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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1482 - (show annotations)
Wed May 28 11:10:58 2014 UTC (5 years, 5 months ago) by ph10
File size: 266760 byte(s)
Error occurred while calculating annotation data.
Fix two bugs concerned with duplicate named patterns.
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
65 Version 8.35 04-April-2014
66 --------------------------
67
68 1. A new flag is set, when property checks are present in an XCLASS.
69 When this flag is not set, PCRE can perform certain optimizations
70 such as studying these XCLASS-es.
71
72 2. The auto-possessification of character sets were improved: a normal
73 and an extended character set can be compared now. Furthermore
74 the JIT compiler optimizes more character set checks.
75
76 3. Got rid of some compiler warnings for potentially uninitialized variables
77 that show up only when compiled with -O2.
78
79 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
80 of a match later then the end of the match. The pcretest program was not
81 handling the case sensibly - it was outputting from the start to the next
82 binary zero. It now reports this situation in a message, and outputs the
83 text from the end to the start.
84
85 5. Fast forward search is improved in JIT. Instead of the first three
86 characters, any three characters with fixed position can be searched.
87 Search order: first, last, middle.
88
89 6. Improve character range checks in JIT. Characters are read by an inprecise
90 function now, which returns with an unknown value if the character code is
91 above a certain treshold (e.g: 256). The only limitation is that the value
92 must be bigger than the treshold as well. This function is useful, when
93 the characters above the treshold are handled in the same way.
94
95 7. The macros whose names start with RAWUCHAR are placeholders for a future
96 mode in which only the bottom 21 bits of 32-bit data items are used. To
97 make this more memorable for those maintaining the code, the names have
98 been changed to start with UCHAR21, and an extensive comment has been added
99 to their definition.
100
101 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
102 to the export list in Makefile.am (they were accidentally omitted from the
103 8.34 tarball).
104
105 9. The informational output from pcretest used the phrase "starting byte set"
106 which is inappropriate for the 16-bit and 32-bit libraries. As the output
107 for "first char" and "need char" really means "non-UTF-char", I've changed
108 "byte" to "char", and slightly reworded the output. The documentation about
109 these values has also been (I hope) clarified.
110
111 10. Another JIT related optimization: use table jumps for selecting the correct
112 backtracking path, when more than four alternatives are present inside a
113 bracket.
114
115 11. Empty match is not possible, when the minimum length is greater than zero,
116 and there is no \K in the pattern. JIT should avoid empty match checks in
117 such cases.
118
119 12. In a caseless character class with UCP support, when a character with more
120 than one alternative case was not the first character of a range, not all
121 the alternative cases were added to the class. For example, s and \x{17f}
122 are both alternative cases for S: the class [RST] was handled correctly,
123 but [R-T] was not.
124
125 13. The configure.ac file always checked for pthread support when JIT was
126 enabled. This is not used in Windows, so I have put this test inside a
127 check for the presence of windows.h (which was already tested for).
128
129 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
130 The algorithm provides a way to skip certain starting offsets, and usually
131 faster than linear prefix searches.
132
133 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
134 as for 'fr_FR' and 'french'. For some reason, however, it then used the
135 Windows-specific input and output files, which have 'french' screwed in.
136 So this could never have worked. One of the problems with locales is that
137 they aren't always the same. I have now updated RunTest so that it checks
138 the output of the locale test (test 3) against three different output
139 files, and it allows the test to pass if any one of them matches. With luck
140 this should make the test pass on some versions of Solaris where it was
141 failing. Because of the uncertainty, the script did not used to stop if
142 test 3 failed; it now does. If further versions of a French locale ever
143 come to light, they can now easily be added.
144
145 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
146 there was a message during ./configure, but it did not stop. This now
147 provokes an error. The invalid example in README has been corrected.
148 If a value less than the minimum is given, the minimum value has always
149 been used, but now a warning is given.
150
151 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
152 was a bug in the test system, which is now fixed. Also, the list of various
153 configurations that are tested for each release did not have one with both
154 16/32 bits and --enable-bar-anycrlf. It now does.
155
156 18. pcretest was missing "-C bsr" for displaying the \R default setting.
157
158 19. Little endian PowerPC systems are supported now by the JIT compiler.
159
160 20. The fast forward newline mechanism could enter to an infinite loop on
161 certain invalid UTF-8 input. Although we don't support these cases
162 this issue can be fixed by a performance optimization.
163
164 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
165 not take account if existing stack usage. There is now a new global
166 variable called pcre_stack_guard that can be set to point to an external
167 function to check stack availability. It is called at the start of
168 processing every parenthesized group.
169
170 22. A typo in the code meant that in ungreedy mode the max/min qualifier
171 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
172 not match "ab".
173
174 23. When UTF was disabled, the JIT program reported some incorrect compile
175 errors. These messages are silenced now.
176
177 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
178 compiler.
179
180 25. Change all the temporary files used in RunGrepTest to be different to those
181 used by RunTest so that the tests can be run simultaneously, for example by
182 "make -j check".
183
184
185 Version 8.34 15-December-2013
186 -----------------------------
187
188 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
189 executable memory. Patch inspired by Carsten Klein.
190
191 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
192 this macro is never tested and has no effect, because the work to support
193 coverage involves only compiling and linking options and special targets in
194 the Makefile. The comment in config.h implied that defining the macro would
195 enable coverage support, which is totally false. There was also support for
196 setting this macro in the CMake files (my fault, I just copied it from
197 configure). SUPPORT_GCOV has now been removed.
198
199 3. Make a small performance improvement in strlen16() and strlen32() in
200 pcretest.
201
202 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
203 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
204 these cases). There was also one in pcretest.c.
205
206 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
207
208 6. In UTF mode, the code for checking whether a group could match an empty
209 string (which is used for indefinitely repeated groups to allow for
210 breaking an infinite loop) was broken when the group contained a repeated
211 negated single-character class with a character that occupied more than one
212 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
213 UTF-8 mode). The effect was undefined: the group might or might not be
214 deemed as matching an empty string, or the program might have crashed.
215
216 7. The code for checking whether a group could match an empty string was not
217 recognizing that \h, \H, \v, \V, and \R must match a character.
218
219 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
220 an empty string. If it can, pcretest shows this in its information output.
221
222 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
223 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
224 matched by pcre_exec() without using JIT:
225
226 (a) If the rest of the pattern did not match after a maximal run of
227 grapheme clusters, the code for backing up to try with fewer of them
228 did not always back up over a full grapheme when characters that do not
229 have the modifier quality were involved, e.g. Hangul syllables.
230
231 (b) If the match point in a subject started with modifier character, and
232 there was no match, the code could incorrectly back up beyond the match
233 point, and potentially beyond the first character in the subject,
234 leading to a segfault or an incorrect match result.
235
236 10. A conditional group with an assertion condition could lead to PCRE
237 recording an incorrect first data item for a match if no other first data
238 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
239 first data item, and therefore matched "ca" after "c" instead of at the
240 start.
241
242 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
243 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
244
245 12. The source of pcregrep now includes z/OS-specific code so that it can be
246 compiled for z/OS as part of the special z/OS distribution.
247
248 13. Added the -T and -TM options to pcretest.
249
250 14. The code in pcre_compile.c for creating the table of named capturing groups
251 has been refactored. Instead of creating the table dynamically during the
252 actual compiling pass, the information is remembered during the pre-compile
253 pass (on the stack unless there are more than 20 named groups, in which
254 case malloc() is used) and the whole table is created before the actual
255 compile happens. This has simplified the code (it is now nearly 150 lines
256 shorter) and prepared the way for better handling of references to groups
257 with duplicate names.
258
259 15. A back reference to a named subpattern when there is more than one of the
260 same name now checks them in the order in which they appear in the pattern.
261 The first one that is set is used for the reference. Previously only the
262 first one was inspected. This change makes PCRE more compatible with Perl.
263
264 16. Unicode character properties were updated from Unicode 6.3.0.
265
266 17. The compile-time code for auto-possessification has been refactored, based
267 on a patch by Zoltan Herczeg. It now happens after instead of during
268 compilation. The code is cleaner, and more cases are handled. The option
269 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
270 options in pcretest are provided to set it. It can also be set by
271 (*NO_AUTO_POSSESS) at the start of a pattern.
272
273 18. The character VT has been added to the default ("C" locale) set of
274 characters that match \s and are generally treated as white space,
275 following this same change in Perl 5.18. There is now no difference between
276 "Perl space" and "POSIX space". Whether VT is treated as white space in
277 other locales depends on the locale.
278
279 19. The code for checking named groups as conditions, either for being set or
280 for being recursed, has been refactored (this is related to 14 and 15
281 above). Processing unduplicated named groups should now be as fast at
282 numerical groups, and processing duplicated groups should be faster than
283 before.
284
285 20. Two patches to the CMake build system, by Alexander Barkov:
286
287 (1) Replace the "source" command by "." in CMakeLists.txt because
288 "source" is a bash-ism.
289
290 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
291 without these the CMake build does not work on Solaris.
292
293 21. Perl has changed its handling of \8 and \9. If there is no previously
294 encountered capturing group of those numbers, they are treated as the
295 literal characters 8 and 9 instead of a binary zero followed by the
296 literals. PCRE now does the same.
297
298 22. Following Perl, added \o{} to specify codepoints in octal, making it
299 possible to specify values greater than 0777 and also making them
300 unambiguous.
301
302 23. Perl now gives an error for missing closing braces after \x{... instead of
303 treating the string as literal. PCRE now does the same.
304
305 24. RunTest used to grumble if an inappropriate test was selected explicitly,
306 but just skip it when running all tests. This make it awkward to run ranges
307 of tests when one of them was inappropriate. Now it just skips any
308 inappropriate tests, as it always did when running all tests.
309
310 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
311 character types such as \d or \w, too many callouts were inserted, and the
312 data that they returned was rubbish.
313
314 26. In UCP mode, \s was not matching two of the characters that Perl matches,
315 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
316 were matched by \h. The code has now been refactored so that the lists of
317 the horizontal and vertical whitespace characters used for \h and \v (which
318 are defined only in one place) are now also used for \s.
319
320 27. Add JIT support for the 64 bit TileGX architecture.
321 Patch by Jiong Wang (Tilera Corporation).
322
323 28. Possessive quantifiers for classes (both explicit and automatically
324 generated) now use special opcodes instead of wrapping in ONCE brackets.
325
326 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
327 (because it's meaningless), this was not happening when PCRE_CASELESS was
328 set. Not wrong, but inefficient.
329
330 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
331 properties for \w, \d, etc) is present in a test regex. Otherwise if the
332 test contains no characters greater than 255, Perl doesn't realise it
333 should be using Unicode semantics.
334
335 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
336 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
337 does in Unicode mode.
338
339 32. Added the "forbid" facility to pcretest so that putting tests into the
340 wrong test files can sometimes be quickly detected.
341
342 33. There is now a limit (default 250) on the depth of nesting of parentheses.
343 This limit is imposed to control the amount of system stack used at compile
344 time. It can be changed at build time by --with-parens-nest-limit=xxx or
345 the equivalent in CMake.
