a change of
1872 specification, with a corresponding code fix. A pattern such as
1873 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1874 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1875 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1876 same bug. Such groups have to be quantified to be useful, or contained
1877 inside another quantified group. (If there's no repetition, the reference
1878 can never match.) The problem arises because, having left the group and
1879 moved on to the rest of the pattern, a later failure that backtracks into
1880 the group uses the captured value from the final iteration of the group
1881 rather than the correct earlier one. I have fixed this in PCRE by forcing
1882 any group that contains a reference to itself to be an atomic group; that
1883 is, there cannot be any backtracking into it once it has completed. This is
1884 similar to recursive and subroutine calls.
1885
1886
1887 Version 8.00 19-Oct-09
1888 ----------------------
1889
1890 1. The table for translating pcre_compile() error codes into POSIX error codes
1891 was out-of-date, and there was no check on the pcre_compile() error code
1892 being within the table. This could lead to an OK return being given in
1893 error.
1894
1895 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1896 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1897 in a Windows environment.
1898
1899 3. The pcregrep --count option prints the count for each file even when it is
1900 zero, as does GNU grep. However, pcregrep was also printing all files when
1901 --files-with-matches was added. Now, when both options are given, it prints
1902 counts only for those files that have at least one match. (GNU grep just
1903 prints the file name in this circumstance, but including the count seems
1904 more useful - otherwise, why use --count?) Also ensured that the
1905 combination -clh just lists non-zero counts, with no names.
1906
1907 4. The long form of the pcregrep -F option was incorrectly implemented as
1908 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1909 but it seems right to fix it, and I didn't think it was worth preserving
1910 the old behaviour.
1911
1912 5. The command line items --regex=pattern and --regexp=pattern were not
1913 recognized by pcregrep, which required --regex pattern or --regexp pattern
1914 (with a space rather than an '='). The man page documented the '=' forms,
1915 which are compatible with GNU grep; these now work.
1916
1917 6. No libpcreposix.pc file was created for pkg-config; there was just
1918 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1919
1920 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1921 when UCP support is not needed, by modifying the Python script that
1922 generates it from Unicode data files. This should not matter if the module
1923 is correctly used as a library, but I received one complaint about 50K of
1924 unwanted data. My guess is that the person linked everything into his
1925 program rather than using a library. Anyway, it does no harm.
1926
1927 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1928 was a minimum greater than 1 for a wide character in a possessive
1929 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1930 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1931 character. Chaos in the form of incorrect output or a compiling loop could
1932 result.
1933
1934 9. The restrictions on what a pattern can contain when partial matching is
1935 requested for pcre_exec() have been removed. All patterns can now be
1936 partially matched by this function. In addition, if there are at least two
1937 slots in the offset vector, the offset of the earliest inspected character
1938 for the match and the offset of the end of the subject are set in them when
1939 PCRE_ERROR_PARTIAL is returned.
1940
1941 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1942 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1943 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1944 and may be more useful for multi-segment matching.
1945
1946 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1947 used to be given if ever the end of the subject was reached; now it is
1948 given only if matching could not proceed because another character was
1949 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1950 string "Z", which now yields "no match" instead of "partial match". In the
1951 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1952 final character ended with (*FAIL).
1953
1954 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1955 if the pattern had a "must contain" character that was already found in the
1956 earlier partial match, unless partial matching was again requested. For
1957 example, with the pattern /dog.(body)?/, the "must contain" character is
1958 "g". If the first part-match was for the string "dog", restarting with
1959 "sbody" failed. This bug has been fixed.
1960
1961 13. The string returned by pcre_dfa_exec() after a partial match has been
1962 changed so that it starts at the first inspected character rather than the
1963 first character of the match. This makes a difference only if the pattern
1964 starts with a lookbehind assertion or \b or \B (\K is not supported by
1965 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1966 matching functions compatible, and I think it's the right thing to do.
1967
1968 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1969 so that the demonstration program is easily available in environments where
1970 PCRE has not been installed from source.
1971
1972 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1973 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1974 library.
1975
1976 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1977 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1978 is not the first non-POSIX option to be added. Clearly some people find
1979 these options useful.
1980
1981 17. If a caller to the POSIX matching function regexec() passes a non-zero
1982 value for nmatch with a NULL value for pmatch, the value of
1983 nmatch is forced to zero.
1984
1985 18. RunGrepTest did not have a test for the availability of the -u option of
1986 the diff command, as RunTest does. It now checks in the same way as
1987 RunTest, and also checks for the -b option.
1988
1989 19. If an odd number of negated classes containing just a single character
1990 interposed, within parentheses, between a forward reference to a named
1991 subpattern and the definition of the subpattern, compilation crashed with
1992 an internal error, complaining that it could not find the referenced
1993 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1994 [The bug was that it was starting one character too far in when skipping
1995 over the character class, thus treating the ] as data rather than
1996 terminating the class. This meant it could skip too much.]
1997
1998 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1999 /g option in pcretest when the pattern contains \K, which makes it possible
2000 to have an empty string match not at the start, even when the pattern is
2001 anchored. Updated pcretest and pcredemo to use this option.
2002
2003 21. If the maximum number of capturing subpatterns in a recursion was greater
2004 than the maximum at the outer level, the higher number was returned, but
2005 with unset values at the outer level. The correct (outer level) value is
2006 now given.
2007
2008 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2009 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2010 make it do so. The string so far is captured, making this feature
2011 compatible with Perl.
2012
2013 23. The tests have been re-organized, adding tests 11 and 12, to make it
2014 possible to check the Perl 5.10 features against Perl 5.10.
2015
2016 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2017 pattern matches a fixed length string. PCRE did not allow this; now it
2018 does. Neither allows recursion.
2019
2020 25. I finally figured out how to implement a request to provide the minimum
2021 length of subject string that was needed in order to match a given pattern.
2022 (It was back references and recursion that I had previously got hung up
2023 on.) This code has now been added to pcre_study(); it finds a lower bound
2024 to the length of subject needed. It is not necessarily the greatest lower
2025 bound, but using it to avoid searching strings that are too short does give
2026 some useful speed-ups. The value is available to calling programs via
2027 pcre_fullinfo().
2028
2029 26. While implementing 25, I discovered to my embarrassment that pcretest had
2030 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2031 study optimizations had never been tested with that matching function.
2032 Oops. What is worse, even when it was passed study data, there was a bug in
2033 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2034 were also very few tests of studied patterns with pcre_dfa_exec().
2035
2036 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2037 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2038 on the other side of the coin, they are no longer allowed to have different
2039 names, because these cannot be distinguished in PCRE, and this has caused
2040 confusion. (This is a difference from Perl.)
2041
2042 28. When duplicate subpattern names are present (necessarily with different
2043 numbers, as required by 27 above), and a test is made by name in a
2044 conditional pattern, either for a subpattern having been matched, or for
2045 recursion in such a pattern, all the associated numbered subpatterns are
2046 tested, and the overall condition is true if the condition is true for any
2047 one of them. This is the way Perl works, and is also more like the way
2048 testing by number works.
2049
2050
2051 Version 7.9 11-Apr-09
2052 ---------------------
2053
2054 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2055 (pcretest), all targets were linked against these libraries. This included
2056 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2057 libraries. This caused unwanted dependencies to be created. This problem
2058 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2059 pcretest is linked with readline.
2060
2061 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2062 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2063 moved outside it again, because FALSE and TRUE are already defined in AIX,
2064 but BOOL is not.
2065
2066 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2067 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2068
2069 4. The pcregrep documentation said spaces were inserted as well as colons (or
2070 hyphens) following file names and line numbers when outputting matching
2071 lines. This is not true; no spaces are inserted. I have also clarified the
2072 wording for the --colour (or --color) option.
2073
2074 5. In pcregrep, when --colour was used with -o, the list of matching strings
2075 was not coloured; this is different to GNU grep, so I have changed it to be
2076 the same.
2077
2078 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2079 each matching line was coloured. Now it goes on to look for further matches
2080 of any of the test patterns, which is the same behaviour as GNU grep.
2081
2082 7. A pattern that could match an empty string could cause pcregrep to loop; it
2083 doesn't make sense to accept an empty string match in pcregrep, so I have
2084 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2085 seems to be how GNU grep behaves. [But see later change 40 for release
2086 8.33.]
2087
2088 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2089 start or after a newline", because the conditional assertion was not being
2090 correctly handled. The rule now is that both the assertion and what follows
2091 in the first alternative must satisfy the test.
2092
2093 9. If auto-callout was enabled in a pattern with a conditional group whose
2094 condition was an assertion, PCRE could crash during matching, both with
2095 pcre_exec() and pcre_dfa_exec().
2096
2097 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2098 used for matching.
2099
2100 11. Unicode property support in character classes was not working for
2101 characters (bytes) greater than 127 when not in UTF-8 mode.
2102
2103 12. Added the -M command line option to pcretest.
2104
2105 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2106
2107 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2108
2109 16. Added comments and documentation about mis-use of no_arg in the C++
2110 wrapper.
2111
2112 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2113 from Martin Jerabek that uses macro names for all relevant character and
2114 string constants.
2115
2116 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2117 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2118 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2119 these, but not everybody uses configure.
2120
2121 19. A conditional group that had only one branch was not being correctly
2122 recognized as an item that could match an empty string. This meant that an
2123 enclosing group might also not be so recognized, causing infinite looping
2124 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2125 with the subject "ab", where knowledge that the repeated group can match
2126 nothing is needed in order to break the loop.
2127
2128 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2129 exec(), but without supplying a callout function, matching went wrong.
2130
2131 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2132 leak if the size of the offset vector was greater than 30. When the vector
2133 is smaller, the saved offsets during recursion go onto a local stack
2134 vector, but for larger vectors malloc() is used. It was failing to free
2135 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2136 error, in fact).
2137
2138 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2139 heapframe that is used only when UTF-8 support is enabled. This caused no
2140 problem, but was untidy.
2141
2142 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2143 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2144 included within another project.
2145
2146 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2147 slightly modified by me:
2148
2149 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2150 not building pcregrep.
2151
2152 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2153 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2154
2155 25. Forward references, both numeric and by name, in patterns that made use of
2156 duplicate group numbers, could behave incorrectly or give incorrect errors,
2157 because when scanning forward to find the reference group, PCRE was not
2158 taking into account the duplicate group numbers. A pattern such as
2159 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2160
2161 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2162 the feature of a custom pointer more persuasive (as requested by a user).
2163
2164 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2165 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2166 pre-defined.
2167
2168 28. Added support for (*UTF8) at the start of a pattern.
2169
2170 29. Arrange for flags added by the "release type" setting in CMake to be shown
2171 in the configuration summary.
2172
2173
2174 Version 7.8 05-Sep-08
2175 ---------------------
2176
2177 1. Replaced UCP searching code with optimized version as implemented for Ad
2178 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2179 stage table and inline lookup instead of a function, giving speed ups of 2
2180 to 5 times on some simple patterns that I tested. Permission was given to
2181 distribute the MultiStage2.py script that generates the tables (it's not in
2182 the tarball, but is in the Subversion repository).
2183
2184 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2185 scripts.
2186
2187 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2188 a group with a zero qualifier. The result of the study could be incorrect,
2189 or the function might crash, depending on the pattern.
2190
2191 4. Caseless matching was not working for non-ASCII characters in back
2192 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2193 It now works when Unicode Property Support is available.
2194
2195 5. In pcretest, an escape such as \x{de} in the data was always generating
2196 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2197 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2198 truncation.
2199
2200 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2201
2202 7. Added two (int) casts to pcregrep when printing the difference of two
2203 pointers, in case they are 64-bit values.
2204
2205 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2206 test 2 if it fails.
2207
2208 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2209 and a #define of that name to empty if it is not externally set. This is to
2210 allow users of MSVC to set it if necessary.
2211
2212 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2213 the convenience functions in the pcre_get.c source file.
2214
2215 11. An option change at the start of a pattern that had top-level alternatives
2216 could cause overwriting and/or a crash. This command provoked a crash in
2217 some environments:
2218
2219 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2220
2221 This potential security problem was recorded as CVE-2008-2371.
2222
2223 12. For a pattern where the match had to start at the beginning or immediately
2224 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2225 pcre_dfa_exec() could read past the end of the passed subject if there was
2226 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2227 pcretest so that it places the subject at the end of its malloc-ed buffer.
2228
2229 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2230 exec() might read past the end of the data buffer in UTF-8 mode.
2231
2232 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2233 the data contained the byte 0x85 as part of a UTF-8 character within its
2234 first line. This applied both to normal and DFA matching.
2235
2236 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2237 /^[^d]*?$/8 failed to match "abc".
2238
2239 16. Added a missing copyright notice to pcrecpp_internal.h.
2240
2241 17. Make it more clear in the documentation that values returned from
2242 pcre_exec() in ovector are byte offsets, not character counts.
2243
2244 18. Tidied a few places to stop certain compilers from issuing warnings.
2245
2246 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2247 supplied by Stefan Weber. I made a further small update for 7.8 because
2248 there is a change of source arrangements: the pcre_searchfuncs.c module is
2249 replaced by pcre_ucd.c.
2250
2251
2252 Version 7.7 07-May-08
2253 ---------------------
2254
2255 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2256 a string to a long long, pretend we don't even have a long long." This is
2257 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2258
2259 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2260 pre-7.6 versions, which defined a global no_arg variable instead of putting
2261 it in the RE class. (See also #8 below.)
2262
2263 3. Remove a line of dead code, identified by coverity and reported by Nuno
2264 Lopes.
2265
2266 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2267
2268 (1) The include/exclude patterns were being applied to the whole pathnames
2269 of files, instead of just to the final components.
2270
2271 (2) If there was more than one level of directory, the subdirectories were
2272 skipped unless they satisfied the include/exclude conditions. This is
2273 inconsistent with GNU grep (and could even be seen as contrary to the
2274 pcregrep specification - which I improved to make it absolutely clear).
2275 The action now is always to scan all levels of directory, and just
2276 apply the include/exclude patterns to regular files.
2277
2278 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2279 --exclude_dir in the tests to avoid scanning .svn directories.
2280
2281 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2282 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2283 doesn't support NULs in patterns.
2284
2285 7. Added some missing "const"s to declarations of static tables in
2286 pcre_compile.c and pcre_dfa_exec.c.
2287
2288 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2289 caused by fix #2 above. (Subsequently also a second patch to fix the
2290 first patch. And a third patch - this was a messy problem.)
2291
2292 9. Applied Craig's patch to remove the use of push_back().
2293
2294 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2295 matching function regexec().
2296
2297 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2298 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2299 references. PCRE supports relative numbers with this syntax (I don't think
2300 Oniguruma does).
2301
2302 12. Previously, a group with a zero repeat such as (...){0} was completely
2303 omitted from the compiled regex. However, this means that if the group
2304 was called as a subroutine from elsewhere in the pattern, things went wrong
2305 (an internal error was given). Such groups are now left in the compiled
2306 pattern, with a new opcode that causes them to be skipped at execution
2307 time.
2308
2309 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2310 to the way PCRE behaves:
2311
2312 (a) A lone ] character is dis-allowed (Perl treats it as data).
2313
2314 (b) A back reference to an unmatched subpattern matches an empty string
2315 (Perl fails the current match path).
2316
2317 (c) A data ] in a character class must be notated as \] because if the
2318 first data character in a class is ], it defines an empty class. (In
2319 Perl it is not possible to have an empty class.) The empty class []
2320 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2321 The negative empty class [^] matches any one character, independently
2322 of the DOTALL setting.
2323
2324 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2325 non-existent subpattern following a character class starting with ']' and
2326 containing () gave an internal compiling error instead of "reference to
2327 non-existent subpattern". Fortunately, when the pattern did exist, the
2328 compiled code was correct. (When scanning forwards to check for the
2329 existence of the subpattern, it was treating the data ']' as terminating
2330 the class, so got the count wrong. When actually compiling, the reference
2331 was subsequently set up correctly.)
2332
2333 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2334 it was being rejected as not supported by pcre_dfa_exec(), even though
2335 other assertions are supported. I have made pcre_dfa_exec() support
2336 (*FAIL).
2337
2338 16. The implementation of 13c above involved the invention of a new opcode,
2339 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2340 cannot be changed at match time, I realized I could make a small
2341 improvement to matching performance by compiling OP_ALLANY instead of
2342 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2343 on the OP_ANY path.
2344
2345 17. Compiling pcretest on Windows with readline support failed without the
2346 following two fixes: (1) Make the unistd.h include conditional on
2347 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2348
2349 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2350 ncurses library to be included for pcretest when ReadLine support is
2351 requested, but also to allow for it to be overridden. This patch came from
2352 Daniel Bergström.
2353
2354 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2355 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2356 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2357 spotting this.
2358
2359
2360 Version 7.6 28-Jan-08
2361 ---------------------
2362
2363 1. A character class containing a very large number of characters with
2364 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2365 overflow.
2366
2367 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2368 HAVE_LONG_LONG is not defined.
2369
2370 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2371 bring them up to date and include new features. This patch includes:
2372
2373 - Fixed PH's badly added libz and libbz2 support.
2374 - Fixed a problem with static linking.
2375 - Added pcredemo. [But later removed - see 7 below.]
2376 - Fixed dftables problem and added an option.
2377 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2378 HAVE_LONG_LONG.
2379 - Added readline support for pcretest.
2380 - Added an listing of the option settings after cmake has run.
2381
2382 4. A user submitted a patch to Makefile that makes it easy to create
2383 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2384 Makefile.am that cause it to include this special target, without
2385 affecting anything else. Note that the same mingw target plus all
2386 the other distribution libraries and programs are now supported
2387 when configuring with CMake (see 6 below) instead of with
2388 Configure/Make.
2389
2390 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2391 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2392 exported in the Windows port". It has not yet been confirmed that the patch
2393 solves the problem, but it does no harm.
2394
2395 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2396 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2397 with CMake, and also correct the comment about stack recursion.
2398
2399 7. Remove the automatic building of pcredemo from the ./configure system and
2400 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2401 of a program that users should build themselves after PCRE is installed, so
2402 building it automatically is not really right. What is more, it gave
2403 trouble in some build environments.
2404
2405 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2406
2407
2408 Version 7.5 10-Jan-08
2409 ---------------------
2410
2411 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2412 values in parens when parsing an RE using the C++ wrapper."
2413
2414 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2415 Characters greater than 255 were excluded from the class instead of being
2416 included.
2417
2418 3. The same bug as (2) above applied to negated POSIX classes such as
2419 [:^space:].
2420
2421 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2422 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2423 I have changed it.
2424
2425 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2426 first named subpattern) and a construct such as (?&a) would reference the
2427 first named subpattern whose name started with "a" (in other words, the
2428 length check was missing). Both these problems are fixed. "Subpattern name
2429 expected" is now given for (?&) (a zero-length name), and this patch also
2430 makes it give the same error for \k'' (previously it complained that that
2431 was a reference to a non-existent subpattern).
2432
2433 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2434 this is right because (?- can be followed by option settings as well as by
2435 digits. I have, however, made the messages clearer.
2436
2437 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2438 than the number used in the conditional) now cause a compile-time error.
2439 This is actually not compatible with Perl, which accepts such patterns, but
2440 treats the conditional as always being FALSE (as PCRE used to), but it
2441 seems to me that giving a diagnostic is better.
2442
2443 8. Change "alphameric" to the more common word "alphanumeric" in comments
2444 and messages.
2445
2446 9. Fix two occurrences of "backslash" in comments that should have been
2447 "backspace".
2448
2449 10. Remove two redundant lines of code that can never be obeyed (their function
2450 was moved elsewhere).
2451
2452 11. The program that makes PCRE's Unicode character property table had a bug
2453 which caused it to generate incorrect table entries for sequences of
2454 characters that have the same character type, but are in different scripts.
2455 It amalgamated them into a single range, with the script of the first of
2456 them. In other words, some characters were in the wrong script. There were
2457 thirteen such cases, affecting characters in the following ranges:
2458
2459 U+002b0 - U+002c1
2460 U+0060c - U+0060d
2461 U+0061e - U+00612
2462 U+0064b - U+0065e
2463 U+0074d - U+0076d
2464 U+01800 - U+01805
2465 U+01d00 - U+01d77
2466 U+01d9b - U+01dbf
2467 U+0200b - U+0200f
2468 U+030fc - U+030fe
2469 U+03260 - U+0327f
2470 U+0fb46 - U+0fbb1
2471 U+10450 - U+1049d
2472
2473 12. The -o option (show only the matching part of a line) for pcregrep was not
2474 compatible with GNU grep in that, if there was more than one match in a
2475 line, it showed only the first of them. It now behaves in the same way as
2476 GNU grep.
2477
2478 13. If the -o and -v options were combined for pcregrep, it printed a blank
2479 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2480 does the same. The return code can be used to tell if there were any
2481 non-matching lines.
2482
2483 14. Added --file-offsets and --line-offsets to pcregrep.
2484
2485 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2486 infinitely looping recursion. The bug was that positive lookaheads were not
2487 being skipped when checking for a possible empty match (negative lookaheads
2488 and both kinds of lookbehind were skipped).
2489
2490 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2491 inclusion of <windows.h> to before rather than after the definition of
2492 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2493
2494 17. Specifying a possessive quantifier with a specific limit for a Unicode
2495 character property caused pcre_compile() to compile bad code, which led at
2496 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2497 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2498 caused the error; without that there was no problem.
2499
2500 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2501
2502 19. Added --enable-pcretest-libreadline.
2503
2504 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2505 RE::GlobalReplace(). As a result, the number of replacements returned was
2506 double what it should be. I removed one of the increments, but Craig sent a
2507 later patch that removed the other one (the right fix) and added unit tests
2508 that check the return values (which was not done before).
2509
2510 21. Several CMake things:
2511
2512 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2513 the names libpcre and libpcreposix, not just pcre and pcreposix.
2514
2515 (2) The above change means that pcretest and pcregrep are now correctly
2516 linked with the newly-built libraries, not previously installed ones.
2517
2518 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2519
2520 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2521 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2522 UTF-8 newline character). The key issue is that the pattern starts .*;
2523 this means that the match must be either at the beginning, or after a
2524 newline. The bug was in the code for advancing after a failed match and
2525 checking that the new position followed a newline. It was not taking
2526 account of UTF-8 characters correctly.
2527
2528 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2529 character classes. PCRE was not treating the sequence [:...:] as a
2530 character class unless the ... were all letters. Perl, however, seems to
2531 allow any characters between [: and :], though of course it rejects as
2532 unknown any "names" that contain non-letters, because all the known class
2533 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2534 for example, whereas PCRE did not - it did not recognize a POSIX character
2535 class. This seemed a bit dangerous, so the code has been changed to be
2536 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2537 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2538 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2539 Perl does, and where it didn't before.
2540
2541 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2542 Windows environments %n is disabled by default.
2543
2544
2545 Version 7.4 21-Sep-07
2546 ---------------------
2547
2548 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2549 means that a class such as [\s] counted as "explicit reference to CR or
2550 LF". That isn't really right - the whole point of the change was to try to
2551 help when there was an actual mention of one of the two characters. So now
2552 the change happens only if \r or \n (or a literal CR or LF) character is
2553 encountered.
2554
2555 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2556 of both had grown to the point where there were only 3 bits left.
2557 Fortunately, there was spare space in the data structure, and so I have
2558 moved the internal flags into a new 16-bit field to free up more option
2559 bits.
2560
2561 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2562 but did not set the internal JCHANGED flag - either of these is enough to
2563 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2564 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2565 start sets both bits.
2566
2567 4. Added options (at build time, compile time, exec time) to change \R from
2568 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2569
2570 5. doc/pcresyntax.html was missing from the distribution.
2571
2572 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2573 compatibility, even though it is no longer used.
2574
2575 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2576 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2577 windows.h file is present (where different names are used). [This was
2578 reversed later after testing - see 16 below.]
2579
2580 8. Changed all #include <config.h> to #include "config.h". There were also
2581 some further <pcre.h> cases that I changed to "pcre.h".
2582
2583 9. When pcregrep was used with the --colour option, it missed the line ending
2584 sequence off the lines that it output.
2585
2586 10. It was pointed out to me that arrays of string pointers cause lots of
2587 relocations when a shared library is dynamically loaded. A technique of
2588 using a single long string with a table of offsets can drastically reduce
2589 these. I have refactored PCRE in four places to do this. The result is
2590 dramatic:
2591
2592 Originally: 290
2593 After changing UCP table: 187
2594 After changing error message table: 43
2595 After changing table of "verbs" 36
2596 After changing table of Posix names 22
2597
2598 Thanks to the folks working on Gregex for glib for this insight.
2599
2600 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2601 unicode-properties was also set.
2602
2603 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2604
2605 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2606 checked only for CRLF.
2607
2608 14. Added casts to pcretest.c to avoid compiler warnings.
2609
2610 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2611
2612 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2613 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2614 entirely. This removes changes made in 7 above.
2615
2616 17. The CMake files have been updated, and there is now more information about
2617 building with CMake in the NON-UNIX-USE document.
2618
2619
2620 Version 7.3 28-Aug-07
2621 ---------------------
2622
2623 1. In the rejigging of the build system that eventually resulted in 7.1, the
2624 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2625 brackets there is not right, since it causes compilers to look for an
2626 installed pcre.h, not the version that is in the source that is being
2627 compiled (which of course may be different). I have changed it back to:
2628
2629 #include "pcre.h"
2630
2631 I have a vague recollection that the change was concerned with compiling in
2632 different directories, but in the new build system, that is taken care of
2633 by the VPATH setting the Makefile.
2634
2635 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2636 when the subject happened to end in the byte 0x85 (e.g. if the last
2637 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2638 characters but of course it shouldn't be taken as a newline when it is part
2639 of another character. The bug was that, for an unlimited repeat of . in
2640 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2641 characters when looking for a newline.
2642
2643 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2644
2645 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2646 in debug output.
2647
2648 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2649 long printing in the pcrecpp unittest when running under MinGW.
2650
2651 6. ESC_K was left out of the EBCDIC table.
2652
2653 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2654 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2655 limit also applies to "virtual nesting" when a pattern is recursive, and in
2656 this case 1000 isn't so big. I have been able to remove this limit at the
2657 expense of backing off one optimization in certain circumstances. Normally,
2658 when pcre_exec() would call its internal match() function recursively and
2659 immediately return the result unconditionally, it uses a "tail recursion"
2660 feature to save stack. However, when a subpattern that can match an empty
2661 string has an unlimited repetition quantifier, it no longer makes this
2662 optimization. That gives it a stack frame in which to save the data for
2663 checking that an empty string has been matched. Previously this was taken
2664 from the 1000-entry workspace that had been reserved. So now there is no
2665 explicit limit, but more stack is used.
2666
2667 8. Applied Daniel's patches to solve problems with the import/export magic
2668 syntax that is required for Windows, and which was going wrong for the
2669 pcreposix and pcrecpp parts of the library. These were overlooked when this
2670 problem was solved for the main library.
2671
2672 9. There were some crude static tests to avoid integer overflow when computing
2673 the size of patterns that contain repeated groups with explicit upper
2674 limits. As the maximum quantifier is 65535, the maximum group length was
2675 set at 30,000 so that the product of these two numbers did not overflow a
2676 32-bit integer. However, it turns out that people want to use groups that
2677 are longer than 30,000 bytes (though not repeat them that many times).
2678 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2679 made it possible to implement the integer overflow checks in a much more
2680 dynamic way, which I have now done. The artificial limitation on group
2681 length has been removed - we now have only the limit on the total length of
2682 the compiled pattern, which depends on the LINK_SIZE setting.
2683
2684 10. Fixed a bug in the documentation for get/copy named substring when
2685 duplicate names are permitted. If none of the named substrings are set, the
2686 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2687 empty string.
2688
2689 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2690 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2691 because the ] is interpreted as the first data character and the
2692 terminating ] is not found. PCRE has been made compatible with Perl in this
2693 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2694 cause memory overwriting.
2695
2696 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2697 string has been matched (to stop an infinite loop). It was not recognizing
2698 a conditional subpattern that could match an empty string if that
2699 subpattern was within another subpattern. For example, it looped when
2700 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2701 condition was not nested. This bug has been fixed.
2702
2703 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2704 past the start of the subject in the presence of bytes with the top bit
2705 set, for example "\x8aBCD".
2706
2707 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2708 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2709
2710 14. Optimized (?!) to (*FAIL).
2711
2712 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2713 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2714 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2715 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2716 does: it's just the validity check that is more restrictive.
2717
2718 16. Inserted checks for integer overflows during escape sequence (backslash)
2719 processing, and also fixed erroneous offset values for syntax errors during
2720 backslash processing.
2721
2722 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2723 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2724
2725 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2726 caused an overrun.
2727
2728 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2729 something other than just ASCII characters) inside a group that had an
2730 unlimited repeat caused a loop at compile time (while checking to see
2731 whether the group could match an empty string).
2732
2733 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2734 [\P{Any}] did so. (Error in the code for printing property names.)
2735
2736 21. An orphan \E inside a character class could cause a crash.
2737
2738 22. A repeated capturing bracket such as (A)? could cause a wild memory
2739 reference during compilation.
2740
2741 23. There are several functions in pcre_compile() that scan along a compiled
2742 expression for various reasons (e.g. to see if it's fixed length for look
2743 behind). There were bugs in these functions when a repeated \p or \P was
2744 present in the pattern. These operators have additional parameters compared
2745 with \d, etc, and these were not being taken into account when moving along
2746 the compiled data. Specifically:
2747
2748 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2749 length.
2750
2751 (b) An item such as \pL+ within a repeated group could cause crashes or
2752 loops.
2753
2754 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2755 "reference to non-existent subpattern" error.
2756
2757 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2758
2759 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2760 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2761
2762 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2763
2764 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2765 character were causing crashes (broken optimization).
2766
2767 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2768 \p or \P) caused a compile-time loop.
2769
2770 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2771 break. For example, the unstudied pattern [\r\n]A does not match the string
2772 "\r\nA" because change 7.0/46 below moves the current point on by two
2773 characters after failing to match at the start. However, the pattern \nA
2774 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2775 the same is true. There doesn't seem any very clean way out of this, but
2776 what I have chosen to do makes the common cases work: PCRE now takes note
2777 of whether there can be an explicit match for \r or \n anywhere in the
2778 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2779 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2780 pattern has explicit CR or LF references.
2781
2782 29. Added (*CR) etc for changing newline setting at start of pattern.
2783
2784
2785 Version 7.2 19-Jun-07
2786 ---------------------
2787
2788 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2789 which is apparently normally available under Windows.
2790
2791 2. Re-jig the pcregrep tests with different newline settings in an attempt
2792 to make them independent of the local environment's newline setting.
2793
2794 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2795
2796 4. Some of the "internals" tests were previously cut out when the link size
2797 was not 2, because the output contained actual offsets. The recent new
2798 "Z" feature of pcretest means that these can be cut out, making the tests
2799 usable with all link sizes.
2800
2801 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2802 stack recursion. This gives a massive performance boost under BSD, but just
2803 a small improvement under Linux. However, it saves one field in the frame
2804 in all cases.