346
347 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
348 errors. Perl warns for these when in warning mode, but PCRE has no facility
349 for giving warnings.
350
351 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
352 However, this was not working for (?!) because it is optimized to (*FAIL),
353 for which PCRE does not allow quantifiers. The optimization is now disabled
354 when a quantifier follows (?!). I can't see any use for this, but it makes
355 things uniform.
356
357 36. Perl no longer allows group names to start with digits, so I have made this
358 change also in PCRE. It simplifies the code a bit.
359
360 37. In extended mode, Perl ignores spaces before a + that indicates a
361 possessive quantifier. PCRE allowed a space before the quantifier, but not
362 before the possessive +. It now does.
363
364 38. The use of \K (reset reported match start) within a repeated possessive
365 group such as (a\Kb)*+ was not working.
366
367 40. Document that the same character tables must be used at compile time and
368 run time, and that the facility to pass tables to pcre_exec() and
369 pcre_dfa_exec() is for use only with saved/restored patterns.
370
371 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
372 features for Builds with MSVC:
373
374 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
375 builds). The .rc files can be used to set FileDescription and many other
376 attributes.
377
378 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
379 This allows higher-level build scripts which want .pdb files to avoid
380 hard-coding the exact files needed."
381
382 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
383 mean "start of word" and "end of word", respectively, as a transition aid.
384
385 43. A minimizing repeat of a class containing codepoints greater than 255 in
386 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
387 compiled to use the heap for recursion.
388
389 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
390 is configured.
391
392
393 Version 8.33 28-May-2013
394 ------------------------
395
396 1. Added 'U' to some constants that are compared to unsigned integers, to
397 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
398 variables that are added to signed variables, to ensure the result is
399 signed and can be negated.
400
401 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
402 CMake config files.
403
404 3. Revise the creation of config.h.generic so that all boolean macros are
405 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
406 overriding via -D on the command line possible.
407
408 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
409 to unsigned int is reported to make a quite noticeable speed difference in
410 a specific Windows environment. Testing on Linux did also appear to show
411 some benefit (and it is clearly not harmful). Also fixed the definition of
412 Xop which should be unsigned.
413
414 5. Related to (4), changing the definition of the intermediate variable cc
415 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
416 improvements.
417
418 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
419 unnecessary spaces.
420
421 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
422 and later.
423
424 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
425
426 9. Optimizing fast_forward_start_bits in JIT.
427
428 10. Adding support for callouts in JIT, and fixing some issues revealed
429 during this work. Namely:
430
431 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
432
433 (b) Minimum length was not checked before the matching is started.
434
435 11. The value of capture_last that is passed to callouts was incorrect in some
436 cases when there was a capture on one path that was subsequently abandoned
437 after a backtrack. Also, the capture_last value is now reset after a
438 recursion, since all captures are also reset in this case.
439
440 12. The interpreter no longer returns the "too many substrings" error in the
441 case when an overflowing capture is in a branch that is subsequently
442 abandoned after a backtrack.
443
444 13. In the pathological case when an offset vector of size 2 is used, pcretest
445 now prints out the matched string after a yield of 0 or 1.
446
447 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
448 Only supported by the JIT compiler at the moment.
449
450 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
451
452 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
453 offset of the starting point of the matching process, provided the offsets
454 vector is large enough.
455
456 17. The \A escape now records a lookbehind value of 1, though its execution
457 does not actually inspect the previous character. This is to ensure that,
458 in partial multi-segment matching, at least one character from the old
459 segment is retained when a new segment is processed. Otherwise, if there
460 are no lookbehinds in the pattern, \A might match incorrectly at the start
461 of a new segment.
462
463 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
464
465 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
466 gives some modest performance improvement in 8-bit mode.
467
468 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
469 be expressed in certain programming languages using Universal Character
470 Names.
471
472 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
473 which points out that "non characters" are not "characters that may not
474 appear in Unicode strings" but rather "characters that are reserved for
475 internal use and have only local meaning".
476
477 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
478 there was a conditional group that depended on an assertion, if the
479 assertion was false, the callout that immediately followed the alternation
480 in the condition was skipped when pcre_exec() was used for matching.
481
482 23. Allow an explicit callout to be inserted before an assertion that is the
483 condition for a conditional group, for compatibility with automatic
484 callouts, which always insert a callout at this point.
485
486 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
487 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
488
489 25. (*PRUNE) is now supported by the JIT compiler.
490
491 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
492
493 27. Fix the case where there are two or more SKIPs with arguments that may be
494 ignored.
495
496 28. (*SKIP) is now supported by the JIT compiler.
497
498 29. (*THEN) is now supported by the JIT compiler.
499
500 30. Update RunTest with additional test selector options.
501
502 31. The way PCRE handles backtracking verbs has been changed in two ways.
503
504 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
505 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
506 backtracking. In some cases this makes it more Perl-compatible, but Perl's
507 rather obscure rules do not always do the same thing.
508
509 (2) Previously, backtracking verbs were confined within assertions. This is
510 no longer the case for positive assertions, except for (*ACCEPT). Again,
511 this sometimes improves Perl compatibility, and sometimes does not.
512
513 32. A number of tests that were in test 2 because Perl did things differently
514 have been moved to test 1, because either Perl or PCRE has changed, and
515 these tests are now compatible.
516
517 32. Backtracking control verbs are now handled in the same way in JIT and
518 interpreter.
519
520 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
521 contained a forward subroutine reference caused a compile error.
522
523 34. Auto-detect and optimize limited repetitions in JIT.
524
525 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
526 blocking (*UTF) etc.
527
528 36. In the interpreter, maximizing pattern repetitions for characters and
529 character types now use tail recursion, which reduces stack usage.
530
531 37. The value of the max lookbehind was not correctly preserved if a compiled
532 and saved regex was reloaded on a host of different endianness.
533
534 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
535 of the compiled pattern block, expand the flags field from 16 to 32 bits
536 because it was almost full.
537
538 39. Try madvise first before posix_madvise.
539
540 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
541 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
542 original change locked out all matches of empty strings. This has been
543 changed so that one match of an empty string per line is recognized.
544 Subsequent searches on the same line (for colouring or for --only-matching,
545 for example) do not recognize empty strings.
546
547 41. Applied a user patch to fix a number of spelling mistakes in comments.
548
549 42. Data lines longer than 65536 caused pcretest to crash.
550
551 43. Clarified the data type for length and startoffset arguments for pcre_exec
552 and pcre_dfa_exec in the function-specific man pages, where they were
553 explicitly stated to be in bytes, never having been updated. I also added
554 some clarification to the pcreapi man page.
555
556 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
557 a segmentation fault.
558
559
560 Version 8.32 30-November-2012
561 -----------------------------
562
563 1. Improved JIT compiler optimizations for first character search and single
564 character iterators.
565
566 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
567 Patch by Daniel Richard G.
568
569 3. Single character iterator optimizations in the JIT compiler.
570
571 4. Improved JIT compiler optimizations for character ranges.
572
573 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
574 Reported by Giuseppe D'Angelo.
575
576 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
577 of a line, was being set incorrectly in cases where .* appeared inside
578 atomic brackets at the start of a pattern, or where there was a subsequent
579 *PRUNE or *SKIP.
580
581 7. Improved instruction cache flush for POWER/PowerPC.
582 Patch by Daniel Richard G.
583
584 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
585 grep:
586
587 (a) There is now no limit to the number of patterns to be matched.
588
589 (b) An error is given if a pattern is too long.
590
591 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
592 are now supported.
593
594 (d) --exclude-from and --include-from (multiple use) have been added.
595
596 (e) Exclusions and inclusions now apply to all files and directories, not
597 just to those obtained from scanning a directory recursively.
598
599 (f) Multiple uses of -f and --file-list are now supported.
600
601 (g) In a Windows environment, the default for -d has been changed from
602 "read" (the GNU grep default) to "skip", because otherwise the presence
603 of a directory in the file list provokes an error.
604
605 (h) The documentation has been revised and clarified in places.
606
607 9. Improve the matching speed of capturing brackets.
608
609 10. Changed the meaning of \X so that it now matches a Unicode extended
610 grapheme cluster.
611
612 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
613 out POSIX threads when JIT support is configured.
614
615 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
616
617 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
618 structure could go wrong in environments where size_t is not the same size
619 as int.
620
621 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
622 set.
623
624 15. The EBCDIC support had decayed; later updates to the code had included
625 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
626 general tidy up of EBCDIC-related issues, and the documentation was also
627 not quite right. There is now a test that can be run on ASCII systems to
628 check some of the EBCDIC-related things (but is it not a full test).
629
630 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
631 in a small tidy to the code.
632
633 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
634
635 18. If the --only-matching (-o) option in pcregrep is specified multiple
636 times, each one causes appropriate output. For example, -o1 -o2 outputs the
637 substrings matched by the 1st and 2nd capturing parentheses. A separating
638 string can be specified by --om-separator (default empty).
639
640 19. Improving the first n character searches.
641
642 20. Turn case lists for horizontal and vertical white space into macros so that
643 they are defined only once.
644
645 21. This set of changes together give more compatible Unicode case-folding
646 behaviour for characters that have more than one other case when UCP
647 support is available.
648
649 (a) The Unicode property table now has offsets into a new table of sets of
650 three or more characters that are case-equivalent. The MultiStage2.py
651 script that generates these tables (the pcre_ucd.c file) now scans
652 CaseFolding.txt instead of UnicodeData.txt for character case
653 information.
654
655 (b) The code for adding characters or ranges of characters to a character
656 class has been abstracted into a generalized function that also handles
657 case-independence. In UTF-mode with UCP support, this uses the new data
658 to handle characters with more than one other case.
659
660 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
661 whose other case is greater than 256 are now correctly matched
662 caselessly. Previously, the high codepoint matched the low one, but not
663 vice versa.
664
665 (d) The processing of \h, \H, \v, and \ in character classes now makes use
666 of the new class addition function, using character lists defined as
667 macros alongside the case definitions of 20 above.
668
669 (e) Caseless back references now work with characters that have more than
670 one other case.
671
672 (f) General caseless matching of characters with more than one other case
673 is supported.
674
675 22. Unicode character properties were updated from Unicode 6.2.0
676
677 23. Improved CMake support under Windows. Patch by Daniel Richard G.
678
679 24. Add support for 32-bit character strings, and UTF-32
680
681 25. Major JIT compiler update (code refactoring and bugfixing).
682 Experimental Sparc 32 support is added.
683
684 26. Applied a modified version of Daniel Richard G's patch to create
685 pcre.h.generic and config.h.generic by "make" instead of in the
686 PrepareRelease script.
687
688 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
689 pcre_compile.c when checking for a zero character.
690
691 28. Introducing a native interface for JIT. Through this interface, the compiled
692 machine code can be directly executed. The purpose of this interface is to
693 provide fast pattern matching, so several sanity checks are not performed.
694 However, feature tests are still performed. The new interface provides
695 1.4x speedup compared to the old one.
696
697 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
698 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
699 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
700
701 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
702 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
703 the "old" RFC 2279). Instead, it ended up passing a negative length to
704 pcre_exec().