2805
2806 6. Added more features from the forthcoming Perl 5.10:
2807
2808 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2809 recursion call. It refers to the nth most recently opened parentheses.
2810
2811 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2812 to be opened parentheses.
2813
2814 (c) Conditions that refer to capturing parentheses can be specified
2815 relatively, for example, (?(-2)... or (?(+3)...
2816
2817 (d) \K resets the start of the current match so that everything before
2818 is not part of it.
2819
2820 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2821
2822 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2823 reference syntax.
2824
2825 (g) (?| introduces a group in which the numbering of parentheses in each
2826 alternative starts with the same number.
2827
2828 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2829
2830 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2831 PCRE_INFO_JCHANGED.
2832
2833 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2834 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2835 for detecting groups that can match an empty string.
2836
2837 9. A pattern with a very large number of alternatives (more than several
2838 hundred) was running out of internal workspace during the pre-compile
2839 phase, where pcre_compile() figures out how much memory will be needed. A
2840 bit of new cunning has reduced the workspace needed for groups with
2841 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2842 workspace instead of running out of the 4096 that are available.
2843
2844 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2845
2846 11. Applied patch from Google to remove an optimization that didn't quite work.
2847 The report of the bug said:
2848
2849 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2850 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2851 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2852
2853 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2854 it matched the wrong number of bytes.
2855
2856
2857 Version 7.1 24-Apr-07
2858 ---------------------
2859
2860 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2861 that is more "standard", making use of automake and other Autotools. There
2862 is some re-arrangement of the files and adjustment of comments consequent
2863 on this.
2864
2865 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2866 for recursive directory scanning broke on some systems because the files
2867 are not scanned in any specific order and on different systems the order
2868 was different. A call to "sort" has been inserted into RunGrepTest for the
2869 approprate test as a short-term fix. In the longer term there may be an
2870 alternative.
2871
2872 3. I had an email from Eric Raymond about problems translating some of PCRE's
2873 man pages to HTML (despite the fact that I distribute HTML pages, some
2874 people do their own conversions for various reasons). The problems
2875 concerned the use of low-level troff macros .br and .in. I have therefore
2876 removed all such uses from the man pages (some were redundant, some could
2877 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2878 HTML has been updated to handle .nf/.fi and to complain if it encounters
2879 .br or .in.
2880
2881 4. Updated comments in configure.ac that get placed in config.h.in and also
2882 arranged for config.h to be included in the distribution, with the name
2883 config.h.generic, for the benefit of those who have to compile without
2884 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2885
2886 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2887 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2888 makevp.bat for the current PCRE, using the additional files
2889 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2890
2891 6. A Windows user reported a minor discrepancy with test 2, which turned out
2892 to be caused by a trailing space on an input line that had got lost in his
2893 copy. The trailing space was an accident, so I've just removed it.
2894
2895 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2896 that is needed.
2897
2898 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2899 as "const" (a) because they are and (b) because it helps the PHP
2900 maintainers who have recently made a script to detect big data structures
2901 in the php code that should be moved to the .rodata section. I remembered
2902 to update Builducptable as well, so it won't revert if ucptable.h is ever
2903 re-created.
2904
2905 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2906 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2907 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2908 support is not required. This saves 1.5-2K of code, which is important in
2909 some applications.
2910
2911 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2912 so as not to refer to the tables, even though these functions will never be
2913 called when UTF-8 support is disabled. Otherwise there are problems with a
2914 shared library.
2915
2916 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2917
2918 (a) It was defining its arguments as char * instead of void *.
2919
2920 (b) It was assuming that all moves were upwards in memory; this was true
2921 a long time ago when I wrote it, but is no longer the case.
2922
2923 The emulated memove() is provided for those environments that have neither
2924 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2925 is clearly not the case, as these two bugs were recently reported.
2926
2927 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2928 and Detrail to create the HTML documentation, the .txt form of the man
2929 pages, and it removes trailing spaces from listed files. It also creates
2930 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2931 case, it wraps all the #defines with #ifndefs. This script should be run
2932 before "make dist".
2933
2934 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2935 with Unicode property support.
2936
2937 (a) For a maximizing quantifier, if the two different cases of the
2938 character were of different lengths in their UTF-8 codings (there are
2939 some cases like this - I found 11), and the matching function had to
2940 back up over a mixture of the two cases, it incorrectly assumed they
2941 were both the same length.
2942
2943 (b) When PCRE was configured to use the heap rather than the stack for
2944 recursion during matching, it was not correctly preserving the data for
2945 the other case of a UTF-8 character when checking ahead for a match
2946 while processing a minimizing repeat. If the check also involved
2947 matching a wide character, but failed, corruption could cause an
2948 erroneous result when trying to check for a repeat of the original
2949 character.
2950
2951 13. Some tidying changes to the testing mechanism:
2952
2953 (a) The RunTest script now detects the internal link size and whether there
2954 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2955 values substituted by "configure". (The RunGrepTest script already did
2956 this for UTF-8.) The configure.ac script no longer substitutes the
2957 relevant variables.
2958
2959 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2960 with length and offset values. This means that the output is different
2961 for different internal link sizes. Test 2 is skipped for link sizes
2962 other than 2 because of this, bypassing the problem. Unfortunately,
2963 there was also a test in test 3 (the locale tests) that used /B and
2964 failed for link sizes other than 2. Rather than cut the whole test out,
2965 I have added a new /Z option to pcretest that replaces the length and
2966 offset values with spaces. This is now used to make test 3 independent
2967 of link size. (Test 2 will be tidied up later.)
2968
2969 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2970 segmentation fault instead of returning the appropriate error message.
2971
2972 15. In multiline mode when the newline sequence was set to "any", the pattern
2973 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2974 This doesn't seem right; it now treats the CRLF combination as the line
2975 ending, and so does not match in that case. It's only a pattern such as ^$
2976 that would hit this one: something like ^ABC$ would have failed after \r
2977 and then tried again after \r\n.
2978
2979 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2980 in an attempt to make files that differ only in their line terminators
2981 compare equal. This works on Linux.
2982
2983 17. Under certain error circumstances pcregrep might try to free random memory
2984 as it exited. This is now fixed, thanks to valgrind.
2985
2986 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2987 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2988 was because its rules for how to advance for /g after matching an empty
2989 string at the end of a line did not allow for this case. They now check for
2990 it specially.
2991
2992 20. pcretest is supposed to handle patterns and data of any length, by
2993 extending its buffers when necessary. It was getting this wrong when the
2994 buffer for a data line had to be extended.
2995
2996 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2997 CRLF as a newline sequence.
2998
2999 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3000 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3001 I have nevertheless tidied it up.
3002
3003 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3004
3005 24. Added a man page for pcre-config.
3006
3007
3008 Version 7.0 19-Dec-06
3009 ---------------------
3010
3011 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3012 moving to gcc 4.1.1.
3013
3014 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3015 sys/time.h, which is documented as needed for this function. It doesn't
3016 seem to matter on Linux, but it showed up on some releases of OS X.
3017
3018 3. It seems that there are systems where bytes whose values are greater than
3019 127 match isprint() in the "C" locale. The "C" locale should be the
3020 default when a C program starts up. In most systems, only ASCII printing
3021 characters match isprint(). This difference caused the output from pcretest
3022 to vary, making some of the tests fail. I have changed pcretest so that:
3023
3024 (a) When it is outputting text in the compiled version of a pattern, bytes
3025 other than 32-126 are always shown as hex escapes.
3026
3027 (b) When it is outputting text that is a matched part of a subject string,
3028 it does the same, unless a different locale has been set for the match
3029 (using the /L modifier). In this case, it uses isprint() to decide.
3030
3031 4. Fixed a major bug that caused incorrect computation of the amount of memory
3032 required for a compiled pattern when options that changed within the
3033 pattern affected the logic of the preliminary scan that determines the
3034 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3035 that the computed length was too small. The symptoms of this bug were
3036 either the PCRE error "internal error: code overflow" from pcre_compile(),
3037 or a glibc crash with a message such as "pcretest: free(): invalid next
3038 size (fast)". Examples of patterns that provoked this bug (shown in
3039 pcretest format) are:
3040
3041 /(?-x: )/x
3042 /(?x)(?-x: \s*#\s*)/
3043 /((?i)[\x{c0}])/8
3044 /(?i:[\x{c0}])/8
3045
3046 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3047 is now done differently.
3048
3049 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3050 wrapper classes; (b) implement a new function in the C++ scanner that is
3051 more efficient than the old way of doing things because it avoids levels of
3052 recursion in the regex matching; (c) add a paragraph to the documentation
3053 for the FullMatch() function.
3054
3055 6. The escape sequence \n was being treated as whatever was defined as
3056 "newline". Not only was this contrary to the documentation, which states
3057 that \n is character 10 (hex 0A), but it also went horribly wrong when
3058 "newline" was defined as CRLF. This has been fixed.
3059
3060 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3061 was being set to -1 for the "end of line" case (supposedly a value that no
3062 character can have). Though this value is never used (the check for end of
3063 line is "zero bytes in current character"), it caused compiler complaints.
3064 I've changed it to 0xffffffff.
3065
3066 8. In pcre_version.c, the version string was being built by a sequence of
3067 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3068 string (as it is for production releases) called a macro with an empty
3069 argument. The C standard says the result of this is undefined. The gcc
3070 compiler treats it as an empty string (which was what was wanted) but it is
3071 reported that Visual C gives an error. The source has been hacked around to
3072 avoid this problem.
3073
3074 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3075 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3076 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3077 of them did).
3078
3079 10. Originally, pcretest opened its input and output without "b"; then I was
3080 told that "b" was needed in some environments, so it was added for release
3081 5.0 to both the input and output. (It makes no difference on Unix-like
3082 systems.) Later I was told that it is wrong for the input on Windows. I've
3083 now abstracted the modes into two macros, to make it easier to fiddle with
3084 them, and removed "b" from the input mode under Windows.
3085
3086 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3087
3088 12. Added -help and --help to pcretest as an official way of being reminded
3089 of the options.
3090
3091 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3092 and pcrecpp.cc because they annoy compilers at high warning levels.
3093
3094 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3095
3096 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3097 scripts are not C programs :-) and which was not noticed because it works
3098 on Linux.
3099
3100 16. pcretest is supposed to handle any length of pattern and data line (as one
3101 line or as a continued sequence of lines) by extending its input buffer if
3102 necessary. This feature was broken for very long pattern lines, leading to
3103 a string of junk being passed to pcre_compile() if the pattern was longer
3104 than about 50K.
3105
3106 17. I have done a major re-factoring of the way pcre_compile() computes the
3107 amount of memory needed for a compiled pattern. Previously, there was code
3108 that made a preliminary scan of the pattern in order to do this. That was
3109 OK when PCRE was new, but as the facilities have expanded, it has become
3110 harder and harder to keep it in step with the real compile phase, and there
3111 have been a number of bugs (see for example, 4 above). I have now found a
3112 cunning way of running the real compile function in a "fake" mode that
3113 enables it to compute how much memory it would need, while actually only
3114 ever using a few hundred bytes of working memory and without too many
3115 tests of the mode. This should make future maintenance and development
3116 easier. A side effect of this work is that the limit of 200 on the nesting
3117 depth of parentheses has been removed (though this was never a serious
3118 limitation, I suspect). However, there is a downside: pcre_compile() now
3119 runs more slowly than before (30% or more, depending on the pattern). I
3120 hope this isn't a big issue. There is no effect on runtime performance.
3121
3122 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3123 newline (only possible for the last line of a file) and it was a
3124 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3125
3126 19. Added additional timing features to pcretest. (1) The -tm option now times
3127 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3128 separate command line item, by a number that specifies the number of
3129 repeats to use when timing. The default is 50000; this gives better
3130 precision, but takes uncomfortably long for very large patterns.
3131
3132 20. Extended pcre_study() to be more clever in cases where a branch of a
3133 subpattern has no definite first character. For example, (a*|b*)[cd] would
3134 previously give no result from pcre_study(). Now it recognizes that the
3135 first character must be a, b, c, or d.
3136
3137 21. There was an incorrect error "recursive call could loop indefinitely" if
3138 a subpattern (or the entire pattern) that was being tested for matching an
3139 empty string contained only one non-empty item after a nested subpattern.
3140 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3141 incorrectly, because the \d was being skipped in the check.
3142
3143 22. The pcretest program now has a new pattern option /B and a command line
3144 option -b, which is equivalent to adding /B to every pattern. This causes
3145 it to show the compiled bytecode, without the additional information that
3146 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3147 is the same as /B/I).
3148
3149 23. A new optimization is now able automatically to treat some sequences such
3150 as a*b as a*+b. More specifically, if something simple (such as a character
3151 or a simple class like \d) has an unlimited quantifier, and is followed by
3152 something that cannot possibly match the quantified thing, the quantifier
3153 is automatically "possessified".
3154
3155 24. A recursive reference to a subpattern whose number was greater than 39
3156 went wrong under certain circumstances in UTF-8 mode. This bug could also
3157 have affected the operation of pcre_study().
3158
3159 25. Realized that a little bit of performance could be had by replacing
3160 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3161
3162 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3163
3164 27. Possessive quantifiers such as a++ were previously implemented by turning
3165 them into atomic groups such as ($>a+). Now they have their own opcodes,
3166 which improves performance. This includes the automatically created ones
3167 from 23 above.
3168
3169 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3170 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3171 the first matched character to be a colon. This applied both to named and
3172 numbered groups.
3173
3174 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3175
3176 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3177 building PCRE on Windows easier, so I have included it in the distribution.
3178
3179 31. There is now a check in pcretest against a ridiculously large number being
3180 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3181 loop, the loop is abandoned.
3182
3183 32. Forward references to subpatterns in conditions such as (?(2)...) where
3184 subpattern 2 is defined later cause pcre_compile() to search forwards in
3185 the pattern for the relevant set of parentheses. This search went wrong
3186 when there were unescaped parentheses in a character class, parentheses
3187 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3188
3189 33. "Subroutine" calls and backreferences were previously restricted to
3190 referencing subpatterns earlier in the regex. This restriction has now
3191 been removed.
3192
3193 34. Added a number of extra features that are going to be in Perl 5.10. On the
3194 whole, these are just syntactic alternatives for features that PCRE had
3195 previously implemented using the Python syntax or my own invention. The
3196 other formats are all retained for compatibility.
3197
3198 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3199 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3200 also .NET compatible.
3201
3202 (b) A recursion or subroutine call to a named group can now be defined as
3203 (?&name) as well as (?P>name).
3204
3205 (c) A backreference to a named group can now be defined as \k<name> or
3206 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3207 5.10, are also .NET compatible.
3208
3209 (d) A conditional reference to a named group can now use the syntax
3210 (?(<name>) or (?('name') as well as (?(name).
3211
3212 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3213 groups (named and numbered) that are never evaluated inline, but can be
3214 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3215 is always false. There may be only one alternative in such a group.
3216
3217 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3218 as the simple (?(R). The condition is true only if the most recent
3219 recursion is that of the given number or name. It does not search out
3220 through the entire recursion stack.
3221
3222 (g) The escape \gN or \g{N} has been added, where N is a positive or
3223 negative number, specifying an absolute or relative reference.
3224
3225 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3226 some "unreachable code" warnings.
3227
3228 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3229 things, this adds five new scripts.
3230
3231 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3232 There were also incompatibilities regarding the handling of \Q..\E inside
3233 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3234 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3235
3236 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3237 matches an empty string, and forcibly breaks the loop. There were bugs in
3238 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3239 against aaaa the result was just "a" rather than "aaaa", for example. Two
3240 separate and independent bugs (that affected different cases) have been
3241 fixed.