705
706 31. Add support for GCC's visibility feature to hide internal functions.
707
708 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
709 "unknown -C option" after outputting 0 or 1.
710
711 33. There is now support for generating a code coverage report for the test
712 suite in environments where gcc is the compiler and lcov is installed. This
713 is mainly for the benefit of the developers.
714
715 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
716 unaddressable using valgrind annotations, allowing valgrind to detect
717 invalid memory accesses. This is mainly for the benefit of the developers.
718
719 25. (*UTF) can now be used to start a pattern in any of the three libraries.
720
721 26. Give configure error if --enable-cpp but no C++ compiler found.
722
723
724 Version 8.31 06-July-2012
725 -------------------------
726
727 1. Fixing a wrong JIT test case and some compiler warnings.
728
729 2. Removed a bashism from the RunTest script.
730
731 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
732 to unsigned type, result still unsigned" that was given by an MS compiler
733 on encountering the code "-sizeof(xxx)".
734
735 4. Partial matching support is added to the JIT compiler.
736
737 5. Fixed several bugs concerned with partial matching of items that consist
738 of more than one character:
739
740 (a) /^(..)\1/ did not partially match "aba" because checking references was
741 done on an "all or nothing" basis. This also applied to repeated
742 references.
743
744 (b) \R did not give a hard partial match if \r was found at the end of the
745 subject.
746
747 (c) \X did not give a hard partial match after matching one or more
748 characters at the end of the subject.
749
750 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
751 a partial match for the string "\r".
752
753 (e) When newline was set to CRLF, the metacharacter "." did not recognize
754 a partial match for a CR character at the end of the subject string.
755
756 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
757 running pcretest, the text "(JIT)" added to the output whenever JIT is
758 actually used to run the match.
759
760 7. Individual JIT compile options can be set in pcretest by following -s+[+]
761 or /S+[+] with a digit between 1 and 7.
762
763 8. OP_NOT now supports any UTF character not just single-byte ones.
764
765 9. (*MARK) control verb is now supported by the JIT compiler.
766
767 10. The command "./RunTest list" lists the available tests without actually
768 running any of them. (Because I keep forgetting what they all are.)
769
770 11. Add PCRE_INFO_MAXLOOKBEHIND.
771
772 12. Applied a (slightly modified) user-supplied patch that improves performance
773 when the heap is used for recursion (compiled with --disable-stack-for-
774 recursion). Instead of malloc and free for each heap frame each time a
775 logical recursion happens, frames are retained on a chain and re-used where
776 possible. This sometimes gives as much as 30% improvement.
777
778 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
779 call.
780
781 14. As documented, (*COMMIT) is now confined to within a positive assertion.
782
783 15. It is now possible to link pcretest with libedit as an alternative to
784 libreadline.
785
786 16. (*COMMIT) control verb is now supported by the JIT compiler.
787
788 17. The Unicode data tables have been updated to Unicode 6.1.0.
789
790 18. Added --file-list option to pcregrep.
791
792 19. Added binary file support to pcregrep, including the -a, --binary-files,
793 -I, and --text options.
794
795 20. The madvise function is renamed for posix_madvise for QNX compatibility
796 reasons. Fixed by Giuseppe D'Angelo.
797
798 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
799
800 22. Changed the option for creating symbolic links for 16-bit man pages from
801 -s to -sf so that re-installing does not cause issues.
802
803 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
804
805 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
806 matches in certain environments (the workspace was not being correctly
807 retained). Also added to pcre_dfa_exec() a simple plausibility check on
808 some of the workspace data at the beginning of a restart.
809
810 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
811 was not doing so when it should - probably a typo introduced by SVN 528
812 (change 8.10/14).
813
814 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
815 \w+ when the character tables indicated that \x{c4} was a word character.
816 There were several related cases, all because the tests for doing a table
817 lookup were testing for characters less than 127 instead of 255.
818
819 27. If a pattern contains capturing parentheses that are not used in a match,
820 their slots in the ovector are set to -1. For those that are higher than
821 any matched groups, this happens at the end of processing. In the case when
822 there were back references that the ovector was too small to contain
823 (causing temporary malloc'd memory to be used during matching), and the
824 highest capturing number was not used, memory off the end of the ovector
825 was incorrectly being set to -1. (It was using the size of the temporary
826 memory instead of the true size.)
827
828 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
829 ovector size, it uses memory at the end of the block that it has got.
830
831 29. Check for an overlong MARK name and give an error at compile time. The
832 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
833
834 30. JIT compiler update.
835
836 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
837 Rill for the patch.
838
839 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
840
841 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
842
843 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
844 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
845 was enabled.
846
847 35. Improve JIT code generation for greedy plus quantifier.
848
849 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
850 1 to "aa" instead of to an empty string. The bug affected repeated groups
851 that could potentially match an empty string.
852
853 37. Optimizing single character iterators in JIT.
854
855 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
856 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
857 codepoints that are too big for the mode are faulted, and in a UTF mode,
858 disallowed codepoints are also faulted.
859
860 39. If PCRE was compiled with UTF support, in three places in the DFA
861 matcher there was code that should only have been obeyed in UTF mode, but
862 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
863 processing when bytes with values greater than 127 were present. In 16-bit
864 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
865 both cases the values are those that cannot be the first data item in a UTF
866 character. The three items that might have provoked this were recursions,
867 possessively repeated groups, and atomic groups.
868
869 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
870 and pcregrep, because some OS require shared objects to be explicitly
871 passed to ld, causing the link step to fail if they are not.
872
873 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
874 mode, patterns that started with \h* or \R* might be incorrectly matched.
875
876
877 Version 8.30 04-February-2012
878 -----------------------------
879
880 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
881 name is defined in ctype.h.
882
883 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
884 only in quite long subpatterns.
885
886 3. Removed the function pcre_info(), which has been obsolete and deprecated
887 since it was replaced by pcre_fullinfo() in February 2000.
888
889 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
890 match a (*MARK), and the match failed at the start of the subject, a
891 reference to memory before the start of the subject could occur. This bug
892 was introduced by fix 17 of release 8.21.
893
894 5. A reference to an unset group with zero minimum repetition was giving
895 totally wrong answers (in non-JavaScript-compatibility mode). For example,
896 /(another)?(\1?)test/ matched against "hello world test". This bug was
897 introduced in release 8.13.
898
899 6. Add support for 16-bit character strings (a large amount of work involving
900 many changes and refactorings).
901
902 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
903 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
904 from a file.
905
906 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
907 rounding is not applied in this particular case).
908
909 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
910 if they appear, or are escaped, in patterns.
911
912 10. Get rid of a number of -Wunused-but-set-variable warnings.
913
914 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
915 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
916 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
917 also returns the mark "x". This bug applied to capturing parentheses,
918 non-capturing parentheses, and atomic parentheses. It also applied to some
919 assertions.
920
921 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
922 information out of configure.ac instead of relying on pcre.h.generic, which
923 is not stored in the repository.
924
925 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
926 -lreadline.
927
928 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
929
930 15. Applied Graycode's patch to put the top-level frame on the stack rather
931 than the heap when not using the stack for recursion. This gives a
932 performance improvement in many cases when recursion is not deep.
933
934 16. Experimental code added to "pcretest -C" to output the stack frame size.
935
936
937 Version 8.21 12-Dec-2011
938 ------------------------
939
940 1. Updating the JIT compiler.
941
942 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
943 are added as well.
944
945 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
946 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
947 calling _pcre_jit_exec. Some extra comments are added.
948
949 4. (*MARK) settings inside atomic groups that do not contain any capturing
950 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
951 was introduced by change 18 for 8.20.
952
953 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
954 ECMA-262 standard.
955
956 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
957 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
958 This bug was probably introduced by change 9 of 8.13.
959
960 7. While fixing 6 above, I noticed that a number of other items were being
961 incorrectly rejected as "not fixed length". This arose partly because newer
962 opcodes had not been added to the fixed-length checking code. I have (a)
963 corrected the bug and added tests for these items, and (b) arranged for an
964 error to occur if an unknown opcode is encountered while checking for fixed
965 length instead of just assuming "not fixed length". The items that were
966 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
967 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
968 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
969
970 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
971 being incorrectly compiled and would have given unpredicatble results.
972
973 9. A possessively repeated subpattern with minimum repeat count greater than
974 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
975 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
976 the first (A) could occur when it should not.
977
978 10. Add a cast and remove a redundant test from the code.
979
980 11. JIT should use pcre_malloc/pcre_free for allocation.
981
982 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
983 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
984 is anything other than /usr, -L is still shown).
985
986 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
987
988 14. Perl does not support \N without a following name in a [] class; PCRE now
989 also gives an error.
990
991 15. If a forward reference was repeated with an upper limit of around 2000,
992 it caused the error "internal error: overran compiling workspace". The
993 maximum number of forward references (including repeats) was limited by the
994 internal workspace, and dependent on the LINK_SIZE. The code has been
995 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
996 the default depends on LINK_SIZE. There is a new upper limit (for safety)
997 of around 200,000 forward references. While doing this, I also speeded up
998 the filling in of repeated forward references.
999
1000 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1001 incorrectly expecting the subject to contain another "a" after the start.
1002
1003 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1004 in the match, the SKIP should be ignored. This was not happening; instead
1005 the SKIP was being treated as NOMATCH. For patterns such as
1006 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1007 tested.
1008
1009 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1010 now much more compatible with Perl, in particular in cases where the result
1011 is a non-match for a non-anchored pattern. For example, if
1012 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1013 "m", where previously it did not return a name. A side effect of this
1014 change is that for partial matches, the last encountered mark name is
1015 returned, as for non matches. A number of tests that were previously not
1016 Perl-compatible have been moved into the Perl-compatible test files. The
1017 refactoring has had the pleasing side effect of removing one argument from
1018 the match() function, thus reducing its stack requirements.
1019
1020 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1021 subsequent uses of /S (without +) incorrectly behaved like /S+.
1022
1023 21. Retrieve executable code size support for the JIT compiler and fixing
1024 some warnings.
1025
1026 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1027 not work when the shorter character appeared right at the end of the
1028 subject string.
1029
1030 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1031 systems.
1032
1033 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1034 output it when the /M option is used in pcretest.
1035
1036 25. The CheckMan script was not being included in the distribution. Also, added
1037 an explicit "perl" to run Perl scripts from the PrepareRelease script
1038 because this is reportedly needed in Windows.
1039
1040 26. If study data was being save in a file and studying had not found a set of
1041 "starts with" bytes for the pattern, the data written to the file (though
1042 never used) was taken from uninitialized memory and so caused valgrind to
1043 complain.
1044
1045 27. Updated RunTest.bat as provided by Sheri Pierce.
1046
1047 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1048
1049 29. Computation of memory usage for the table of capturing group names was
1050 giving an unnecessarily large value.
1051
1052
1053 Version 8.20 21-Oct-2011
1054 ------------------------
1055
1056 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1057 a POSIX class. After further experiments with Perl, which convinced me that
1058 Perl has bugs and confusions, a closing square bracket is no longer allowed
1059 in a POSIX name. This bug also affected patterns with classes that started
1060 with full stops.