3242
3243 39. Refactored the code to abolish the use of different opcodes for small
3244 capturing bracket numbers. This is a tidy that I avoided doing when I
3245 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3246 The new approach is not only tidier, it makes it possible to reduce the
3247 memory needed to fix the previous bug (38).
3248
3249 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3250 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3251 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3252 mode.
3253
3254 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3255 report.
3256
3257 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3258 copy construction and assignment in the C++ wrapper.
3259
3260 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3261 couple of bugs that could have given wrong results in the "--newline=crlf"
3262 case.
3263
3264 44. Added a number of casts and did some reorganization of signed/unsigned int
3265 variables following suggestions from Dair Grant. Also renamed the variable
3266 "this" as "item" because it is a C++ keyword.
3267
3268 45. Arranged for dftables to add
3269
3270 #include "pcre_internal.h"
3271
3272 to pcre_chartables.c because without it, gcc 4.x may remove the array
3273 definition from the final binary if PCRE is built into a static library and
3274 dead code stripping is activated.
3275
3276 46. For an unanchored pattern, if a match attempt fails at the start of a
3277 newline sequence, and the newline setting is CRLF or ANY, and the next two
3278 characters are CRLF, advance by two characters instead of one.
3279
3280
3281 Version 6.7 04-Jul-06
3282 ---------------------
3283
3284 1. In order to handle tests when input lines are enormously long, pcretest has
3285 been re-factored so that it automatically extends its buffers when
3286 necessary. The code is crude, but this _is_ just a test program. The
3287 default size has been increased from 32K to 50K.
3288
3289 2. The code in pcre_study() was using the value of the re argument before
3290 testing it for NULL. (Of course, in any sensible call of the function, it
3291 won't be NULL.)
3292
3293 3. The memmove() emulation function in pcre_internal.h, which is used on
3294 systems that lack both memmove() and bcopy() - that is, hardly ever -
3295 was missing a "static" storage class specifier.
3296
3297 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3298 containing an extended class (one that cannot be represented by a bitmap
3299 because it contains high-valued characters or Unicode property items, e.g.
3300 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3301 pattern, but PCRE should not loop if you do not (it no longer does).
3302 [Detail: two cases were found: (a) a repeated subpattern containing an
3303 extended class; (b) a recursive reference to a subpattern that followed a
3304 previous extended class. It wasn't skipping over the extended class
3305 correctly when UTF-8 mode was not set.]
3306
3307 5. A negated single-character class was not being recognized as fixed-length
3308 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3309 compile error "lookbehind assertion is not fixed length".
3310
3311 6. The RunPerlTest auxiliary script was showing an unexpected difference
3312 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3313 write a Perl script that can interpret lines of an input file either as
3314 byte characters or as UTF-8, which is what "perltest" was being required to
3315 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3316 can't do is switch easily at run time between having the "use utf8;" pragma
3317 or not. In the end, I fudged it by using the RunPerlTest script to insert
3318 "use utf8;" explicitly for the UTF-8 tests.
3319
3320 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3321 the end of the subject string, contrary to the documentation and to what
3322 Perl does. This was true of both matching functions. Now it matches only at
3323 the start of the subject and immediately after *internal* newlines.
3324
3325 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3326 a pointer to an int instead of a pointer to an unsigned long int. This
3327 caused problems on 64-bit systems.
3328
3329 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3330 instance of the 'standard' template library not being so standard".
3331
3332 10. There was no check on the number of named subpatterns nor the maximum
3333 length of a subpattern name. The product of these values is used to compute
3334 the size of the memory block for a compiled pattern. By supplying a very
3335 long subpattern name and a large number of named subpatterns, the size
3336 computation could be caused to overflow. This is now prevented by limiting
3337 the length of names to 32 characters, and the number of named subpatterns
3338 to 10,000.
3339
3340 11. Subpatterns that are repeated with specific counts have to be replicated in
3341 the compiled pattern. The size of memory for this was computed from the
3342 length of the subpattern and the repeat count. The latter is limited to
3343 65535, but there was no limit on the former, meaning that integer overflow
3344 could in principle occur. The compiled length of a repeated subpattern is
3345 now limited to 30,000 bytes in order to prevent this.
3346
3347 12. Added the optional facility to have named substrings with the same name.
3348
3349 13. Added the ability to use a named substring as a condition, using the
3350 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3351 are numbers (not recommended). Forward references are permitted.
3352
3353 14. Added forward references in named backreferences (if you see what I mean).
3354
3355 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3356 pattern could run off the end of the subject. For example, the pattern
3357 "(?s)(.{1,5})"8 did this with the subject "ab".
3358
3359 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3360 PCRE_CASELESS was set when matching characters that were quantified with ?
3361 or *.
3362
3363 17. A character class other than a single negated character that had a minimum
3364 but no maximum quantifier - for example [ab]{6,} - was not handled
3365 correctly by pce_dfa_exec(). It would match only one character.
3366
3367 18. A valid (though odd) pattern that looked like a POSIX character
3368 class but used an invalid character after [ (for example [[,abc,]]) caused
3369 pcre_compile() to give the error "Failed: internal error: code overflow" or
3370 in some cases to crash with a glibc free() error. This could even happen if
3371 the pattern terminated after [[ but there just happened to be a sequence of
3372 letters, a binary zero, and a closing ] in the memory that followed.
3373
3374 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3375 over the years. Originally (before any Unicode support), just the bottom 8
3376 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3377 output from "man perlunicode" includes this:
3378
3379 The regular expression compiler produces polymorphic opcodes. That
3380 is, the pattern adapts to the data and automatically switches to
3381 the Unicode character scheme when presented with Unicode data--or
3382 instead uses a traditional byte scheme when presented with byte
3383 data.
3384
3385 Sadly, a wide octal escape does not cause a switch, and in a string with
3386 no other multibyte characters, these octal escapes are treated as before.
3387 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3388 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3389 Unicode string.
3390
3391 I have not perpetrated such confusion in PCRE. Up till now, it took just
3392 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3393 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3394 translate to the appropriate multibyte character.
3395
3396 29. Applied some refactoring to reduce the number of warnings from Microsoft
3397 and Borland compilers. This has included removing the fudge introduced
3398 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3399 a warning about an unused variable.
3400
3401 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3402 characters since release 4.0, because Perl (from release 5.004) does not.
3403 [Or at least, is documented not to: some releases seem to be in conflict
3404 with the documentation.] However, when a pattern was studied with
3405 pcre_study() and all its branches started with \s, PCRE still included VT
3406 as a possible starting character. Of course, this did no harm; it just
3407 caused an unnecessary match attempt.
3408
3409 22. Removed a now-redundant internal flag bit that recorded the fact that case
3410 dependency changed within the pattern. This was once needed for "required
3411 byte" processing, but is no longer used. This recovers a now-scarce options
3412 bit. Also moved the least significant internal flag bit to the most-
3413 significant bit of the word, which was not previously used (hangover from
3414 the days when it was an int rather than a uint) to free up another bit for
3415 the future.
3416
3417 23. Added support for CRLF line endings as well as CR and LF. As well as the
3418 default being selectable at build time, it can now be changed at runtime
3419 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3420 specify that it is scanning data with non-default line endings.
3421
3422 24. Changed the definition of CXXLINK to make it agree with the definition of
3423 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3424
3425 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3426 recursions. This makes a big different to stack usage for some patterns.
3427
3428 26. If a subpattern containing a named recursion or subroutine reference such
3429 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3430 the space required for the compiled pattern went wrong and gave too small a
3431 value. Depending on the environment, this could lead to "Failed: internal
3432 error: code overflow at offset 49" or "glibc detected double free or
3433 corruption" errors.
3434
3435 27. Applied patches from Google (a) to support the new newline modes and (b) to
3436 advance over multibyte UTF-8 characters in GlobalReplace.
3437
3438 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3439 difference for some implementation of PCRE in some Windows version.
3440
3441 29. Added some extra testing facilities to pcretest:
3442
3443 \q<number> in a data line sets the "match limit" value
3444 \Q<number> in a data line sets the "match recursion limt" value
3445 -S <number> sets the stack size, where <number> is in megabytes
3446
3447 The -S option isn't available for Windows.
3448
3449
3450 Version 6.6 06-Feb-06
3451 ---------------------
3452
3453 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3454 in pcreposix.h. I have copied the definition from pcre.h.
3455
3456 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3457 because pcre.h is no longer a built file.
3458
3459 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3460 not normally included in the compiled code.
3461
3462
3463 Version 6.5 01-Feb-06
3464 ---------------------
3465
3466 1. When using the partial match feature with pcre_dfa_exec(), it was not
3467 anchoring the second and subsequent partial matches at the new starting
3468 point. This could lead to incorrect results. For example, with the pattern
3469 /1234/, partially matching against "123" and then "a4" gave a match.
3470
3471 2. Changes to pcregrep:
3472
3473 (a) All non-match returns from pcre_exec() were being treated as failures
3474 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3475 error message is output. Some extra information is given for the
3476 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3477 probably the only errors that are likely to be caused by users (by
3478 specifying a regex that has nested indefinite repeats, for instance).
3479 If there are more than 20 of these errors, pcregrep is abandoned.
3480
3481 (b) A binary zero was treated as data while matching, but terminated the
3482 output line if it was written out. This has been fixed: binary zeroes
3483 are now no different to any other data bytes.
3484
3485 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3486 used to set a locale for matching. The --locale=xxxx long option has
3487 been added (no short equivalent) to specify a locale explicitly on the
3488 pcregrep command, overriding the environment variables.
3489
3490 (d) When -B was used with -n, some line numbers in the output were one less
3491 than they should have been.
3492
3493 (e) Added the -o (--only-matching) option.
3494
3495 (f) If -A or -C was used with -c (count only), some lines of context were
3496 accidentally printed for the final match.
3497
3498 (g) Added the -H (--with-filename) option.
3499
3500 (h) The combination of options -rh failed to suppress file names for files
3501 that were found from directory arguments.
3502
3503 (i) Added the -D (--devices) and -d (--directories) options.
3504
3505 (j) Added the -F (--fixed-strings) option.
3506
3507 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3508
3509 (l) Added the --colo(u)r option.
3510
3511 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3512 is not present by default.
3513
3514 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3515 items such as (?R) or (?1), when the recursion could match a number of
3516 alternatives. If it matched one of the alternatives, but subsequently,
3517 outside the recursion, there was a failure, the code tried to back up into
3518 the recursion. However, because of the way PCRE is implemented, this is not
3519 possible, and the result was an incorrect result from the match.
3520
3521 In order to prevent this happening, the specification of recursion has
3522 been changed so that all such subpatterns are automatically treated as
3523 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3524
3525 4. I had overlooked the fact that, in some locales, there are characters for
3526 which isalpha() is true but neither isupper() nor islower() are true. In
3527 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3528 and ordfeminine) are like this. This affected the treatment of \w and \W
3529 when they appeared in character classes, but not when they appeared outside
3530 a character class. The bit map for "word" characters is now created
3531 separately from the results of isalnum() instead of just taking it from the
3532 upper, lower, and digit maps. (Plus the underscore character, of course.)
3533
3534 5. The above bug also affected the handling of POSIX character classes such as
3535 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3536 permanent tables. Instead, the bit maps for such a class were previously
3537 created as the appropriate unions of the upper, lower, and digit bitmaps.
3538 Now they are created by subtraction from the [[:word:]] class, which has
3539 its own bitmap.
3540
3541 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3542 It is created by subtracting the vertical space characters (\x09, \x0a,
3543 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3544 subtraction was done in the overall bitmap for a character class, meaning
3545 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3546 be recognized. This bug has been fixed.
3547
3548 7. Patches from the folks at Google:
3549
3550 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3551 real life, but is still worth protecting against".
3552
3553 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3554 regular expressions".
3555
3556 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3557 have it.
3558
3559 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3560 "configure" and the latter not, in order to fix a problem somebody had
3561 with compiling the Arg class on HP-UX.
3562
3563 (e) Improve the error-handling of the C++ wrapper a little bit.
3564
3565 (f) New tests for checking recursion limiting.
3566
3567 8. The pcre_memmove() function, which is used only if the environment does not
3568 have a standard memmove() function (and is therefore rarely compiled),
3569 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3570 returning a result (though PCRE never actually uses the result).
3571
3572 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3573 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3574 returned instead of calling malloc() with an overflowing number that would
3575 most likely cause subsequent chaos.
3576
3577 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3578
3579 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3580 with this option is matched, the nmatch and pmatch options of regexec() are
3581 ignored.
3582
3583 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3584 provided in case anyone wants to the the POSIX interface with UTF-8
3585 strings.
3586
3587 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3588 C++ linking (needed for some HP-UX environments).
3589
3590 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3591 (unused parameter) and in the pcre_printint() function (omitted "default"
3592 switch label when the default is to do nothing).
3593
3594 15. Added some code to make it possible, when PCRE is compiled as a C++
3595 library, to replace subject pointers for pcre_exec() with a smart pointer
3596 class, thus making it possible to process discontinuous strings.
3597
3598 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3599 much the same function. They were added by different people who were trying
3600 to make PCRE easy to compile on non-Unix systems. It has been suggested
3601 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3602 for compiling on Windows systems. I have therefore replaced it with
3603 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3604 defaults to "extern" for C or "extern C" for C++, which works fine on
3605 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3606 SCOPE with something explicit in config.h. In addition:
3607
3608 (a) pcreposix.h still had just "extern" instead of either of these macros;
3609 I have replaced it with PCRE_DATA_SCOPE.
3610
3611 (b) Functions such as _pcre_xclass(), which are internal to the library,
3612 but external in the C sense, all had PCRE_EXPORT in their definitions.
3613 This is apparently wrong for the Windows case, so I have removed it.
3614 (It makes no difference on Unix-like systems.)
3615
3616 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3617 of recursive calls to match(). This is different to MATCH_LIMIT because
3618 that limits the total number of calls to match(), not all of which increase
3619 the depth of recursion. Limiting the recursion depth limits the amount of
3620 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3621 when PCRE is compiled, and changed at run time. A patch from Google adds
3622 this functionality to the C++ interface.
3623
3624 18. Changes to the handling of Unicode character properties:
3625
3626 (a) Updated the table to Unicode 4.1.0.
3627
3628 (b) Recognize characters that are not in the table as "Cn" (undefined).
3629
3630 (c) I revised the way the table is implemented to a much improved format
3631 which includes recognition of ranges. It now supports the ranges that
3632 are defined in UnicodeData.txt, and it also amalgamates other
3633 characters into ranges. This has reduced the number of entries in the
3634 table from around 16,000 to around 3,000, thus reducing its size
3635 considerably. I realized I did not need to use a tree structure after
3636 all - a binary chop search is just as efficient. Having reduced the
3637 number of entries, I extended their size from 6 bytes to 8 bytes to
3638 allow for more data.
3639
3640 (d) Added support for Unicode script names via properties such as \p{Han}.
3641
3642 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3643 matching that character.
3644
3645 20. When matching a repeated Unicode property with a minimum greater than zero,
3646 (for example \pL{2,}), PCRE could look past the end of the subject if it
3647 reached it while seeking the minimum number of characters. This could
3648 happen only if some of the characters were more than one byte long, because
3649 there is a check for at least the minimum number of bytes.
3650
3651 21. Refactored the implementation of \p and \P so as to be more general, to
3652 allow for more different types of property in future. This has changed the
3653 compiled form incompatibly. Anybody with saved compiled patterns that use
3654 \p or \P will have to recompile them.
3655
3656 22. Added "Any" and "L&" to the supported property types.
3657
3658 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3659 but give a compile time error if the value is greater than 0xff.
3660
3661 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3662 accidentally not being installed or uninstalled.