1061
1062 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1063 captured substring, but while checking the failing first alternative,
1064 substring 1 is temporarily captured. If the output vector supplied to
1065 pcre_exec() was not big enough for this capture, the yield of the function
1066 was still zero ("insufficient space for captured substrings"). This cannot
1067 be totally fixed without adding another stack variable, which seems a lot
1068 of expense for a edge case. However, I have improved the situation in cases
1069 such as /(a)(b)x|abc/ matched against "abc", where the return code
1070 indicates that fewer than the maximum number of slots in the ovector have
1071 been set.
1072
1073 3. Related to (2) above: when there are more back references in a pattern than
1074 slots in the output vector, pcre_exec() uses temporary memory during
1075 matching, and copies in the captures as far as possible afterwards. It was
1076 using the entire output vector, but this conflicts with the specification
1077 that only 2/3 is used for passing back captured substrings. Now it uses
1078 only the first 2/3, for compatibility. This is, of course, another edge
1079 case.
1080
1081 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1082 main code base, and can be used by building with --enable-jit. When this is
1083 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1084 runtime --no-jit option is given.
1085
1086 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1087 ovector, the return from the function was zero, implying that there were
1088 other matches that did not fit. The correct "exactly full" value is now
1089 returned.
1090
1091 6. If a subpattern that was called recursively or as a subroutine contained
1092 (*PRUNE) or any other control that caused it to give a non-standard return,
1093 invalid errors such as "Error -26 (nested recursion at the same subject
1094 position)" or even infinite loops could occur.
1095
1096 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1097 computing the minimum length on reaching *ACCEPT, and so ended up with the
1098 wrong value of 1 rather than 0. Further investigation indicates that
1099 computing a minimum subject length in the presence of *ACCEPT is difficult
1100 (think back references, subroutine calls), and so I have changed the code
1101 so that no minimum is registered for a pattern that contains *ACCEPT.
1102
1103 8. If (*THEN) was present in the first (true) branch of a conditional group,
1104 it was not handled as intended. [But see 16 below.]
1105
1106 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1107 Sheri Pierce.
1108
1109 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1110 the first byte in a match must be "a".
1111
1112 11. Change 17 for 8.13 increased the recursion depth for patterns like
1113 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1114 pattern contains any instances of (*THEN). If it does not, the old
1115 optimizations are restored. It would be nice to do this on a per-group
1116 basis, but at the moment that is not feasible.
1117
1118 12. In some environments, the output of pcretest -C is CRLF terminated. This
1119 broke RunTest's code that checks for the link size. A single white space
1120 character after the value is now allowed for.
1121
1122 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1123 For "fr", it uses the Windows-specific input and output files.
1124
1125 14. If (*THEN) appeared in a group that was called recursively or as a
1126 subroutine, it did not work as intended. [But see next item.]
1127
1128 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1129 pattern fragments (but not containing any | characters). If A and B are
1130 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1131 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1132 D. In other words, Perl considers parentheses that do not contain any |
1133 characters to be part of a surrounding alternative, whereas PCRE was
1134 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1135 differently. PCRE now behaves in the same way as Perl, except in the case
1136 of subroutine/recursion calls such as (?1) which have in any case always
1137 been different (but PCRE had them first :-).
1138
1139 16. Related to 15 above: Perl does not treat the | in a conditional group as
1140 creating alternatives. Such a group is treated in the same way as an
1141 ordinary group without any | characters when processing (*THEN). PCRE has
1142 been changed to match Perl's behaviour.
1143
1144 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1145 RunGrepTest script failed.
1146
1147 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1148 inevitable for groups that contain captures, but it can lead to a lot of
1149 stack use in large patterns. The old behaviour has been restored for atomic
1150 groups that do not contain any capturing parentheses.
1151
1152 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1153 suppress the check for a minimum subject length at run time. (If it was
1154 given to pcre_exec() or pcre_dfa_exec() it did work.)
1155
1156 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1157 fail to work when decoding hex characters in data strings in EBCDIC
1158 environments.
1159
1160 21. It appears that in at least one Mac OS environment, the isxdigit() function
1161 is implemented as a macro that evaluates to its argument more than once,
1162 contravening the C 90 Standard (I haven't checked a later standard). There
1163 was an instance in pcretest which caused it to go wrong when processing
1164 \x{...} escapes in subject strings. The has been rewritten to avoid using
1165 things like p++ in the argument of isxdigit().
1166
1167
1168 Version 8.13 16-Aug-2011
1169 ------------------------
1170
1171 1. The Unicode data tables have been updated to Unicode 6.0.0.
1172
1173 2. Two minor typos in pcre_internal.h have been fixed.
1174
1175 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1176 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1177 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1178
1179 4. There were a number of related bugs in the code for matching backrefences
1180 caselessly in UTF-8 mode when codes for the characters concerned were
1181 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1182 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1183 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1184 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1185 2-byte code at the end of the subject (it thought there wasn't enough data
1186 left).
1187
1188 5. Comprehensive information about what went wrong is now returned by
1189 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1190 as the output vector has at least 2 elements. The offset of the start of
1191 the failing character and a reason code are placed in the vector.
1192
1193 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1194 now returned is for the first byte of the failing character, instead of the
1195 last byte inspected. This is an incompatible change, but I hope it is small
1196 enough not to be a problem. It makes the returned offset consistent with
1197 pcre_exec() and pcre_dfa_exec().
1198
1199 7. pcretest now gives a text phrase as well as the error number when
1200 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1201 failure, the offset and reason code are output.
1202
1203 8. When \R was used with a maximizing quantifier it failed to skip backwards
1204 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1205 back over a single character (\n). This seems wrong (because it treated the
1206 two characters as a single entity when going forwards), conflicts with the
1207 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1208 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1209 has been changed.
1210
1211 9. Some internal refactoring has changed the processing so that the handling
1212 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1213 time (the PCRE_DOTALL option was changed this way some time ago: version
1214 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1215 which was always a bit of a fudge. It also means that there is one less
1216 argument for the match() function, which reduces its stack requirements
1217 slightly. This change also fixes an incompatibility with Perl: the pattern
1218 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1219
1220 10. More internal refactoring has drastically reduced the number of recursive
1221 calls to match() for possessively repeated groups such as (abc)++ when
1222 using pcre_exec().
1223
1224 11. While implementing 10, a number of bugs in the handling of groups were
1225 discovered and fixed:
1226
1227 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1228 (a|)*(?1) gave a compile-time internal error.
1229 ((a|)+)+ did not notice that the outer group could match an empty string.
1230 (^a|^)+ was not marked as anchored.
1231 (.*a|.*)+ was not marked as matching at start or after a newline.
1232
1233 12. Yet more internal refactoring has removed another argument from the match()
1234 function. Special calls to this function are now indicated by setting a
1235 value in a variable in the "match data" data block.
1236
1237 13. Be more explicit in pcre_study() instead of relying on "default" for
1238 opcodes that mean there is no starting character; this means that when new
1239 ones are added and accidentally left out of pcre_study(), testing should
1240 pick them up.
1241
1242 14. The -s option of pcretest has been documented for ages as being an old
1243 synonym of -m (show memory usage). I have changed it to mean "force study
1244 for every regex", that is, assume /S for every regex. This is similar to -i
1245 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1246 using it. It makes it easier to run collections of tests with and without
1247 study enabled, and thereby test pcre_study() more easily. All the standard
1248 tests are now run with and without -s (but some patterns can be marked as
1249 "never study" - see 20 below).
1250
1251 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1252 restoration of the capturing data to the outer values was not happening
1253 correctly.
1254
1255 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1256 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1257 pattern had matched an empty string, and so incorrectly returned a no
1258 match.
1259
1260 17. There was optimizing code for the last branch of non-capturing parentheses,
1261 and also for the obeyed branch of a conditional subexpression, which used
1262 tail recursion to cut down on stack usage. Unfortunately, now that there is
1263 the possibility of (*THEN) occurring in these branches, tail recursion is
1264 no longer possible because the return has to be checked for (*THEN). These
1265 two optimizations have therefore been removed. [But see 8.20/11 above.]
1266
1267 18. If a pattern containing \R was studied, it was assumed that \R always
1268 matched two bytes, thus causing the minimum subject length to be
1269 incorrectly computed because \R can also match just one byte.
1270
1271 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1272 was incorrectly computed.
1273
1274 20. If /S is present twice on a test pattern in pcretest input, it now
1275 *disables* studying, thereby overriding the use of -s on the command line
1276 (see 14 above). This is necessary for one or two tests to keep the output
1277 identical in both cases.
1278
1279 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1280 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1281
1282 22. When an atomic group that contained a capturing parenthesis was
1283 successfully matched, but the branch in which it appeared failed, the
1284 capturing was not being forgotten if a higher numbered group was later
1285 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1286 group 1 to "a", when in fact it should be unset. This applied to multi-
1287 branched capturing and non-capturing groups, repeated or not, and also to
1288 positive assertions (capturing in negative assertions does not happen
1289 in PCRE) and also to nested atomic groups.
1290
1291 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1292 subject after a captured substring, to make it easier to tell which of a
1293 number of identical substrings has been captured.
1294
1295 24. The way atomic groups are processed by pcre_exec() has been changed so that
1296 if they are repeated, backtracking one repetition now resets captured
1297 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1298 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1299 "aaa". Previously, it would have been "a". As part of this code
1300 refactoring, the way recursive calls are handled has also been changed.
1301
1302 25. If an assertion condition captured any substrings, they were not passed
1303 back unless some other capturing happened later. For example, if
1304 (?(?=(a))a) was matched against "a", no capturing was returned.
1305
1306 26. When studying a pattern that contained subroutine calls or assertions,
1307 the code for finding the minimum length of a possible match was handling
1308 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1309 group 1 called group 2 while simultaneously a separate group 2 called group
1310 1). A stack overflow occurred in this case. I have fixed this by limiting
1311 the recursion depth to 10.
1312
1313 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1314 Fortmann. This supports explicit test numbers on the command line, and has
1315 argument validation and error reporting.
1316
1317 28. An instance of \X with an unlimited repeat could fail if at any point the
1318 first character it looked at was a mark character.
1319
1320 29. Some minor code refactoring concerning Unicode properties and scripts
1321 should reduce the stack requirement of match() slightly.
1322
1323 30. Added the '=' option to pcretest to check the setting of unused capturing
1324 slots at the end of the pattern, which are documented as being -1, but are
1325 not included in the return count.
1326
1327 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1328 compiled something random. Now it gives a compile-time error (as does
1329 Perl).
1330
1331 32. A *MARK encountered during the processing of a positive assertion is now
1332 recorded and passed back (compatible with Perl).
1333
1334 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1335 had alternative anchored branches, the search for a second match in a line
1336 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1337 matched the line "0102" twice. The same bug affected patterns that started
1338 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1339 twice.
1340
1341 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1342 does, and because of capturing effects, quantifying parenthesized
1343 assertions may at times be useful. Quantifiers are now allowed for
1344 parenthesized assertions.
1345
1346 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1347
1348 36. \g was being checked for fancy things in a character class, when it should
1349 just be a literal "g".