3663
3664 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3665 made were to insert the current release number. This seemed silly, because
3666 it made things harder for people building PCRE on systems that don't run
3667 "configure". I have turned pcre.h into a distributed file, no longer built
3668 by "configure", with the version identification directly included. There is
3669 no longer a pcre.h.in file.
3670
3671 However, this change necessitated a change to the pcre-config script as
3672 well. It is built from pcre-config.in, and one of the substitutions was the
3673 release number. I have updated configure.ac so that ./configure now finds
3674 the release number by grepping pcre.h.
3675
3676 26. Added the ability to run the tests under valgrind.
3677
3678
3679 Version 6.4 05-Sep-05
3680 ---------------------
3681
3682 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3683 "--" to be printed when multiple files were scanned, even when none of the
3684 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3685 consider it to be a bug, and have restored the previous behaviour.
3686
3687 2. A couple of code tidies to get rid of compiler warnings.
3688
3689 3. The pcretest program used to cheat by referring to symbols in the library
3690 whose names begin with _pcre_. These are internal symbols that are not
3691 really supposed to be visible externally, and in some environments it is
3692 possible to suppress them. The cheating is now confined to including
3693 certain files from the library's source, which is a bit cleaner.
3694
3695 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3696 file's purpose clearer.
3697
3698 5. Reorganized pcre_ucp_findchar().
3699
3700
3701 Version 6.3 15-Aug-05
3702 ---------------------
3703
3704 1. The file libpcre.pc.in did not have general read permission in the tarball.
3705
3706 2. There were some problems when building without C++ support:
3707
3708 (a) If C++ support was not built, "make install" and "make test" still
3709 tried to test it.
3710
3711 (b) There were problems when the value of CXX was explicitly set. Some
3712 changes have been made to try to fix these, and ...
3713
3714 (c) --disable-cpp can now be used to explicitly disable C++ support.
3715
3716 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3717 backslash in a target when C++ was disabled. This confuses some
3718 versions of "make", apparently. Using an intermediate variable solves
3719 this. (Same for CPP_LOBJ.)
3720
3721 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3722 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3723 necessary on certain architectures.
3724
3725 4. Added a setting of -export-symbols-regex to the link command to remove
3726 those symbols that are exported in the C sense, but actually are local
3727 within the library, and not documented. Their names all begin with
3728 "_pcre_". This is not a perfect job, because (a) we have to except some
3729 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3730 available (and never for static libraries). I have made a note to try to
3731 find a way round (a) in the future.
3732
3733
3734 Version 6.2 01-Aug-05
3735 ---------------------
3736
3737 1. There was no test for integer overflow of quantifier values. A construction
3738 such as {1111111111111111} would give undefined results. What is worse, if
3739 a minimum quantifier for a parenthesized subpattern overflowed and became
3740 negative, the calculation of the memory size went wrong. This could have
3741 led to memory overwriting.
3742
3743 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3744
3745 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3746 operating environments where this matters.
3747
3748 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3749 PCRE options from within the C++ wrapper.
3750
3751 5. Named capturing subpatterns were not being correctly counted when a pattern
3752 was compiled. This caused two problems: (a) If there were more than 100
3753 such subpatterns, the calculation of the memory needed for the whole
3754 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3755 back references of the form \12, where the number was greater than 9, were
3756 not recognized as back references, even though there were sufficient
3757 previous subpatterns.
3758
3759 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3760 versions of gcc, e.g. 2.95.4.
3761
3762
3763 Version 6.1 21-Jun-05
3764 ---------------------
3765
3766 1. There was one reference to the variable "posix" in pcretest.c that was not
3767 surrounded by "#if !defined NOPOSIX".
3768
3769 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3770 the cross-check on the old pcre_info() function, for the benefit of the
3771 cut-down version of PCRE that is currently imported into Exim.
3772
3773 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3774 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3775 patterns that start (?i) but I don't think that matters. The use of (?i) is
3776 just an example; this all applies to the other options as well.
3777
3778 4. Since libtool seems to echo the compile commands it is issuing, the output
3779 from "make" can be reduced a bit by putting "@" in front of each libtool
3780 compile command.
3781
3782 5. Patch from the folks at Google for configure.in to be a bit more thorough
3783 in checking for a suitable C++ installation before trying to compile the
3784 C++ stuff. This should fix a reported problem when a compiler was present,
3785 but no suitable headers.
3786
3787 6. The man pages all had just "PCRE" as their title. I have changed them to
3788 be the relevant file name. I have also arranged that these names are
3789 retained in the file doc/pcre.txt, which is a concatenation in text format
3790 of all the man pages except the little individual ones for each function.
3791
3792 7. The NON-UNIX-USE file had not been updated for the different set of source
3793 files that come with release 6. I also added a few comments about the C++
3794 wrapper.
3795
3796
3797 Version 6.0 07-Jun-05
3798 ---------------------
3799
3800 1. Some minor internal re-organization to help with my DFA experiments.
3801
3802 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3803 didn't matter for the library itself when fully configured, but did matter
3804 when compiling without UCP support, or within Exim, where the ucp files are
3805 not imported.
3806
3807 3. Refactoring of the library code to split up the various functions into
3808 different source modules. The addition of the new DFA matching code (see
3809 below) to a single monolithic source would have made it really too
3810 unwieldy, quite apart from causing all the code to be include in a
3811 statically linked application, when only some functions are used. This is
3812 relevant even without the DFA addition now that patterns can be compiled in
3813 one application and matched in another.
3814
3815 The downside of splitting up is that there have to be some external
3816 functions and data tables that are used internally in different modules of
3817 the library but which are not part of the API. These have all had their
3818 names changed to start with "_pcre_" so that they are unlikely to clash
3819 with other external names.
3820
3821 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3822 a different (DFA) algorithm. Although it is slower than the original
3823 function, it does have some advantages for certain types of matching
3824 problem.
3825
3826 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3827 including restarting after a partial match.
3828
3829 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3830 defined when compiling for Windows was sent to me. I have put it into the
3831 code, though I have no means of testing or verifying it.
3832
3833 7. Added the pcre_refcount() auxiliary function.
3834
3835 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3836 match before or at the first newline in the subject string. In pcretest,
3837 the /f option on a pattern can be used to set this.
3838
3839 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3840 would behave wrongly. This has been present in PCRE since release 4.0.
3841
3842 10. A number of changes to the pcregrep command:
3843
3844 (a) Refactored how -x works; insert ^(...)$ instead of setting
3845 PCRE_ANCHORED and checking the length, in preparation for adding
3846 something similar for -w.
3847
3848 (b) Added the -w (match as a word) option.
3849
3850 (c) Refactored the way lines are read and buffered so as to have more
3851 than one at a time available.
3852
3853 (d) Implemented a pcregrep test script.
3854
3855 (e) Added the -M (multiline match) option. This allows patterns to match
3856 over several lines of the subject. The buffering ensures that at least
3857 8K, or the rest of the document (whichever is the shorter) is available
3858 for matching (and similarly the previous 8K for lookbehind assertions).
3859
3860 (f) Changed the --help output so that it now says
3861
3862 -w, --word-regex(p)
3863
3864 instead of two lines, one with "regex" and the other with "regexp"
3865 because that confused at least one person since the short forms are the
3866 same. (This required a bit of code, as the output is generated
3867 automatically from a table. It wasn't just a text change.)
3868
3869 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3870 option but starts with a hyphen. Could be a pattern or a path name
3871 starting with a hyphen, for instance.
3872
3873 (h) "-" can be given as a file name to represent stdin.
3874
3875 (i) When file names are being printed, "(standard input)" is used for
3876 the standard input, for compatibility with GNU grep. Previously
3877 "<stdin>" was used.
3878
3879 (j) The option --label=xxx can be used to supply a name to be used for
3880 stdin when file names are being printed. There is no short form.
3881
3882 (k) Re-factored the options decoding logic because we are going to add
3883 two more options that take data. Such options can now be given in four
3884 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3885
3886 (l) Added the -A, -B, and -C options for requesting that lines of context
3887 around matches be printed.
3888
3889 (m) Added the -L option to print the names of files that do not contain
3890 any matching lines, that is, the complement of -l.
3891
3892 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3893 continue to scan other files.
3894
3895 (o) The -s option was incorrectly implemented. For compatibility with other
3896 greps, it now suppresses the error message for a non-existent or non-
3897 accessible file (but not the return code). There is a new option called
3898 -q that suppresses the output of matching lines, which was what -s was
3899 previously doing.
3900
3901 (p) Added --include and --exclude options to specify files for inclusion
3902 and exclusion when recursing.
3903
3904 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3905 Hopefully, it now does.
3906
3907 12. Missing cast in pcre_study().
3908
3909 13. Added an "uninstall" target to the makefile.
3910
3911 14. Replaced "extern" in the function prototypes in Makefile.in with
3912 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3913 world, but is set differently for Windows.
3914
3915 15. Added a second compiling function called pcre_compile2(). The only
3916 difference is that it has an extra argument, which is a pointer to an
3917 integer error code. When there is a compile-time failure, this is set
3918 non-zero, in addition to the error test pointer being set to point to an
3919 error message. The new argument may be NULL if no error number is required
3920 (but then you may as well call pcre_compile(), which is now just a
3921 wrapper). This facility is provided because some applications need a
3922 numeric error indication, but it has also enabled me to tidy up the way
3923 compile-time errors are handled in the POSIX wrapper.
3924
3925 16. Added VPATH=.libs to the makefile; this should help when building with one
3926 prefix path and installing with another. (Or so I'm told by someone who
3927 knows more about this stuff than I do.)
3928
3929 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3930 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3931 match everything, including newlines. This is not POSIX-compatible, but
3932 somebody wanted the feature. From pcretest it can be activated by using
3933 both the P and the s flags.
3934
3935 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3936
3937 19. libpcre.pc was being incorrectly installed as executable.
3938
3939 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3940 it now also looks for '\r' so that it will work unmodified on Windows.
3941
3942 21. Added Google's contributed C++ wrapper to the distribution.
3943
3944 22. Added some untidy missing memory free() calls in pcretest, to keep
3945 Electric Fence happy when testing.
3946
3947
3948
3949 Version 5.0 13-Sep-04
3950 ---------------------
3951
3952 1. Internal change: literal characters are no longer packed up into items
3953 containing multiple characters in a single byte-string. Each character
3954 is now matched using a separate opcode. However, there may be more than one
3955 byte in the character in UTF-8 mode.
3956
3957 2. The pcre_callout_block structure has two new fields: pattern_position and
3958 next_item_length. These contain the offset in the pattern to the next match
3959 item, and its length, respectively.
3960
3961 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3962 insertion of callouts before each pattern item. Added the /C option to
3963 pcretest to make use of this.
3964
3965 4. On the advice of a Windows user, the lines
3966
3967 #if defined(_WIN32) || defined(WIN32)
3968 _setmode( _fileno( stdout ), 0x8000 );
3969 #endif /* defined(_WIN32) || defined(WIN32) */
3970
3971 have been added to the source of pcretest. This apparently does useful
3972 magic in relation to line terminators.
3973
3974 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3975 for the benefit of those environments where the "b" makes a difference.
3976
3977 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3978 to know about it. I have put a hack into configure.in that adds in code
3979 to set GCC=yes if CC=icc. This seems to end up at a point in the
3980 generated configure script that is early enough to affect the setting of
3981 compiler options, which is what is needed, but I have no means of testing
3982 whether it really works. (The user who reported this had patched the
3983 generated configure script, which of course I cannot do.)
3984
3985 LATER: After change 22 below (new libtool files), the configure script
3986 seems to know about icc (and also ecc). Therefore, I have commented out
3987 this hack in configure.in.
3988
3989 7. Added support for pkg-config (2 patches were sent in).
3990
3991 8. Negated POSIX character classes that used a combination of internal tables
3992 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3993 [[:^ascii]]. Typically, they would match almost any characters. The other
3994 POSIX classes were not broken in this way.
3995
3996 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3997 to find the match, as PCRE was deluded into thinking that the match had to
3998 start at the start point or following a newline. The same bug applied to
3999 patterns with negative forward assertions or any backward assertions
4000 preceding ".*" at the start, unless the pattern required a fixed first
4001 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4002
4003 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4004 starting at the last subject character, bytes beyond the end of the subject
4005 string were read.
4006
4007 11. Renamed the variable "class" as "classbits" to make life easier for C++
4008 users. (Previously there was a macro definition, but it apparently wasn't
4009 enough.)
4010
4011 12. Added the new field "tables" to the extra data so that tables can be passed
4012 in at exec time, or the internal tables can be re-selected. This allows
4013 a compiled regex to be saved and re-used at a later time by a different
4014 program that might have everything at different addresses.
4015
4016 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4017 -R library as well as a -L library.
4018
4019 14. The debugging options of pcretest (-d on the command line or D on a
4020 pattern) showed incorrect output for anything following an extended class
4021 that contained multibyte characters and which was followed by a quantifier.
4022
4023 15. Added optional support for general category Unicode character properties
4024 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4025 support. It adds about 90K to the size of the library. The meanings of the
4026 inbuilt class escapes such as \d and \s have NOT been changed.
4027
4028 16. Updated pcredemo.c to include calls to free() to release the memory for the
4029 compiled pattern.
4030
4031 17. The generated file chartables.c was being created in the source directory
4032 instead of in the building directory. This caused the build to fail if the
4033 source directory was different from the building directory, and was
4034 read-only.
4035
4036 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4037 file. No doubt somebody will tell me if they don't make sense... Also added
4038 Dan Mooney's comments about building on OpenVMS.
4039
4040 19. Added support for partial matching via the PCRE_PARTIAL option for
4041 pcre_exec() and the \P data escape in pcretest.
4042
4043 20. Extended pcretest with 3 new pattern features:
4044
4045 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4046 write the compiled pattern to the file whose name is "rest-of-line".
4047 This is a straight binary dump of the data, with the saved pointer to
4048 the character tables forced to be NULL. The study data, if any, is
4049 written too. After writing, pcretest reads a new pattern.
4050
4051 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4052 compiled pattern from the given file. There must not be any
4053 occurrences of "<" in the file name (pretty unlikely); if there are,
4054 pcretest will instead treat the initial "<" as a pattern delimiter.
4055 After reading in the pattern, pcretest goes on to read data lines as
4056 usual.
4057
4058 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4059 and 16-bit fields in a compiled pattern, to simulate a pattern that
4060 was compiled on a host of opposite endianness.
4061
4062 21. The pcre-exec() function can now cope with patterns that were compiled on
4063 hosts of opposite endianness, with this restriction:
4064
4065 As for any compiled expression that is saved and used later, the tables
4066 pointer field cannot be preserved; the extra_data field in the arguments
4067 to pcre_exec() should be used to pass in a tables address if a value
4068 other than the default internal tables were used at compile time.
4069
4070 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4071 now diagnosed as an error. Previously, most of the time, a negative number
4072 would have been treated as zero, but if in addition "ovector" was passed as
4073 NULL, a crash could occur.
4074
4075 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4076 new versions from the libtool 1.5 distribution (the last one is a copy of
4077 a file called libtool.m4). This seems to have fixed the need to patch
4078 "configure" to support Darwin 1.3 (which I used to do). However, I still
4079 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4080 workstation).
4081
4082 24. Changed the PCRE licence to be the more standard "BSD" licence.
4083
4084
4085 Version 4.5 01-Dec-03
4086 ---------------------
4087
4088 1. There has been some re-arrangement of the code for the match() function so
4089 that it can be compiled in a version that does not call itself recursively.
4090 Instead, it keeps those local variables that need separate instances for
4091 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4092 needs to "recurse". Keeping track of where control must go is done by means
4093 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4094 hide most of the details from the main code, and operates only if
4095 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4096 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4097 operating.