1350
1351 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1352 appearance of a nested POSIX class supersedes an apparent external class.
1353 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1354 unescaped square brackets may also appear as part of class names. For
1355 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1356 more like Perl. (But see 8.20/1 above.)
1357
1358 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1359 was because it thought it was \N{name}, which is not supported).
1360
1361 39. Add minix to OS list not supporting the -S option in pcretest.
1362
1363 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1364 cannot analyze patterns in sufficient detail to catch mutual recursions
1365 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1366 subgroup is called recursively as a subpattern for a second time at the
1367 same position in the subject string. In previous releases this might have
1368 been caught by the recursion limit, or it might have run out of stack.
1369
1370 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1371 happen only once. PCRE was, however incorrectly giving a compile time error
1372 "recursive call could loop indefinitely" because it cannot analyze the
1373 pattern in sufficient detail. The compile time test no longer happens when
1374 PCRE is compiling a conditional subpattern, but actual runaway loops are
1375 now caught at runtime (see 40 above).
1376
1377 42. It seems that Perl allows any characters other than a closing parenthesis
1378 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1379 has been changed to be the same.
1380
1381 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1382 as not to get warnings when autogen.sh is called. Also changed
1383 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1384
1385 44. To help people who use pcregrep to scan files containing exceedingly long
1386 lines, the following changes have been made:
1387
1388 (a) The default value of the buffer size parameter has been increased from
1389 8K to 20K. (The actual buffer used is three times this size.)
1390
1391 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1392 PCRE is built.
1393
1394 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1395 to be set at run time.
1396
1397 (d) Numerical values in pcregrep options can be followed by K or M, for
1398 example --buffer-size=50K.
1399
1400 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1401 given and the return code is set to 2.
1402
1403 45. Add a pointer to the latest mark to the callout data block.
1404
1405 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1406 partial match of an empty string instead of no match. This was specific to
1407 the use of ".".
1408
1409 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1410 complete match instead of a partial match. This bug was dependent on both
1411 the PCRE_UTF8 and PCRE_DOTALL options being set.
1412
1413 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1414 starting byte set, because \b was not being ignored.
1415
1416
1417 Version 8.12 15-Jan-2011
1418 ------------------------
1419
1420 1. Fixed some typos in the markup of the man pages, and wrote a script that
1421 checks for such things as part of the documentation building process.
1422
1423 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1424 --match-limit and --recursion-limit options (added for 8.11). In
1425 particular, this made one of the standard tests fail. (The integer value
1426 went into the wrong half of a long int.)
1427
1428 3. If the --colour option was given to pcregrep with -v (invert match), it
1429 did strange things, either producing crazy output, or crashing. It should,
1430 of course, ignore a request for colour when reporting lines that do not
1431 match.
1432
1433 4. Another pcregrep bug caused similar problems if --colour was specified with
1434 -M (multiline) and the pattern match finished with a line ending.
1435
1436 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1437 matched in multiline mode, the following line was shown as part of the
1438 match. This seems wrong, so I have changed it.
1439
1440 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1441 the check for further matches in the same line (so they could be coloured)
1442 to overrun the end of the current line. If another match was found, it was
1443 incorrectly shown (and then shown again when found in the next line).
1444
1445 7. If pcregrep was compiled under Windows, there was a reference to the
1446 function pcregrep_exit() before it was defined. I am assuming this was
1447 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1448 reported by a user. I've moved the definition above the reference.
1449
1450
1451 Version 8.11 10-Dec-2010
1452 ------------------------
1453
1454 1. (*THEN) was not working properly if there were untried alternatives prior
1455 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1456 backtracked to try for "b" instead of moving to the next alternative branch
1457 at the same level (in this case, to look for "c"). The Perl documentation
1458 is clear that when (*THEN) is backtracked onto, it goes to the "next
1459 alternative in the innermost enclosing group".
1460
1461 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1462 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1463 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1464 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1465 (*THEN).
1466
1467 3. If \s appeared in a character class, it removed the VT character from
1468 the class, even if it had been included by some previous item, for example
1469 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1470 of \s, but is part of the POSIX "space" class.)
1471
1472 4. A partial match never returns an empty string (because you can always
1473 match an empty string at the end of the subject); however the checking for
1474 an empty string was starting at the "start of match" point. This has been
1475 changed to the "earliest inspected character" point, because the returned
1476 data for a partial match starts at this character. This means that, for
1477 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1478 (previously it gave "no match").
1479
1480 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1481 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1482 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1483 has an implication that the given string is incomplete (because a partial
1484 match is preferred over a full match). For this reason, these items now
1485 give a partial match in this situation. [Aside: previously, the one case
1486 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1487 match rather than a full match, which was wrong by the old rules, but is
1488 now correct.]
1489
1490 6. There was a bug in the handling of #-introduced comments, recognized when
1491 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1492 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1493 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1494 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1495 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1496 places in pcre_compile().
1497
1498 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1499 comments when looking ahead for named forward references to subpatterns,
1500 the only newline sequence it recognized was NL. It now handles newlines
1501 according to the set newline convention.
1502
1503 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1504 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1505 cater for a lack of strerror(). These oversights have been fixed.
1506
1507 9. Added --match-limit and --recursion-limit to pcregrep.
1508
1509 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1510
1511 11. When the -o option was used, pcregrep was setting a return code of 1, even
1512 when matches were found, and --line-buffered was not being honoured.
1513
1514 12. Added an optional parentheses number to the -o and --only-matching options
1515 of pcregrep.
1516
1517 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1518 can match an empty string. The code to do it in pcretest and pcredemo
1519 needed fixing:
1520
1521 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1522 only one byte after an empty string match just before CRLF (this case
1523 just got forgotten; "any" and "anycrlf" were OK).
1524
1525 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1526 mode when an empty string match preceded an ASCII character followed by
1527 a non-ASCII character. (The code for advancing by one character rather
1528 than one byte was nonsense.)
1529
1530 (c) The pcredemo.c sample program did not have any code at all to handle
1531 the cases when CRLF is a valid newline sequence.
1532
1533 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1534 as a starting offset was within the subject string. There is now a new
1535 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1536 negative or greater than the length of the string. In order to test this,
1537 pcretest is extended to allow the setting of negative starting offsets.
1538
1539 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1540 starting offset points to the beginning of a UTF-8 character was
1541 unnecessarily clumsy. I tidied it up.
1542
1543 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1544 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1545
1546 17. Nobody had reported that the --include_dir option, which was added in
1547 release 7.7 should have been called --include-dir (hyphen, not underscore)
1548 for compatibility with GNU grep. I have changed it to --include-dir, but
1549 left --include_dir as an undocumented synonym, and the same for
1550 --exclude-dir, though that is not available in GNU grep, at least as of
1551 release 2.5.4.
1552
1553 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1554 characters from a string of bytes) have been redefined so as not to use
1555 loops, in order to improve performance in some environments. At the same
1556 time, I abstracted some of the common code into auxiliary macros to save
1557 repetition (this should not affect the compiled code).
1558
1559 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1560 compile-time error is now given if \c is not followed by an ASCII
1561 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1562 different, and any byte value is allowed.)
1563
1564 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1565 START_OPTIMIZE option, which is now allowed at compile time - but just
1566 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1567 to pcregrep and other applications that have no direct access to PCRE
1568 options. The new /Y option in pcretest sets this option when calling
1569 pcre_compile().
1570
1571 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1572 back references. Groups containing recursive back references were forced to
1573 be atomic by that change, but in the case of named groups, the amount of
1574 memory required was incorrectly computed, leading to "Failed: internal
1575 error: code overflow". This has been fixed.
1576
1577 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1578 pcretest.c, to avoid build problems in some Borland environments.
1579
1580
1581 Version 8.10 25-Jun-2010
1582 ------------------------
1583
1584 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1585 THEN.
1586
1587 2. (*ACCEPT) was not working when inside an atomic group.
1588
1589 3. Inside a character class, \B is treated as a literal by default, but
1590 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1591 causes the error). The code is unchanged, but I tidied the documentation.
1592
1593 4. Inside a character class, PCRE always treated \R and \X as literals,
1594 whereas Perl faults them if its -w option is set. I have changed PCRE so
1595 that it faults them when PCRE_EXTRA is set.
1596
1597 5. Added support for \N, which always matches any character other than
1598 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1599
1600 6. When compiling pcregrep with newer versions of gcc which may have
1601 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1602 declared with attribute warn_unused_result" were given. Just casting the
1603 result to (void) does not stop the warnings; a more elaborate fudge is
1604 needed. I've used a macro to implement this.
1605
1606 7. Minor change to pcretest.c to avoid a compiler warning.
1607
1608 8. Added four artifical Unicode properties to help with an option to make
1609 \s etc use properties (see next item). The new properties are: Xan
1610 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1611
1612 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1613 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1614 this option. Modified pcretest to add /W to test this facility. Added
1615 REG_UCP to make it available via the POSIX interface.
1616
1617 10. Added --line-buffered to pcregrep.
1618
1619 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1620 studied, and the match started with a letter with a code point greater than
1621 127 whose first byte was different to the first byte of the other case of
1622 the letter, the other case of this starting letter was not recognized
1623 (#976).
1624
1625 12. If a pattern that was studied started with a repeated Unicode property
1626 test, for example, \p{Nd}+, there was the theoretical possibility of
1627 setting up an incorrect bitmap of starting bytes, but fortunately it could
1628 not have actually happened in practice until change 8 above was made (it
1629 added property types that matched character-matching opcodes).
1630
1631 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1632 possible starting bytes for non-anchored patterns.
1633
1634 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1635 \R, and also a number of cases that involve Unicode properties, both
1636 explicit and implicit when PCRE_UCP is set.
1637
1638 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1639 input, it could crash or give wrong results if characters with values
1640 greater than 0xc0 were present in the subject string. (Detail: it assumed
1641 UTF-8 input when processing these items.)
1642
1643 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1644 size_t is 64-bit (#991).
1645
1646 17. Added a check for running out of memory when PCRE is compiled with
1647 --disable-stack-for-recursion (#990).
1648
1649 18. If the last data line in a file for pcretest does not have a newline on
1650 the end, a newline was missing in the output.
1651
1652 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1653 less than 128) in its various bitmaps. However, there is a facility for
1654 generating tables according to the current locale when PCRE is compiled. It
1655 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1656 characters, are recognized by isspace() and therefore were getting set in
1657 these tables, and indeed these tables seem to approximate to ISO 8859. This
1658 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1659 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1660 which of course cannot start UTF-8 characters. I have changed the code so
1661 that only real ASCII characters (less than 128) and the correct starting
1662 bytes for UTF-8 encodings are set for characters greater than 127 when in
1663 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1664 altogether.)
1665
1666 20. Added the /T option to pcretest so as to be able to run tests with non-
1667 standard character tables, thus making it possible to include the tests
1668 used for 19 above in the standard set of tests.
1669
1670 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1671 reference to a subpattern the other side of a comment that contains an
1672 opening parenthesis caused either an internal compiling error, or a
1673 reference to the wrong subpattern.