4098
4099 To make it easier for callers to provide specially tailored get/free
4100 functions for this usage, two new functions, pcre_stack_malloc, and
4101 pcre_stack_free, are used. They are always called in strict stacking order,
4102 and the size of block requested is always the same.
4103
4104 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4105 PCRE has been compiled to use the stack or the heap for recursion. The
4106 -C option of pcretest uses this to show which version is compiled.
4107
4108 A new data escape \S, is added to pcretest; it causes the amounts of store
4109 obtained and freed by both kinds of malloc/free at match time to be added
4110 to the output.
4111
4112 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4113 what's available on my current Linux desktop machine.
4114
4115 3. When matching a UTF-8 string, the test for a valid string at the start has
4116 been extended. If start_offset is not zero, PCRE now checks that it points
4117 to a byte that is the start of a UTF-8 character. If not, it returns
4118 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4119 this is necessary because there may be backward assertions in the pattern.
4120 When matching the same subject several times, it may save resources to use
4121 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4122
4123 4. The code for checking the validity of UTF-8 strings has been tightened so
4124 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4125 containing "overlong sequences".
4126
4127 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4128 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4129 should have been "&", but it just so happened that all the cases this let
4130 through by mistake were picked up later in the function.
4131
4132 6. I had used a variable called "isblank" - this is a C99 function, causing
4133 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4134
4135 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4136 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4137 so the version is shown; (c) stop "make test" echoing "./RunTest".
4138
4139 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4140
4141 9. The prototype for memmove() for systems that don't have it was using
4142 size_t, but the inclusion of the header that defines size_t was later. I've
4143 moved the #includes for the C headers earlier to avoid this.
4144
4145 10. Added some adjustments to the code to make it easier to compiler on certain
4146 special systems:
4147
4148 (a) Some "const" qualifiers were missing.
4149 (b) Added the macro EXPORT before all exported functions; by default this
4150 is defined to be empty.
4151 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4152 that it reads its output file name as an argument instead of writing
4153 to the standard output and assuming this can be redirected.
4154
4155 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4156 class containing characters with values greater than 255, PCRE compilation
4157 went into a loop.
4158
4159 12. A recursive reference to a subpattern that was within another subpattern
4160 that had a minimum quantifier of zero caused PCRE to crash. For example,
4161 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4162 recursion. If the recursively-called subpattern itself had a zero repeat,
4163 that was OK.
4164
4165 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4166 buffer into which it was copied (for escape processing) was still set at
4167 1024, so long lines caused crashes.
4168
4169 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4170 "internal error: code overflow...". This applied to any character class
4171 that was followed by a possessive quantifier.
4172
4173 15. Modified the Makefile to add libpcre.la as a prerequisite for
4174 libpcreposix.la because I was told this is needed for a parallel build to
4175 work.
4176
4177 16. If a pattern that contained .* following optional items at the start was
4178 studied, the wrong optimizing data was generated, leading to matching
4179 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4180 matching string must start with a or b or c. The correct conclusion for
4181 this pattern is that a match can start with any character.
4182
4183
4184 Version 4.4 13-Aug-03
4185 ---------------------
4186
4187 1. In UTF-8 mode, a character class containing characters with values between
4188 127 and 255 was not handled correctly if the compiled pattern was studied.
4189 In fixing this, I have also improved the studying algorithm for such
4190 classes (slightly).
4191
4192 2. Three internal functions had redundant arguments passed to them. Removal
4193 might give a very teeny performance improvement.
4194
4195 3. Documentation bug: the value of the capture_top field in a callout is *one
4196 more than* the number of the hightest numbered captured substring.
4197
4198 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4199 in incorrectly linking with a previously installed version. They now link
4200 explicitly with libpcre.la.
4201
4202 5. configure.in no longer needs to recognize Cygwin specially.
4203
4204 6. A problem in pcre.in for Windows platforms is fixed.
4205
4206 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4207 pcretest, it used to include the size of the study block as part of its
4208 output. Unfortunately, the structure contains a field that has a different
4209 size on different hardware architectures. This meant that the tests that
4210 showed this size failed. As the block is currently always of a fixed size,
4211 this information isn't actually particularly useful in pcretest output, so
4212 I have just removed it.
4213
4214 8. Three pre-processor statements accidentally did not start in column 1.
4215 Sadly, there are *still* compilers around that complain, even though
4216 standard C has not required this for well over a decade. Sigh.
4217
4218 9. In pcretest, the code for checking callouts passed small integers in the
4219 callout_data field, which is a void * field. However, some picky compilers
4220 complained about the casts involved for this on 64-bit systems. Now
4221 pcretest passes the address of the small integer instead, which should get
4222 rid of the warnings.
4223
4224 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4225 both compile and run time, and gives an error if an invalid UTF-8 sequence
4226 is found. There is a option for disabling this check in cases where the
4227 string is known to be correct and/or the maximum performance is wanted.
4228
4229 11. In response to a bug report, I changed one line in Makefile.in from
4230
4231 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4232 to
4233 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4234
4235 to look similar to other lines, but I have no way of telling whether this
4236 is the right thing to do, as I do not use Windows. No doubt I'll get told
4237 if it's wrong...
4238
4239
4240 Version 4.3 21-May-03
4241 ---------------------
4242
4243 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4244 Makefile.
4245
4246 2. Some refactoring to improve the quality of the code:
4247
4248 (i) The utf8_table... variables are now declared "const".
4249
4250 (ii) The code for \cx, which used the "case flipping" table to upper case
4251 lower case letters, now just substracts 32. This is ASCII-specific,
4252 but the whole concept of \cx is ASCII-specific, so it seems
4253 reasonable.
4254
4255 (iii) PCRE was using its character types table to recognize decimal and
4256 hexadecimal digits in the pattern. This is silly, because it handles
4257 only 0-9, a-f, and A-F, but the character types table is locale-
4258 specific, which means strange things might happen. A private
4259 table is now used for this - though it costs 256 bytes, a table is
4260 much faster than multiple explicit tests. Of course, the standard
4261 character types table is still used for matching digits in subject
4262 strings against \d.
4263
4264 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4265 ending in _t are). So I've renamed it as ESC_tee.
4266
4267 3. The first argument for regexec() in the POSIX wrapper should have been
4268 defined as "const".
4269
4270 4. Changed pcretest to use malloc() for its buffers so that they can be
4271 Electric Fenced for debugging.
4272
4273 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4274 to read one or more bytes before the start of the subject string. Often this
4275 had no effect on PCRE's behaviour, but in some circumstances it could
4276 provoke a segmentation fault.
4277
4278 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4279 to try to read one or more bytes before the start of the subject string.
4280
4281 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4282 UTF-8 support could misbehave in various ways if the subject string
4283 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4284 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4285 back over UTF-8 characters.)
4286
4287
4288 Version 4.2 14-Apr-03
4289 ---------------------
4290
4291 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4292
4293 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4294 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4295 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4296 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4297 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4298 and BUILD_EXEEXT
4299 Note: automatic setting of the BUILD variables is not yet working
4300 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4301 compile-time but not at link-time
4302 [LINK]: use for linking executables only
4303 make different versions for Windows and non-Windows
4304 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4305 libraries
4306 [LINK_FOR_BUILD]: new variable
4307 [OBJEXT]: use throughout
4308 [EXEEXT]: use throughout
4309 <winshared>: new target
4310 <wininstall>: new target
4311 <dftables.o>: use native compiler
4312 <dftables>: use native linker
4313 <install>: handle Windows platform correctly
4314 <clean>: ditto
4315 <check>: ditto
4316 copy DLL to top builddir before testing
4317
4318 As part of these changes, -no-undefined was removed again. This was reported
4319 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4320 in any case.
4321
4322 3. Some tidies to get rid of compiler warnings:
4323
4324 . In the match_data structure, match_limit was an unsigned long int, whereas
4325 match_call_count was an int. I've made them both unsigned long ints.
4326
4327 . In pcretest the fact that a const uschar * doesn't automatically cast to
4328 a void * provoked a warning.
4329
4330 . Turning on some more compiler warnings threw up some "shadow" variables
4331 and a few more missing casts.
4332
4333 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4334 option, a class that contained a single character with a value between 128
4335 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4336
4337 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4338 option, a class that contained several characters, but with at least one
4339 whose value was between 128 and 255 caused PCRE to crash.
4340
4341
4342 Version 4.1 12-Mar-03
4343 ---------------------
4344
4345 1. Compiling with gcc -pedantic found a couple of places where casts were
4346 needed, and a string in dftables.c that was longer than standard compilers are
4347 required to support.
4348
4349 2. Compiling with Sun's compiler found a few more places where the code could
4350 be tidied up in order to avoid warnings.
4351
4352 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4353 first of these names is deprecated in the latest Autoconf in favour of the name
4354 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4355 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4356 analogy I have changed it to CFLAGS_FOR_BUILD.
4357
4358 4. Added -no-undefined to the linking command in the Makefile, because this is
4359 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4360 linking step for the pcreposix library.
4361
4362 5. PCRE was failing to diagnose the case of two named groups with the same
4363 name.
4364
4365 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4366 literal character that is needed in the subject for a match, and scans along to
4367 ensure that it is present before embarking on the full matching process. This
4368 saves time in cases of nested unlimited repeats that are never going to match.
4369 Problem: the scan can take a lot of time if the subject is very long (e.g.
4370 megabytes), thus penalizing straightforward matches. It is now done only if the
4371 amount of subject to be scanned is less than 1000 bytes.
4372
4373 7. A lesser problem with the same optimization is that it was recording the
4374 first character of an anchored pattern as "needed", thus provoking a search
4375 right along the subject, even when the first match of the pattern was going to
4376 fail. The "needed" character is now not set for anchored patterns, unless it
4377 follows something in the pattern that is of non-fixed length. Thus, it still
4378 fulfils its original purpose of finding quick non-matches in cases of nested
4379 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4380
4381
4382 Version 4.0 17-Feb-03
4383 ---------------------
4384
4385 1. If a comment in an extended regex that started immediately after a meta-item
4386 extended to the end of string, PCRE compiled incorrect data. This could lead to
4387 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4388
4389 2. Moved to autoconf 2.53 and libtool 1.4.2.
4390
4391 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4392 the special perltest8 script is no longer needed - all the tests can be run
4393 from a single perltest script.
4394
4395 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4396 by \s. It has now been removed in PCRE. This means it isn't recognized as
4397 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4398 class [:space:] *does* include VT, thereby creating a mess.
4399
4400 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4401 space and tab.
4402
4403 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4404 its new features into the main test script, reducing the number of scripts.
4405
4406 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4407 were backward compatible, and made the (?i) apply to the whole pattern, as if
4408 /i were given. Now it behaves more logically, and applies the option setting
4409 only to what follows. PCRE has been changed to follow suit. However, if it
4410 finds options settings right at the start of the pattern, it extracts them into
4411 the global options, as before. Thus, they show up in the info data.
4412
4413 8. Added support for the \Q...\E escape sequence. Characters in between are
4414 treated as literals. This is slightly different from Perl in that $ and @ are
4415 also handled as literals inside the quotes. In Perl, they will cause variable
4416 interpolation. Note the following examples:
4417
4418 Pattern PCRE matches Perl matches
4419
4420 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4421 \Qabc\$xyz\E abc\$xyz abc\$xyz
4422 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4423
4424 For compatibility with Perl, \Q...\E sequences are recognized inside character
4425 classes as well as outside them.
4426
4427 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4428 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4429 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4430 signed/unsigned warnings.
4431
4432 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4433 option for pcretest, so I've replaced it by a simple function that does just
4434 that job.
4435
4436 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4437 "pcregrep -".
4438
4439 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4440 Java package. This provides some syntactic sugar for simple cases of what my
4441 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4442 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4443 item, you can use this simplified notation. Note that only makes sense with
4444 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4445 greediness, whatever the setting of the PCRE_UNGREEDY option.
4446
4447 13. A change of greediness default within a pattern was not taking effect at
4448 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4449 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4450 was abstracted outside.
4451
4452 14. PCRE now supports the \G assertion. It is true when the current matching
4453 position is at the start point of the match. This differs from \A when the
4454 starting offset is non-zero. Used with the /g option of pcretest (or similar
4455 code), it works in the same way as it does for Perl's /g option. If all
4456 alternatives of a regex begin with \G, the expression is anchored to the start
4457 match position, and the "anchored" flag is set in the compiled expression.
4458
4459 15. Some bugs concerning the handling of certain option changes within patterns
4460 have been fixed. These applied to options other than (?ims). For example,
4461 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4462 been the other way round. Some of this was related to change 7 above.
4463
4464 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4465 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4466 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4467 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4468
4469 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4470 mode. Unlike ".", it always matches newline, whatever the setting of
4471 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4472 assertions. Perl allows it, but it doesn't (in general) work because it can't
4473 calculate the length of the lookbehind. At least, that's the case for Perl
4474 5.8.0 - I've been told they are going to document that it doesn't work in
4475 future.
4476
4477 18. Added an error diagnosis for escapes that PCRE does not support: these are
4478 \L, \l, \N, \P, \p, \U, \u, and \X.
4479
4480 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4481 reading past the end of the pattern in cases such as /[abcd/.
4482
4483 20. PCRE was getting more memory than necessary for patterns with classes that
4484 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4485
4486 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4487 compiling PCRE for use with Virtual Pascal.
4488
4489 22. Small fix to the Makefile to make it work properly if the build is done
4490 outside the source tree.
4491
4492 23. Added a new extension: a condition to go with recursion. If a conditional
4493 subpattern starts with (?(R) the "true" branch is used if recursion has
4494 happened, whereas the "false" branch is used only at the top level.
4495
4496 24. When there was a very long string of literal characters (over 255 bytes
4497 without UTF support, over 250 bytes with UTF support), the computation of how
4498 much memory was required could be incorrect, leading to segfaults or other
4499 strange effects.
4500
4501 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4502 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4503 there was a subsequent back reference to those brackets. This meant that, for
4504 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4505 possible to check for precisely this case. All we can do is abandon the
4506 optimization if .* occurs inside capturing brackets when there are any back
4507 references whatsoever. (See below for a better fix that came later.)
4508
4509 26. The handling of the optimization for finding the first character of a
4510 non-anchored pattern, and for finding a character that is required later in the
4511 match were failing in some cases. This didn't break the matching; it just
4512 failed to optimize when it could. The way this is done has been re-implemented.
4513
4514 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4515
4516 28. Added a new feature that provides some of the functionality that Perl
4517 provides with (?{...}). The facility is termed a "callout". The way it is done
4518 in PCRE is for the caller to provide an optional function, by setting
4519 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4520 global variable. By default it is unset, which disables all calling out. To get
4521 the function called, the regex must include (?C) at appropriate points. This
4522 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4523 This provides a means of identifying different callout points. When PCRE
4524 reaches such a point in the regex, if pcre_callout has been set, the external
4525 function is called. It is provided with data in a structure called
4526 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4527 matching continues; if it returns a non-zero value, the match at the current
4528 point fails. However, backtracking will occur if possible. [This was changed
4529 later and other features added - see item 49 below.]
4530
4531 29. pcretest is upgraded to test the callout functionality. It provides a
4532 callout function that displays information. By default, it shows the start of
4533 the match and the current position in the text. There are some new data escapes
4534 to vary what happens:
4535
4536 \C+ in addition, show current contents of captured substrings
4537 \C- do not supply a callout function
4538 \C!n return 1 when callout number n is reached
4539 \C!n!m return 1 when callout number n is reached for the mth time
4540
4541 30. If pcregrep was called with the -l option and just a single file name, it
4542 output "<stdin>" if a match was found, instead of the file name.