1674
1675
1676 Version 8.02 19-Mar-2010
1677 ------------------------
1678
1679 1. The Unicode data tables have been updated to Unicode 5.2.0.
1680
1681 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1682 configured.
1683
1684 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1685 original author of that file, following a query about its status.
1686
1687 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1688 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1689
1690 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1691 quantifier applied to a forward-referencing subroutine call, could compile
1692 incorrect code or give the error "internal error: previously-checked
1693 referenced subpattern not found".
1694
1695 6. Both MS Visual Studio and Symbian OS have problems with initializing
1696 variables to point to external functions. For these systems, therefore,
1697 pcre_malloc etc. are now initialized to local functions that call the
1698 relevant global functions.
1699
1700 7. There were two entries missing in the vectors called coptable and poptable
1701 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1702 I've fixed the data, and added a kludgy way of testing at compile time that
1703 the lengths are correct (equal to the number of opcodes).
1704
1705 8. Following on from 7, I added a similar kludge to check the length of the
1706 eint vector in pcreposix.c.
1707
1708 9. Error texts for pcre_compile() are held as one long string to avoid too
1709 much relocation at load time. To find a text, the string is searched,
1710 counting zeros. There was no check for running off the end of the string,
1711 which could happen if a new error number was added without updating the
1712 string.
1713
1714 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1715
1716 11. \K was not working if it appeared in an atomic group or in a group that
1717 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1718 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1719 the assertion is positive, but not if it is negative.
1720
1721 12. Change 11 fortuitously reduced the size of the stack frame used in the
1722 "match()" function of pcre_exec.c by one pointer. Forthcoming
1723 implementation of support for (*MARK) will need an extra pointer on the
1724 stack; I have reserved it now, so that the stack frame size does not
1725 decrease.
1726
1727 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1728 item in branch that calls a recursion is a subroutine call - as in the
1729 second branch in the above example - was incorrectly given the compile-
1730 time error "recursive call could loop indefinitely" because pcre_compile()
1731 was not correctly checking the subroutine for matching a non-empty string.
1732
1733 14. The checks for overrunning compiling workspace could trigger after an
1734 overrun had occurred. This is a "should never occur" error, but it can be
1735 triggered by pathological patterns such as hundreds of nested parentheses.
1736 The checks now trigger 100 bytes before the end of the workspace.
1737
1738 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1739
1740
1741 Version 8.01 19-Jan-2010
1742 ------------------------
1743
1744 1. If a pattern contained a conditional subpattern with only one branch (in
1745 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1746 computed the wrong minimum data length (which is of course zero for such
1747 subpatterns). This could cause incorrect "no match" results.
1748
1749 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1750 the pattern is reset in the first branch, pcre_compile() failed with
1751 "internal error: code overflow at offset...". This happened only when
1752 the reset was to the original external option setting. (An optimization
1753 abstracts leading options settings into an external setting, which was the
1754 cause of this.)
1755
1756 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1757 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1758 assertion pattern did not match (meaning that the assertion was true), it
1759 was incorrectly treated as false if the SKIP had been reached during the
1760 matching. This also applied to assertions used as conditions.
1761
1762 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1763 assertion subpattern, including such a pattern used as a condition,
1764 unpredictable results occurred, instead of the error return
1765 PCRE_ERROR_DFA_UITEM.
1766
1767 5. The C++ GlobalReplace function was not working like Perl for the special
1768 situation when an empty string is matched. It now does the fancy magic
1769 stuff that is necessary.
1770
1771 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1772 removed. (These were left over from very, very early versions of PCRE.)
1773
1774 7. Some cosmetic changes to the code to make life easier when compiling it
1775 as part of something else:
1776
1777 (a) Change DEBUG to PCRE_DEBUG.
1778
1779 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1780 called "current" as "current_branch", to prevent a collision with the
1781 Linux macro when compiled as a kernel module.
1782
1783 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1784 prevent a collision with the Linux macro when compiled as a kernel
1785 module.
1786
1787 8. In pcre_compile() there are some checks for integer overflows that used to
1788 cast potentially large values to (double). This has been changed to that
1789 when building, a check for int64_t is made, and if it is found, it is used
1790 instead, thus avoiding the use of floating point arithmetic. (There is no
1791 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1792 double.
1793
1794 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1795 2005 (difference between two addresses compared to an unsigned value).
1796
1797 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1798 custom one, because of the following reported problem in Windows:
1799
1800 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1801 under Win32.
1802 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1803 therefore missing the function definition.
1804 - The compiler thus generates a "C" signature for the test function.
1805 - The linker fails to find the "C" function.
1806 - PCRE fails to configure if asked to do so against libbz2.
1807
1808 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1809 messages were output:
1810
1811 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1812 rerunning libtoolize, to keep the correct libtool macros in-tree.
1813 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1814
1815 I have done both of these things.
1816
1817 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1818 most of the time, it *can* run out if it is given a pattern that contains a
1819 runaway infinite recursion. I updated the discussion in the pcrestack man
1820 page.
1821
1822 13. Now that we have gone to the x.xx style of version numbers, the minor
1823 version may start with zero. Using 08 or 09 is a bad idea because users
1824 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1825 interpreted as invalid octal numbers. I've updated the previous comment in
1826 configure.ac, and also added a check that gives an error if 08 or 09 are
1827 used.
1828
1829 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1830 causing partial matching to fail when the end of the subject matched \W
1831 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1832
1833 15. There were some discrepancies between the declarations in pcre_internal.h
1834 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1835 their definitions. The declarations used "const uschar *" and the
1836 definitions used USPTR. Even though USPTR is normally defined as "const
1837 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1838 reported that: "This difference in casting confuses some C++ compilers, for
1839 example, SunCC recognizes above declarations as different functions and
1840 generates broken code for hbpcre." I have changed the declarations to use
1841 USPTR.
1842
1843 16. GNU libtool is named differently on some systems. The autogen.sh script now
1844 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1845 (FreeBSD).
1846
1847 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1848 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1849 comment: "Figure out how to create a longlong from a string: strtoll and
1850 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1851 instance, but it only takes 2 args instead of 3!"
1852
1853 18. A subtle bug concerned with back references has been fixed by a change of
1854 specification, with a corresponding code fix. A pattern such as
1855 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1856 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1857 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1858 same bug. Such groups have to be quantified to be useful, or contained
1859 inside another quantified group. (If there's no repetition, the reference
1860 can never match.) The problem arises because, having left the group and
1861 moved on to the rest of the pattern, a later failure that backtracks into
1862 the group uses the captured value from the final iteration of the group
1863 rather than the correct earlier one. I have fixed this in PCRE by forcing
1864 any group that contains a reference to itself to be an atomic group; that
1865 is, there cannot be any backtracking into it once it has completed. This is
1866 similar to recursive and subroutine calls.
1867
1868
1869 Version 8.00 19-Oct-09
1870 ----------------------
1871
1872 1. The table for translating pcre_compile() error codes into POSIX error codes
1873 was out-of-date, and there was no check on the pcre_compile() error code
1874 being within the table. This could lead to an OK return being given in
1875 error.
1876
1877 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1878 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1879 in a Windows environment.
1880
1881 3. The pcregrep --count option prints the count for each file even when it is
1882 zero, as does GNU grep. However, pcregrep was also printing all files when
1883 --files-with-matches was added. Now, when both options are given, it prints
1884 counts only for those files that have at least one match. (GNU grep just
1885 prints the file name in this circumstance, but including the count seems
1886 more useful - otherwise, why use --count?) Also ensured that the
1887 combination -clh just lists non-zero counts, with no names.
1888
1889 4. The long form of the pcregrep -F option was incorrectly implemented as
1890 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1891 but it seems right to fix it, and I didn't think it was worth preserving
1892 the old behaviour.
1893
1894 5. The command line items --regex=pattern and --regexp=pattern were not
1895 recognized by pcregrep, which required --regex pattern or --regexp pattern
1896 (with a space rather than an '='). The man page documented the '=' forms,
1897 which are compatible with GNU grep; these now work.
1898
1899 6. No libpcreposix.pc file was created for pkg-config; there was just
1900 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1901
1902 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1903 when UCP support is not needed, by modifying the Python script that
1904 generates it from Unicode data files. This should not matter if the module
1905 is correctly used as a library, but I received one complaint about 50K of
1906 unwanted data. My guess is that the person linked everything into his
1907 program rather than using a library. Anyway, it does no harm.
1908
1909 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1910 was a minimum greater than 1 for a wide character in a possessive
1911 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1912 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1913 character. Chaos in the form of incorrect output or a compiling loop could
1914 result.
1915
1916 9. The restrictions on what a pattern can contain when partial matching is
1917 requested for pcre_exec() have been removed. All patterns can now be
1918 partially matched by this function. In addition, if there are at least two
1919 slots in the offset vector, the offset of the earliest inspected character
1920 for the match and the offset of the end of the subject are set in them when
1921 PCRE_ERROR_PARTIAL is returned.
1922
1923 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1924 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1925 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1926 and may be more useful for multi-segment matching.
1927
1928 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1929 used to be given if ever the end of the subject was reached; now it is
1930 given only if matching could not proceed because another character was
1931 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1932 string "Z", which now yields "no match" instead of "partial match". In the
1933 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1934 final character ended with (*FAIL).
1935
1936 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1937 if the pattern had a "must contain" character that was already found in the
1938 earlier partial match, unless partial matching was again requested. For
1939 example, with the pattern /dog.(body)?/, the "must contain" character is
1940 "g". If the first part-match was for the string "dog", restarting with
1941 "sbody" failed. This bug has been fixed.
1942
1943 13. The string returned by pcre_dfa_exec() after a partial match has been
1944 changed so that it starts at the first inspected character rather than the
1945 first character of the match. This makes a difference only if the pattern
1946 starts with a lookbehind assertion or \b or \B (\K is not supported by
1947 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1948 matching functions compatible, and I think it's the right thing to do.
1949
1950 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1951 so that the demonstration program is easily available in environments where
1952 PCRE has not been installed from source.
1953
1954 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1955 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1956 library.
1957
1958 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1959 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1960 is not the first non-POSIX option to be added. Clearly some people find
1961 these options useful.
1962
1963 17. If a caller to the POSIX matching function regexec() passes a non-zero
1964 value for nmatch with a NULL value for pmatch, the value of
1965 nmatch is forced to zero.
1966
1967 18. RunGrepTest did not have a test for the availability of the -u option of
1968 the diff command, as RunTest does. It now checks in the same way as
1969 RunTest, and also checks for the -b option.
1970
1971 19. If an odd number of negated classes containing just a single character
1972 interposed, within parentheses, between a forward reference to a named
1973 subpattern and the definition of the subpattern, compilation crashed with
1974 an internal error, complaining that it could not find the referenced
1975 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1976 [The bug was that it was starting one character too far in when skipping
1977 over the character class, thus treating the ] as data rather than
1978 terminating the class. This meant it could skip too much.]
1979
1980 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1981 /g option in pcretest when the pattern contains \K, which makes it possible
1982 to have an empty string match not at the start, even when the pattern is
1983 anchored. Updated pcretest and pcredemo to use this option.