4543
4544 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4545 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4546 pcre_exec(). This saves a malloc/free per call. The default value of
4547 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4548 when configuring.
4549
4550 32. The default maximum size of a compiled pattern is 64K. There have been a
4551 few cases of people hitting this limit. The code now uses macros to handle the
4552 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4553 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4554 configuring. Tests 2 and 5 work only with 2-byte links because they output
4555 debugging information about compiled patterns.
4556
4557 33. Internal code re-arrangements:
4558
4559 (a) Moved the debugging function for printing out a compiled regex into
4560 its own source file (printint.c) and used #include to pull it into
4561 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4562 separate copies.
4563
4564 (b) Defined the list of op-code names for debugging as a macro in
4565 internal.h so that it is next to the definition of the opcodes.
4566
4567 (c) Defined a table of op-code lengths for simpler skipping along compiled
4568 code. This is again a macro in internal.h so that it is next to the
4569 definition of the opcodes.
4570
4571 34. Added support for recursive calls to individual subpatterns, along the
4572 lines of Robin Houston's patch (but implemented somewhat differently).
4573
4574 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4575 allow it to read and process whole directories in Win32. This code was
4576 contributed by Lionel Fourquaux; it has not been tested by me.
4577
4578 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4579 used to name a group. Names consist of alphanumerics and underscores, and must
4580 be unique. Back references use the syntax (?P=name) and recursive calls use
4581 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4582 numbers. The function pcre_fullinfo() can be used after compilation to extract
4583 a name/number map. There are three relevant calls:
4584
4585 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4586 PCRE_INFO_NAMECOUNT yields the number of entries
4587 PCRE_INFO_NAMETABLE yields a pointer to the map.
4588
4589 The map is a vector of fixed-size entries. The size of each entry depends on
4590 the length of the longest name used. The first two bytes of each entry are the
4591 group number, most significant byte first. There follows the corresponding
4592 name, zero terminated. The names are in alphabetical order.
4593
4594 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4595 case instead of 255. Making it the same both with and without UTF-8 support
4596 means that the same test output works with both.
4597
4598 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4599 calling malloc() with a zero argument.
4600
4601 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4602 optimization. I've improved things by keeping a bitmap of backreferences with
4603 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4604 fact referenced, the optimization can be applied. It is unlikely that a
4605 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4606 the match to follow \n) will appear inside brackets with a number greater than
4607 31, but if it does, any back reference > 31 suppresses the optimization.
4608
4609 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4610 of disabling numbered capturing parentheses. Any opening parenthesis that is
4611 not followed by ? behaves as if it were followed by ?: but named parentheses
4612 can still be used for capturing (and they will acquire numbers in the usual
4613 way).
4614
4615 41. Redesigned the return codes from the match() function into yes/no/error so
4616 that errors can be passed back from deep inside the nested calls. A malloc
4617 failure while inside a recursive subpattern call now causes the
4618 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4619
4620 42. It is now possible to set a limit on the number of times the match()
4621 function is called in a call to pcre_exec(). This facility makes it possible to
4622 limit the amount of recursion and backtracking, though not in a directly
4623 obvious way, because the match() function is used in a number of different
4624 circumstances. The count starts from zero for each position in the subject
4625 string (for non-anchored patterns). The default limit is, for compatibility, a
4626 large number, namely 10 000 000. You can change this in two ways:
4627
4628 (a) When configuring PCRE before making, you can use --with-match-limit=n
4629 to set a default value for the compiled library.
4630
4631 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4632 a different value is set. See 45 below.
4633
4634 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4635
4636 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4637 of things that can be changed at compile time. The first argument specifies
4638 what is wanted and the second points to where the information is to be placed.
4639 The current list of available information is:
4640
4641 PCRE_CONFIG_UTF8
4642
4643 The output is an integer that is set to one if UTF-8 support is available;
4644 otherwise it is set to zero.
4645
4646 PCRE_CONFIG_NEWLINE
4647
4648 The output is an integer that it set to the value of the code that is used for
4649 newline. It is either LF (10) or CR (13).
4650
4651 PCRE_CONFIG_LINK_SIZE
4652
4653 The output is an integer that contains the number of bytes used for internal
4654 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4655
4656 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4657
4658 The output is an integer that contains the threshold above which the POSIX
4659 interface uses malloc() for output vectors. See item 31 above.
4660
4661 PCRE_CONFIG_MATCH_LIMIT
4662
4663 The output is an unsigned integer that contains the default limit of the number
4664 of match() calls in a pcre_exec() execution. See 42 above.
4665
4666 44. pcretest has been upgraded by the addition of the -C option. This causes it
4667 to extract all the available output from the new pcre_config() function, and to
4668 output it. The program then exits immediately.
4669
4670 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4671 order to support additional features. One way would have been to define
4672 pcre_exec2() (for example) with extra arguments, but this would not have been
4673 extensible, and would also have required all calls to the original function to
4674 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4675 is used for passing in "extra" data from pcre_study().
4676
4677 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4678 contains the following fields:
4679
4680 flags a bitmap indicating which of the following fields are set
4681 study_data opaque data from pcre_study()
4682 match_limit a way of specifying a limit on match() calls for a specific
4683 call to pcre_exec()
4684 callout_data data for callouts (see 49 below)
4685
4686 The flag bits are also defined in pcre.h, and are
4687
4688 PCRE_EXTRA_STUDY_DATA
4689 PCRE_EXTRA_MATCH_LIMIT
4690 PCRE_EXTRA_CALLOUT_DATA
4691
4692 The pcre_study() function now returns one of these new pcre_extra blocks, with
4693 the actual study data pointed to by the study_data field, and the
4694 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4695 before. That is, this change is entirely upwards-compatible and requires no
4696 change to existing code.
4697
4698 If you want to pass in additional data to pcre_exec(), you can either place it
4699 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4700 block.
4701
4702 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4703 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4704 times with different match limits, until it finds the minimum value needed for
4705 pcre_exec() to complete. The value is then output. This can be instructive; for
4706 most simple matches the number is quite small, but for pathological cases it
4707 gets very large very quickly.
4708
4709 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4710 returns the size of the data block pointed to by the study_data field in a
4711 pcre_extra block, that is, the value that was passed as the argument to
4712 pcre_malloc() when PCRE was getting memory in which to place the information
4713 created by pcre_study(). The fourth argument should point to a size_t variable.
4714 pcretest has been extended so that this information is shown after a successful
4715 pcre_study() call when information about the compiled regex is being displayed.
4716
4717 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4718 because what follows is always an absolute path. (Later: it turns out that this
4719 is more than cosmetic for MinGW, because it doesn't like empty path
4720 components.)
4721
4722 49. Some changes have been made to the callout feature (see 28 above):
4723
4724 (i) A callout function now has three choices for what it returns:
4725
4726 0 => success, carry on matching
4727 > 0 => failure at this point, but backtrack if possible
4728 < 0 => serious error, return this value from pcre_exec()
4729
4730 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4731 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4732 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4733 use by callout functions. It will never be used by PCRE itself.
4734
4735 (ii) The pcre_extra structure (see 45 above) has a void * field called
4736 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4737 pcre_callout_block structure has a field of the same name. The contents of
4738 the field passed in the pcre_extra structure are passed to the callout
4739 function in the corresponding field in the callout block. This makes it
4740 easier to use the same callout-containing regex from multiple threads. For
4741 testing, the pcretest program has a new data escape
4742
4743 \C*n pass the number n (may be negative) as callout_data
4744
4745 If the callout function in pcretest receives a non-zero value as
4746 callout_data, it returns that value.
4747
4748 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4749 there were some redundant $(CFLAGS) in commands that are now specified as
4750 $(LINK), which already includes $(CFLAGS).
4751
4752 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4753 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4754 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4755 one-byte characters throughout. Note that case-insensitive matching applies
4756 only to characters whose values are less than 256. PCRE doesn't support the
4757 notion of cases for higher-valued characters.
4758
4759 (i) A character class whose characters are all within 0-255 is handled as
4760 a bit map, and the map is inverted for negative classes. Previously, a
4761 character > 255 always failed to match such a class; however it should
4762 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4763
4764 (ii) A negated character class with a single character < 255 is coded as
4765 "not this character" (OP_NOT). This wasn't working properly when the test
4766 character was multibyte, either singly or repeated.
4767
4768 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4769 mode, for example: \x{100}{2,3}.
4770
4771 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4772 singly or repeated) now correctly test multibyte characters. However,
4773 PCRE doesn't recognize any characters with values greater than 255 as
4774 digits, spaces, or word characters. Such characters always match \D, \S,
4775 and \W, and never match \d, \s, or \w.
4776
4777 (v) Classes may now contain characters and character ranges with values
4778 greater than 255. For example: [ab\x{100}-\x{400}].
4779
4780 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4781 PCRE in UTF-8 mode.
4782
4783 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4784 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4785 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4786 value.)
4787
4788 53. The single man page has become too large. I have therefore split it up into
4789 a number of separate man pages. These also give rise to individual HTML pages;
4790 these are now put in a separate directory, and there is an index.html page that
4791 lists them all. Some hyperlinking between the pages has been installed.
4792
4793 54. Added convenience functions for handling named capturing parentheses.
4794
4795 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4796 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4797 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4798 are faulted.
4799
4800 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4801 calling configure. These values are used when compiling the dftables.c program
4802 which is run to generate the source of the default character tables. They
4803 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4804 you will need to set these values.
4805
4806 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4807
4808
4809 Version 3.9 02-Jan-02
4810 ---------------------
4811
4812 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4813
4814 2. If --disable-static was given, the building process failed when trying to
4815 build pcretest and pcregrep. (For some reason it was using libtool to compile
4816 them, which is not right, as they aren't part of the library.)
4817
4818
4819 Version 3.8 18-Dec-01
4820 ---------------------
4821
4822 1. The experimental UTF-8 code was completely screwed up. It was packing the
4823 bytes in the wrong order. How dumb can you get?
4824
4825
4826 Version 3.7 29-Oct-01
4827 ---------------------
4828
4829 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4830 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4831 this didn't happen under Solaris 8, where I normally test things.
4832
4833 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4834 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4835 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4836 this fix an hour or so after the initial 3.7 release.)
4837
4838
4839 Version 3.6 23-Oct-01
4840 ---------------------
4841
4842 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4843 offsets passed as NULL with zero offset count.
4844
4845 2. The config.guess and config.sub files had not been updated when I moved to
4846 the latest autoconf.
4847
4848
4849 Version 3.5 15-Aug-01
4850 ---------------------
4851
4852 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4853 had been forgotten.
4854
4855 2. By using declared but undefined structures, we can avoid using "void"
4856 definitions in pcre.h while keeping the internal definitions of the structures
4857 private.
4858
4859 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4860 user point of view, this means that both static and shared libraries are built
4861 by default, but this can be individually controlled. More of the work of
4862 handling this static/shared cases is now inside libtool instead of PCRE's make
4863 file.
4864
4865 4. The pcretest utility is now installed along with pcregrep because it is
4866 useful for users (to test regexs) and by doing this, it automatically gets
4867 relinked by libtool. The documentation has been turned into a man page, so
4868 there are now .1, .txt, and .html versions in /doc.
4869
4870 5. Upgrades to pcregrep:
4871 (i) Added long-form option names like gnu grep.
4872 (ii) Added --help to list all options with an explanatory phrase.
4873 (iii) Added -r, --recursive to recurse into sub-directories.
4874 (iv) Added -f, --file to read patterns from a file.
4875
4876 6. pcre_exec() was referring to its "code" argument before testing that
4877 argument for NULL (and giving an error if it was NULL).
4878
4879 7. Upgraded Makefile.in to allow for compiling in a different directory from
4880 the source directory.
4881
4882 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4883 options bits, the pointer it was passed was to an int instead of to an unsigned
4884 long int. This mattered only on 64-bit systems.
4885
4886 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4887 generated) instead of pcre.in, which it its source. Also made the same change
4888 in several of the .c files.
4889
4890 10. A new release of gcc defines printf() as a macro, which broke pcretest
4891 because it had an ifdef in the middle of a string argument for printf(). Fixed
4892 by using separate calls to printf().
4893
4894 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4895 script, to force use of CR or LF instead of \n in the source. On non-Unix
4896 systems, the value can be set in config.h.
4897
4898 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4899 absolute limit. Changed the text of the error message to make this clear, and
4900 likewise updated the man page.
4901
4902 13. The limit of 99 on the number of capturing subpatterns has been removed.
4903 The new limit is 65535, which I hope will not be a "real" limit.
4904
4905
4906 Version 3.4 22-Aug-00
4907 ---------------------
4908
4909 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4910
4911 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4912
4913
4914 Version 3.3 01-Aug-00
4915 ---------------------
4916
4917 1. If an octal character was given, but the value was greater than \377, it
4918 was not getting masked to the least significant bits, as documented. This could
4919 lead to crashes in some systems.
4920
4921 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4922 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4923
4924 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4925 These just pass their arguments on to (pcre_free)(), but they are provided
4926 because some uses of PCRE bind it to non-C systems that can call its functions,
4927 but cannot call free() or pcre_free() directly.
4928
4929 4. Add "make test" as a synonym for "make check". Corrected some comments in
4930 the Makefile.
4931
4932 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4933 Makefile.
4934
4935 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4936 command called pgrep for grepping around the active processes.
4937
4938 7. Added the beginnings of support for UTF-8 character strings.
4939
4940 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4941 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4942 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4943 out for the ar command.)
4944
4945
4946 Version 3.2 12-May-00
4947 ---------------------
4948
4949 This is purely a bug fixing release.
4950
4951 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4952 of ZA. This was just one example of several cases that could provoke this bug,
4953 which was introduced by change 9 of version 2.00. The code for breaking
4954 infinite loops after an iteration that matches an empty string was't working
4955 correctly.
4956
4957 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4958 when matched against abbab (for example). After matching an empty string, it
4959 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4960 caused it to match further down the string than it should.
4961
4962 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4963 was there because it doesn't seem to be needed, and it causes trouble on some
4964 systems, as it is not a Standard C header. It has been removed.
4965
4966 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4967 were reported on the Macintosh. The changes were from
4968
4969 while ((c = *(++ptr)) != 0 && c != '\n');
4970 to
4971 while ((c = *(++ptr)) != 0 && c != '\n') ;
4972
4973 Totally extraordinary, but if that's what it takes...
4974
4975 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4976 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4977 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4978 assumes the way PCRE uses memmove() (always moving upwards).
4979
4980 6. PCRE is being used in one environment where strchr() is not available. There
4981 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4982 faster code anyway.
4983
4984
4985 Version 3.1 09-Feb-00
4986 ---------------------
4987
4988 The only change in this release is the fixing of some bugs in Makefile.in for
4989 the "install" target:
4990
4991 (1) It was failing to install pcreposix.h.
4992
4993 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4994
4995
4996 Version 3.0 01-Feb-00
4997 ---------------------
4998
4999 1. Add support for the /+ modifier to perltest (to output $` like it does in
5000 pcretest).
5001
5002 2. Add support for the /g modifier to perltest.
5003
5004 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5005 matches null strings.
5006
5007 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5008 pattern. Perl treats empty patterns specially - it reuses the most recent
5009 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5010 effect.
5011
5012 5. The POSIX interface was broken in that it was just handing over the POSIX
5013 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5014 required a bigger vector, with some working space on the end. This means that
5015 the POSIX wrapper now has to get and free some memory, and copy the results.
5016
5017 6. Added some simple autoconf support, placing the test data and the
5018 documentation in separate directories, re-organizing some of the
5019 information files, and making it build pcre-config (a GNU standard). Also added
5020 libtool support for building PCRE as a shared library, which is now the
5021 default.
5022
5023 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5024 09 are not valid octal constants. Single digits will be used for minor values
5025 less than 10.