1984
1985 21. If the maximum number of capturing subpatterns in a recursion was greater
1986 than the maximum at the outer level, the higher number was returned, but
1987 with unset values at the outer level. The correct (outer level) value is
1988 now given.
1989
1990 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1991 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1992 make it do so. The string so far is captured, making this feature
1993 compatible with Perl.
1994
1995 23. The tests have been re-organized, adding tests 11 and 12, to make it
1996 possible to check the Perl 5.10 features against Perl 5.10.
1997
1998 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1999 pattern matches a fixed length string. PCRE did not allow this; now it
2000 does. Neither allows recursion.
2001
2002 25. I finally figured out how to implement a request to provide the minimum
2003 length of subject string that was needed in order to match a given pattern.
2004 (It was back references and recursion that I had previously got hung up
2005 on.) This code has now been added to pcre_study(); it finds a lower bound
2006 to the length of subject needed. It is not necessarily the greatest lower
2007 bound, but using it to avoid searching strings that are too short does give
2008 some useful speed-ups. The value is available to calling programs via
2009 pcre_fullinfo().
2010
2011 26. While implementing 25, I discovered to my embarrassment that pcretest had
2012 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2013 study optimizations had never been tested with that matching function.
2014 Oops. What is worse, even when it was passed study data, there was a bug in
2015 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2016 were also very few tests of studied patterns with pcre_dfa_exec().
2017
2018 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2019 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2020 on the other side of the coin, they are no longer allowed to have different
2021 names, because these cannot be distinguished in PCRE, and this has caused
2022 confusion. (This is a difference from Perl.)
2023
2024 28. When duplicate subpattern names are present (necessarily with different
2025 numbers, as required by 27 above), and a test is made by name in a
2026 conditional pattern, either for a subpattern having been matched, or for
2027 recursion in such a pattern, all the associated numbered subpatterns are
2028 tested, and the overall condition is true if the condition is true for any
2029 one of them. This is the way Perl works, and is also more like the way
2030 testing by number works.
2031
2032
2033 Version 7.9 11-Apr-09
2034 ---------------------
2035
2036 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2037 (pcretest), all targets were linked against these libraries. This included
2038 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2039 libraries. This caused unwanted dependencies to be created. This problem
2040 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2041 pcretest is linked with readline.
2042
2043 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2044 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2045 moved outside it again, because FALSE and TRUE are already defined in AIX,
2046 but BOOL is not.
2047
2048 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2049 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2050
2051 4. The pcregrep documentation said spaces were inserted as well as colons (or
2052 hyphens) following file names and line numbers when outputting matching
2053 lines. This is not true; no spaces are inserted. I have also clarified the
2054 wording for the --colour (or --color) option.
2055
2056 5. In pcregrep, when --colour was used with -o, the list of matching strings
2057 was not coloured; this is different to GNU grep, so I have changed it to be
2058 the same.
2059
2060 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2061 each matching line was coloured. Now it goes on to look for further matches
2062 of any of the test patterns, which is the same behaviour as GNU grep.
2063
2064 7. A pattern that could match an empty string could cause pcregrep to loop; it
2065 doesn't make sense to accept an empty string match in pcregrep, so I have
2066 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2067 seems to be how GNU grep behaves. [But see later change 40 for release
2068 8.33.]
2069
2070 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2071 start or after a newline", because the conditional assertion was not being
2072 correctly handled. The rule now is that both the assertion and what follows
2073 in the first alternative must satisfy the test.
2074
2075 9. If auto-callout was enabled in a pattern with a conditional group whose
2076 condition was an assertion, PCRE could crash during matching, both with
2077 pcre_exec() and pcre_dfa_exec().
2078
2079 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2080 used for matching.
2081
2082 11. Unicode property support in character classes was not working for
2083 characters (bytes) greater than 127 when not in UTF-8 mode.
2084
2085 12. Added the -M command line option to pcretest.
2086
2087 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2088
2089 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2090
2091 16. Added comments and documentation about mis-use of no_arg in the C++
2092 wrapper.
2093
2094 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2095 from Martin Jerabek that uses macro names for all relevant character and
2096 string constants.
2097
2098 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2099 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2100 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2101 these, but not everybody uses configure.
2102
2103 19. A conditional group that had only one branch was not being correctly
2104 recognized as an item that could match an empty string. This meant that an
2105 enclosing group might also not be so recognized, causing infinite looping
2106 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2107 with the subject "ab", where knowledge that the repeated group can match
2108 nothing is needed in order to break the loop.
2109
2110 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2111 exec(), but without supplying a callout function, matching went wrong.
2112
2113 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2114 leak if the size of the offset vector was greater than 30. When the vector
2115 is smaller, the saved offsets during recursion go onto a local stack
2116 vector, but for larger vectors malloc() is used. It was failing to free
2117 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2118 error, in fact).
2119
2120 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2121 heapframe that is used only when UTF-8 support is enabled. This caused no
2122 problem, but was untidy.
2123
2124 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2125 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2126 included within another project.
2127
2128 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2129 slightly modified by me:
2130
2131 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2132 not building pcregrep.
2133
2134 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2135 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2136
2137 25. Forward references, both numeric and by name, in patterns that made use of
2138 duplicate group numbers, could behave incorrectly or give incorrect errors,
2139 because when scanning forward to find the reference group, PCRE was not
2140 taking into account the duplicate group numbers. A pattern such as
2141 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2142
2143 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2144 the feature of a custom pointer more persuasive (as requested by a user).
2145
2146 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2147 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2148 pre-defined.
2149
2150 28. Added support for (*UTF8) at the start of a pattern.
2151
2152 29. Arrange for flags added by the "release type" setting in CMake to be shown
2153 in the configuration summary.
2154
2155
2156 Version 7.8 05-Sep-08
2157 ---------------------
2158
2159 1. Replaced UCP searching code with optimized version as implemented for Ad
2160 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2161 stage table and inline lookup instead of a function, giving speed ups of 2
2162 to 5 times on some simple patterns that I tested. Permission was given to
2163 distribute the MultiStage2.py script that generates the tables (it's not in
2164 the tarball, but is in the Subversion repository).
2165
2166 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2167 scripts.
2168
2169 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2170 a group with a zero qualifier. The result of the study could be incorrect,
2171 or the function might crash, depending on the pattern.
2172
2173 4. Caseless matching was not working for non-ASCII characters in back
2174 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2175 It now works when Unicode Property Support is available.
2176
2177 5. In pcretest, an escape such as \x{de} in the data was always generating
2178 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2179 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2180 truncation.
2181
2182 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2183
2184 7. Added two (int) casts to pcregrep when printing the difference of two
2185 pointers, in case they are 64-bit values.
2186
2187 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2188 test 2 if it fails.
2189
2190 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2191 and a #define of that name to empty if it is not externally set. This is to
2192 allow users of MSVC to set it if necessary.
2193
2194 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2195 the convenience functions in the pcre_get.c source file.
2196
2197 11. An option change at the start of a pattern that had top-level alternatives
2198 could cause overwriting and/or a crash. This command provoked a crash in
2199 some environments:
2200
2201 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2202
2203 This potential security problem was recorded as CVE-2008-2371.
2204
2205 12. For a pattern where the match had to start at the beginning or immediately
2206 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2207 pcre_dfa_exec() could read past the end of the passed subject if there was
2208 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2209 pcretest so that it places the subject at the end of its malloc-ed buffer.
2210
2211 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2212 exec() might read past the end of the data buffer in UTF-8 mode.
2213
2214 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2215 the data contained the byte 0x85 as part of a UTF-8 character within its
2216 first line. This applied both to normal and DFA matching.
2217
2218 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2219 /^[^d]*?$/8 failed to match "abc".
2220
2221 16. Added a missing copyright notice to pcrecpp_internal.h.
2222
2223 17. Make it more clear in the documentation that values returned from
2224 pcre_exec() in ovector are byte offsets, not character counts.
2225
2226 18. Tidied a few places to stop certain compilers from issuing warnings.
2227
2228 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2229 supplied by Stefan Weber. I made a further small update for 7.8 because
2230 there is a change of source arrangements: the pcre_searchfuncs.c module is
2231 replaced by pcre_ucd.c.
2232
2233
2234 Version 7.7 07-May-08
2235 ---------------------
2236
2237 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2238 a string to a long long, pretend we don't even have a long long." This is
2239 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2240
2241 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2242 pre-7.6 versions, which defined a global no_arg variable instead of putting
2243 it in the RE class. (See also #8 below.)
2244
2245 3. Remove a line of dead code, identified by coverity and reported by Nuno
2246 Lopes.
2247
2248 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2249
2250 (1) The include/exclude patterns were being applied to the whole pathnames
2251 of files, instead of just to the final components.
2252
2253 (2) If there was more than one level of directory, the subdirectories were
2254 skipped unless they satisfied the include/exclude conditions. This is
2255 inconsistent with GNU grep (and could even be seen as contrary to the
2256 pcregrep specification - which I improved to make it absolutely clear).
2257 The action now is always to scan all levels of directory, and just
2258 apply the include/exclude patterns to regular files.
2259
2260 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2261 --exclude_dir in the tests to avoid scanning .svn directories.
2262
2263 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2264 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2265 doesn't support NULs in patterns.
2266
2267 7. Added some missing "const"s to declarations of static tables in
2268 pcre_compile.c and pcre_dfa_exec.c.
2269
2270 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2271 caused by fix #2 above. (Subsequently also a second patch to fix the
2272 first patch. And a third patch - this was a messy problem.)
2273
2274 9. Applied Craig's patch to remove the use of push_back().
2275
2276 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2277 matching function regexec().
2278
2279 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2280 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2281 references. PCRE supports relative numbers with this syntax (I don't think
2282 Oniguruma does).
2283
2284 12. Previously, a group with a zero repeat such as (...){0} was completely
2285 omitted from the compiled regex. However, this means that if the group
2286 was called as a subroutine from elsewhere in the pattern, things went wrong
2287 (an internal error was given). Such groups are now left in the compiled
2288 pattern, with a new opcode that causes them to be skipped at execution
2289 time.
2290
2291 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2292 to the way PCRE behaves:
2293
2294 (a) A lone ] character is dis-allowed (Perl treats it as data).
2295
2296 (b) A back reference to an unmatched subpattern matches an empty string
2297 (Perl fails the current match path).
2298
2299 (c) A data ] in a character class must be notated as \] because if the
2300 first data character in a class is ], it defines an empty class. (In
2301 Perl it is not possible to have an empty class.) The empty class []
2302 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2303 The negative empty class [^] matches any one character, independently
2304 of the DOTALL setting.
2305
2306 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2307 non-existent subpattern following a character class starting with ']' and
2308 containing () gave an internal compiling error instead of "reference to
2309 non-existent subpattern". Fortunately, when the pattern did exist, the
2310 compiled code was correct. (When scanning forwards to check for the
2311 existence of the subpattern, it was treating the data ']' as terminating
2312 the class, so got the count wrong. When actually compiling, the reference
2313 was subsequently set up correctly.)
2314
2315 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2316 it was being rejected as not supported by pcre_dfa_exec(), even though
2317 other assertions are supported. I have made pcre_dfa_exec() support
2318 (*FAIL).