5026
5027 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5028 existing programs that set these in the POSIX interface can use PCRE without
5029 modification.
5030
5031 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5032 return all that pcre_info() returns, plus additional data. The pcre_info()
5033 function is retained for compatibility, but is considered to be obsolete.
5034
5035 10. Added experimental recursion feature (?R) to handle one common case that
5036 Perl 5.6 will be able to do with (?p{...}).
5037
5038 11. Added support for POSIX character classes like [:alpha:], which Perl is
5039 adopting.
5040
5041
5042 Version 2.08 31-Aug-99
5043 ----------------------
5044
5045 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5046 trying to match at the startoffset position, but instead was moving forward to
5047 the next newline as if a previous match had failed.
5048
5049 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5050 and could get into a loop if a null string was matched other than at the start
5051 of the subject.
5052
5053 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5054 be distinguished at compile time, and for completeness also added PCRE_DATE.
5055
5056 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5057 in GnuWin32 environments.
5058
5059
5060 Version 2.07 29-Jul-99
5061 ----------------------
5062
5063 1. The documentation is now supplied in plain text form and HTML as well as in
5064 the form of man page sources.
5065
5066 2. C++ compilers don't like assigning (void *) values to other pointer types.
5067 In particular this affects malloc(). Although there is no problem in Standard
5068 C, I've put in casts to keep C++ compilers happy.
5069
5070 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5071 should be (const char *).
5072
5073 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5074 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5075 However, I haven't made this a standard facility. The documentation doesn't
5076 mention it, and the Makefile doesn't support it.
5077
5078 5. The Makefile now contains an "install" target, with editable destinations at
5079 the top of the file. The pcretest program is not installed.
5080
5081 6. pgrep -V now gives the PCRE version number and date.
5082
5083 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5084 causing the entire string to be ignored, instead of just the last character.
5085
5086 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5087 non-matching string, it can take a very, very long time, even for strings of
5088 quite modest length, because of the nested recursion. PCRE now does better in
5089 some of these cases. It does this by remembering the last required literal
5090 character in the pattern, and pre-searching the subject to ensure it is present
5091 before running the real match. In other words, it applies a heuristic to detect
5092 some types of certain failure quickly, and in the above example, if presented
5093 with a string that has no trailing " it gives "no match" very quickly.
5094
5095 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5096 other alternatives are tried instead.
5097
5098
5099 Version 2.06 09-Jun-99
5100 ----------------------
5101
5102 1. Change pcretest's output for amount of store used to show just the code
5103 space, because the remainder (the data block) varies in size between 32-bit and
5104 64-bit systems.
5105
5106 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5107 start matching at. This allows lookbehinds to work when searching for multiple
5108 occurrences in a string.
5109
5110 3. Added additional options to pcretest for testing multiple occurrences:
5111
5112 /+ outputs the rest of the string that follows a match
5113 /g loops for multiple occurrences, using the new startoffset argument
5114 /G loops for multiple occurrences by passing an incremented pointer
5115
5116 4. PCRE wasn't doing the "first character" optimization for patterns starting
5117 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5118 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5119 the letter 'x'. On long subject strings, this gives a significant speed-up.
5120
5121
5122 Version 2.05 21-Apr-99
5123 ----------------------
5124
5125 1. Changed the type of magic_number from int to long int so that it works
5126 properly on 16-bit systems.
5127
5128 2. Fixed a bug which caused patterns starting with .* not to work correctly
5129 when the subject string contained newline characters. PCRE was assuming
5130 anchoring for such patterns in all cases, which is not correct because .* will
5131 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5132 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5133 must be retried after every newline in the subject.
5134
5135
5136 Version 2.04 18-Feb-99
5137 ----------------------
5138
5139 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5140 computation of the store needed to hold the pattern was incorrect (too large).
5141 If such patterns were nested a few deep, this could multiply and become a real
5142 problem.
5143
5144 2. Added /M option to pcretest to show the memory requirement of a specific
5145 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5146
5147 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5148 compiled in such a way that the backtracking after subsequent failure was
5149 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5150 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5151
5152
5153 Version 2.03 02-Feb-99
5154 ----------------------
5155
5156 1. Fixed typo and small mistake in man page.
5157
5158 2. Added 4th condition (GPL supersedes if conflict) and created separate
5159 LICENCE file containing the conditions.
5160
5161 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5162 Perl, that is the internal \ allows the delimiter to be included in the
5163 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5164 the final delimiter, add \ to the end of the pattern (to test the error).
5165
5166 4. Added the convenience functions for extracting substrings after a successful
5167 match. Updated pcretest to make it able to test these functions.
5168
5169
5170 Version 2.02 14-Jan-99
5171 ----------------------
5172
5173 1. Initialized the working variables associated with each extraction so that
5174 their saving and restoring doesn't refer to uninitialized store.
5175
5176 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5177 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5178 fix the problem.
5179
5180 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5181 calls, and wasn't printing the correct value for compiling calls. Increased the
5182 default value of LOOPREPEAT, and the number of significant figures in the
5183 times.
5184
5185 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5186
5187 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5188 a building problem on Windows NT with a FAT file system.
5189
5190
5191 Version 2.01 21-Oct-98
5192 ----------------------
5193
5194 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5195 to character tables built by pcre_maketables() in the current locale. If NULL
5196 is passed, the default tables are used.
5197
5198
5199 Version 2.00 24-Sep-98
5200 ----------------------
5201
5202 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5203 it any more.
5204
5205 2. Allow quantification of (?>) groups, and make it work correctly.
5206
5207 3. The first character computation wasn't working for (?>) groups.
5208
5209 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5210 end of the subject) and add 5.005's \z, which really does match only at the
5211 very end of the subject.
5212
5213 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5214
5215 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5216 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5217 localized options. All options to pcre_study() were also removed.
5218
5219 7. Add other new features from 5.005:
5220
5221 $(?<= positive lookbehind
5222 $(?<! negative lookbehind
5223 (?imsx-imsx) added the unsetting capability
5224 such a setting is global if at outer level; local otherwise
5225 (?imsx-imsx:) non-capturing groups with option setting
5226 (?(cond)re|re) conditional pattern matching
5227
5228 A backreference to itself in a repeated group matches the previous
5229 captured string.
5230
5231 8. General tidying up of studying (both automatic and via "study")
5232 consequential on the addition of new assertions.
5233
5234 9. As in 5.005, unlimited repeated groups that could match an empty substring
5235 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5236 runtime if any iteration does actually match an empty substring.
5237
5238 10. Include the RunTest script in the distribution.
5239
5240 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5241 discrepancies, some of which were old and were also with respect to 5.004. They
5242 have now been fixed.
5243
5244
5245 Version 1.09 28-Apr-98
5246 ----------------------
5247
5248 1. A negated single character class followed by a quantifier with a minimum
5249 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5250 program crashes, or just wrong answers. This did not apply to negated classes
5251 containing more than one character, or to minima other than one.
5252
5253
5254 Version 1.08 27-Mar-98
5255 ----------------------
5256
5257 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5258
5259 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5260 latter must appear before anything that relies on it in the pattern.
5261
5262
5263 Version 1.07 16-Feb-98
5264 ----------------------
5265
5266 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5267 repeat of a potentially empty string).
5268
5269
5270 Version 1.06 23-Jan-98
5271 ----------------------
5272
5273 1. Added Markus Oberhumer's little patches for C++.
5274
5275 2. Literal strings longer than 255 characters were broken.
5276
5277
5278 Version 1.05 23-Dec-97
5279 ----------------------
5280
5281 1. Negated character classes containing more than one character were failing if
5282 PCRE_CASELESS was set at run time.
5283
5284
5285 Version 1.04 19-Dec-97
5286 ----------------------
5287
5288 1. Corrected the man page, where some "const" qualifiers had been omitted.
5289
5290 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5291 input syntax.
5292
5293 3. Fixed memory leak which occurred when a regex with back references was
5294 matched with an offsets vector that wasn't big enough. The temporary memory
5295 that is used in this case wasn't being freed if the match failed.
5296
5297 4. Tidied pcretest to ensure it frees memory that it gets.
5298
5299 5. Temporary memory was being obtained in the case where the passed offsets
5300 vector was exactly big enough.
5301
5302 6. Corrected definition of offsetof() from change 5 below.
5303
5304 7. I had screwed up change 6 below and broken the rules for the use of
5305 setjmp(). Now fixed.
5306
5307
5308 Version 1.03 18-Dec-97
5309 ----------------------
5310
5311 1. A erroneous regex with a missing opening parenthesis was correctly
5312 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5313 on some systems.
5314
5315 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5316 it was reported that one broken compiler failed on the former because "code" is
5317 also an independent variable.
5318
5319 3. The erroneous regex a[]b caused an array overrun reference.
5320
5321 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5322 fail on data ending with that character. (It was going on too far, and checking
5323 the next character, typically a binary zero.) This was specific to the
5324 optimized code for single-character negative classes.
5325
5326 5. Added a contributed patch from the TIN world which does the following:
5327
5328 + Add an undef for memmove, in case the the system defines a macro for it.
5329
5330 + Add a definition of offsetof(), in case there isn't one. (I don't know
5331 the reason behind this - offsetof() is part of the ANSI standard - but
5332 it does no harm).
5333
5334 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5335 most of the places where whitespace preceded '#'. I have given up and
5336 allowed the remaining 2 cases to be at the margin.
5337
5338 + Rename some variables in pcre to eliminate shadowing. This seems very
5339 pedantic, but does no harm, of course.
5340
5341 6. Moved the call to setjmp() into its own function, to get rid of warnings
5342 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5343
5344 7. Constructs such as \d{8,} were compiling into the equivalent of
5345 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5346 outcome, but in this particular case used more store than had been allocated,
5347 which caused the bug to be discovered because it threw up an internal error.
5348
5349 8. The debugging code in both pcre and pcretest for outputting the compiled
5350 form of a regex was going wrong in the case of back references followed by
5351 curly-bracketed repeats.
5352
5353
5354 Version 1.02 12-Dec-97
5355 ----------------------
5356
5357 1. Typos in pcre.3 and comments in the source fixed.
5358
5359 2. Applied a contributed patch to get rid of places where it used to remove
5360 'const' from variables, and fixed some signed/unsigned and uninitialized
5361 variable warnings.
5362
5363 3. Added the "runtest" target to Makefile.
5364
5365 4. Set default compiler flag to -O2 rather than just -O.
5366
5367
5368 Version 1.01 19-Nov-97
5369 ----------------------
5370
5371 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5372 like /([ab]*)*/, that is, for classes with more than one character in them.
5373
5374 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5375 as /((?>a*))*/ (a PCRE_EXTRA facility).
5376
5377
5378 Version 1.00 18-Nov-97
5379 ----------------------
5380
5381 1. Added compile-time macros to support systems such as SunOS4 which don't have
5382 memmove() or strerror() but have other things that can be used instead.
5383
5384 2. Arranged that "make clean" removes the executables.
5385
5386
5387 Version 0.99 27-Oct-97
5388 ----------------------
5389
5390 1. Fixed bug in code for optimizing classes with only one character. It was
5391 initializing a 32-byte map regardless, which could cause it to run off the end
5392 of the memory it had got.
5393
5394 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5395
5396
5397 Version 0.98 22-Oct-97
5398 ----------------------
5399
5400 1. Fixed bug in code for handling temporary memory usage when there are more
5401 back references than supplied space in the ovector. This could cause segfaults.
5402
5403
5404 Version 0.97 21-Oct-97
5405 ----------------------
5406
5407 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5408
5409 2. Optimized negated single characters not to use a bit map.
5410
5411 3. Brought error texts together as macro definitions; clarified some of them;
5412 fixed one that was wrong - it said "range out of order" when it meant "invalid
5413 escape sequence".
5414
5415 4. Changed some char * arguments to const char *.
5416
5417 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5418
5419 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5420 pcretest.
5421
5422
5423 Version 0.96 16-Oct-97
5424 ----------------------
5425
5426 1. Added a simple "pgrep" utility to the distribution.
5427
5428 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5429 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5430 where "ddd" means "one or more decimal digits".
5431
5432 3. Fixed serious bug. If a pattern had a back reference, but the call to
5433 pcre_exec() didn't supply a large enough ovector to record the related
5434 identifying subpattern, the match always failed. PCRE now remembers the number
5435 of the largest back reference, and gets some temporary memory in which to save
5436 the offsets during matching if necessary, in order to ensure that
5437 backreferences always work.
5438
5439 4. Increased the compatibility with Perl in a number of ways:
5440
5441 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5442 to request this handling. The option can be set at compile or exec time.
5443
5444 (b) $ matches before a terminating newline by default; an option
5445 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5446 mode). The option can be set at compile or exec time.
5447
5448 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5449 the same as Perl's. If the decimal number it represents is less than 10
5450 or there aren't that many previous left capturing parentheses, an octal
5451 escape is read. Inside a character class, it's always an octal escape,
5452 even if it is a single digit.
5453
5454 (d) An escaped but undefined alphabetic character is taken as a literal,
5455 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5456 escapes.
5457
5458 (e) {0} is now permitted. (The previous item is removed from the compiled
5459 pattern).
5460
5461 5. Changed all the names of code files so that the basic parts are no longer
5462 than 10 characters, and abolished the teeny "globals.c" file.
5463
5464 6. Changed the handling of character classes; they are now done with a 32-byte
5465 bit map always.
5466
5467 7. Added the -d and /D options to pcretest to make it possible to look at the
5468 internals of compilation without having to recompile pcre.
5469
5470
5471 Version 0.95 23-Sep-97
5472 ----------------------
5473
5474 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5475 \x20 at the start of a run of normal characters. These were being treated as
5476 real characters, instead of the source characters being re-checked.
5477
5478
5479 Version 0.94 18-Sep-97
5480 ----------------------
5481
5482 1. The functions are now thread-safe, with the caveat that the global variables
5483 containing pointers to malloc() and free() or alternative functions are the
5484 same for all threads.
5485
5486 2. Get pcre_study() to generate a bitmap of initial characters for non-
5487 anchored patterns when this is possible, and use it if passed to pcre_exec().
5488
5489
5490 Version 0.93 15-Sep-97
5491 ----------------------
5492
5493 1. /(b)|(:+)/ was computing an incorrect first character.
5494
5495 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5496 but not actually doing anything yet.
5497
5498 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5499 as Perl does (e.g. [-az] or [az-]).
5500
5501 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5502 all possible positions.
5503
5504 5. Split up into different modules to avoid including unneeded functions in a
5505 compiled binary. However, compile and exec are still in one module. The "study"
5506 function is split off.
5507
5508 6. The character tables are now in a separate module whose source is generated
5509 by an auxiliary program - but can then be edited by hand if required. There are
5510 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5511 toupper() in the code.
5512
5513 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5514 make them global. Abolish the function for setting them, as the caller can now
5515 set them directly.
5516
5517
5518 Version 0.92 11-Sep-97
5519 ----------------------
5520
5521 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5522 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5523
5524 2. Caseless matching was not working in character classes if the characters in
5525 the pattern were in upper case.
5526
5527 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5528
5529 4. Make PCRE_ANCHORED public and accept as a compile option.
5530
5531 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5532 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5533 pass them.
5534
5535 6. Give an error if bad option bits passed at compile or run time.
5536
5537 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5538 pcretest to cause it to pass that flag.
5539
5540 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5541 options, and the first character, if set.
5542
5543 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5544
5545
5546 Version 0.91 10-Sep-97
5547 ----------------------
5548
5549 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5550 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5551
5552 2. PCRE was looping on encountering an indefinitely repeated back reference to
5553 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5554 Perl does - treats the match as successful.
5555
5556 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5