2319
2320 16. The implementation of 13c above involved the invention of a new opcode,
2321 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2322 cannot be changed at match time, I realized I could make a small
2323 improvement to matching performance by compiling OP_ALLANY instead of
2324 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2325 on the OP_ANY path.
2326
2327 17. Compiling pcretest on Windows with readline support failed without the
2328 following two fixes: (1) Make the unistd.h include conditional on
2329 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2330
2331 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2332 ncurses library to be included for pcretest when ReadLine support is
2333 requested, but also to allow for it to be overridden. This patch came from
2334 Daniel Bergström.
2335
2336 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2337 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2338 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2339 spotting this.
2340
2341
2342 Version 7.6 28-Jan-08
2343 ---------------------
2344
2345 1. A character class containing a very large number of characters with
2346 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2347 overflow.
2348
2349 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2350 HAVE_LONG_LONG is not defined.
2351
2352 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2353 bring them up to date and include new features. This patch includes:
2354
2355 - Fixed PH's badly added libz and libbz2 support.
2356 - Fixed a problem with static linking.
2357 - Added pcredemo. [But later removed - see 7 below.]
2358 - Fixed dftables problem and added an option.
2359 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2360 HAVE_LONG_LONG.
2361 - Added readline support for pcretest.
2362 - Added an listing of the option settings after cmake has run.
2363
2364 4. A user submitted a patch to Makefile that makes it easy to create
2365 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2366 Makefile.am that cause it to include this special target, without
2367 affecting anything else. Note that the same mingw target plus all
2368 the other distribution libraries and programs are now supported
2369 when configuring with CMake (see 6 below) instead of with
2370 Configure/Make.
2371
2372 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2373 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2374 exported in the Windows port". It has not yet been confirmed that the patch
2375 solves the problem, but it does no harm.
2376
2377 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2378 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2379 with CMake, and also correct the comment about stack recursion.
2380
2381 7. Remove the automatic building of pcredemo from the ./configure system and
2382 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2383 of a program that users should build themselves after PCRE is installed, so
2384 building it automatically is not really right. What is more, it gave
2385 trouble in some build environments.
2386
2387 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2388
2389
2390 Version 7.5 10-Jan-08
2391 ---------------------
2392
2393 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2394 values in parens when parsing an RE using the C++ wrapper."
2395
2396 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2397 Characters greater than 255 were excluded from the class instead of being
2398 included.
2399
2400 3. The same bug as (2) above applied to negated POSIX classes such as
2401 [:^space:].
2402
2403 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2404 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2405 I have changed it.
2406
2407 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2408 first named subpattern) and a construct such as (?&a) would reference the
2409 first named subpattern whose name started with "a" (in other words, the
2410 length check was missing). Both these problems are fixed. "Subpattern name
2411 expected" is now given for (?&) (a zero-length name), and this patch also
2412 makes it give the same error for \k'' (previously it complained that that
2413 was a reference to a non-existent subpattern).
2414
2415 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2416 this is right because (?- can be followed by option settings as well as by
2417 digits. I have, however, made the messages clearer.
2418
2419 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2420 than the number used in the conditional) now cause a compile-time error.
2421 This is actually not compatible with Perl, which accepts such patterns, but
2422 treats the conditional as always being FALSE (as PCRE used to), but it
2423 seems to me that giving a diagnostic is better.
2424
2425 8. Change "alphameric" to the more common word "alphanumeric" in comments
2426 and messages.
2427
2428 9. Fix two occurrences of "backslash" in comments that should have been
2429 "backspace".
2430
2431 10. Remove two redundant lines of code that can never be obeyed (their function
2432 was moved elsewhere).
2433
2434 11. The program that makes PCRE's Unicode character property table had a bug
2435 which caused it to generate incorrect table entries for sequences of
2436 characters that have the same character type, but are in different scripts.
2437 It amalgamated them into a single range, with the script of the first of
2438 them. In other words, some characters were in the wrong script. There were
2439 thirteen such cases, affecting characters in the following ranges:
2440
2441 U+002b0 - U+002c1
2442 U+0060c - U+0060d
2443 U+0061e - U+00612
2444 U+0064b - U+0065e
2445 U+0074d - U+0076d
2446 U+01800 - U+01805
2447 U+01d00 - U+01d77
2448 U+01d9b - U+01dbf
2449 U+0200b - U+0200f
2450 U+030fc - U+030fe
2451 U+03260 - U+0327f
2452 U+0fb46 - U+0fbb1
2453 U+10450 - U+1049d
2454
2455 12. The -o option (show only the matching part of a line) for pcregrep was not
2456 compatible with GNU grep in that, if there was more than one match in a
2457 line, it showed only the first of them. It now behaves in the same way as
2458 GNU grep.
2459
2460 13. If the -o and -v options were combined for pcregrep, it printed a blank
2461 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2462 does the same. The return code can be used to tell if there were any
2463 non-matching lines.
2464
2465 14. Added --file-offsets and --line-offsets to pcregrep.
2466
2467 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2468 infinitely looping recursion. The bug was that positive lookaheads were not
2469 being skipped when checking for a possible empty match (negative lookaheads
2470 and both kinds of lookbehind were skipped).
2471
2472 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2473 inclusion of <windows.h> to before rather than after the definition of
2474 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2475
2476 17. Specifying a possessive quantifier with a specific limit for a Unicode
2477 character property caused pcre_compile() to compile bad code, which led at
2478 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2479 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2480 caused the error; without that there was no problem.
2481
2482 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2483
2484 19. Added --enable-pcretest-libreadline.
2485
2486 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2487 RE::GlobalReplace(). As a result, the number of replacements returned was
2488 double what it should be. I removed one of the increments, but Craig sent a
2489 later patch that removed the other one (the right fix) and added unit tests
2490 that check the return values (which was not done before).
2491
2492 21. Several CMake things:
2493
2494 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2495 the names libpcre and libpcreposix, not just pcre and pcreposix.
2496
2497 (2) The above change means that pcretest and pcregrep are now correctly
2498 linked with the newly-built libraries, not previously installed ones.
2499
2500 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2501
2502 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2503 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2504 UTF-8 newline character). The key issue is that the pattern starts .*;
2505 this means that the match must be either at the beginning, or after a
2506 newline. The bug was in the code for advancing after a failed match and
2507 checking that the new position followed a newline. It was not taking
2508 account of UTF-8 characters correctly.
2509
2510 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2511 character classes. PCRE was not treating the sequence [:...:] as a
2512 character class unless the ... were all letters. Perl, however, seems to
2513 allow any characters between [: and :], though of course it rejects as
2514 unknown any "names" that contain non-letters, because all the known class
2515 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2516 for example, whereas PCRE did not - it did not recognize a POSIX character
2517 class. This seemed a bit dangerous, so the code has been changed to be
2518 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2519 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2520 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2521 Perl does, and where it didn't before.
2522
2523 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2524 Windows environments %n is disabled by default.
2525
2526
2527 Version 7.4 21-Sep-07
2528 ---------------------
2529
2530 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2531 means that a class such as [\s] counted as "explicit reference to CR or
2532 LF". That isn't really right - the whole point of the change was to try to
2533 help when there was an actual mention of one of the two characters. So now
2534 the change happens only if \r or \n (or a literal CR or LF) character is
2535 encountered.
2536
2537 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2538 of both had grown to the point where there were only 3 bits left.
2539 Fortunately, there was spare space in the data structure, and so I have
2540 moved the internal flags into a new 16-bit field to free up more option
2541 bits.
2542
2543 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2544 but did not set the internal JCHANGED flag - either of these is enough to
2545 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2546 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2547 start sets both bits.
2548
2549 4. Added options (at build time, compile time, exec time) to change \R from
2550 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2551
2552 5. doc/pcresyntax.html was missing from the distribution.
2553
2554 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2555 compatibility, even though it is no longer used.
2556
2557 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2558 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2559 windows.h file is present (where different names are used). [This was
2560 reversed later after testing - see 16 below.]
2561
2562 8. Changed all #include <config.h> to #include "config.h". There were also
2563 some further <pcre.h> cases that I changed to "pcre.h".
2564
2565 9. When pcregrep was used with the --colour option, it missed the line ending
2566 sequence off the lines that it output.
2567
2568 10. It was pointed out to me that arrays of string pointers cause lots of
2569 relocations when a shared library is dynamically loaded. A technique of
2570 using a single long string with a table of offsets can drastically reduce
2571 these. I have refactored PCRE in four places to do this. The result is
2572 dramatic:
2573
2574 Originally: 290
2575 After changing UCP table: 187
2576 After changing error message table: 43
2577 After changing table of "verbs" 36
2578 After changing table of Posix names 22
2579
2580 Thanks to the folks working on Gregex for glib for this insight.
2581
2582 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2583 unicode-properties was also set.
2584
2585 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2586
2587 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2588 checked only for CRLF.
2589
2590 14. Added casts to pcretest.c to avoid compiler warnings.
2591
2592 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2593
2594 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2595 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2596 entirely. This removes changes made in 7 above.
2597
2598 17. The CMake files have been updated, and there is now more information about
2599 building with CMake in the NON-UNIX-USE document.
2600
2601
2602 Version 7.3 28-Aug-07
2603 ---------------------
2604
2605 1. In the rejigging of the build system that eventually resulted in 7.1, the
2606 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2607 brackets there is not right, since it causes compilers to look for an
2608 installed pcre.h, not the version that is in the source that is being
2609 compiled (which of course may be different). I have changed it back to:
2610
2611 #include "pcre.h"
2612
2613 I have a vague recollection that the change was concerned with compiling in
2614 different directories, but in the new build system, that is taken care of
2615 by the VPATH setting the Makefile.
2616
2617 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2618 when the subject happened to end in the byte 0x85 (e.g. if the last
2619 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2620 characters but of course it shouldn't be taken as a newline when it is part
2621 of another character. The bug was that, for an unlimited repeat of . in
2622 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2623 characters when looking for a newline.
2624
2625 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2626
2627 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2628 in debug output.
2629
2630 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2631 long printing in the pcrecpp unittest when running under MinGW.
2632
2633 6. ESC_K was left out of the EBCDIC table.
2634
2635 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2636 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2637 limit also applies to "virtual nesting" when a pattern is recursive, and in
2638 this case 1000 isn't so big. I have been able to remove this limit at the
2639 expense of backing off one optimization in certain circumstances. Normally,
2640 when pcre_exec() would call its internal match() function recursively and
2641 immediately return the result unconditionally, it uses a "tail recursion"
2642 feature to save stack. However, when a subpattern that can match an empty
2643 string has an unlimited repetition quantifier, it no longer makes this
2644 optimization. That gives it a stack frame in which to save the data for
2645 checking that an empty string has been matched. Previously this was taken
2646 from the 1000-entry workspace that had been reserved. So now there is no
2647 explicit limit, but more stack is used.
2648
2649 8. Applied Daniel's patches to solve problems with the import/export magic
2650 syntax that is required for Windows, and which was going wrong for the
2651 pcreposix and pcrecpp parts of the library. These were overlooked when this
2652 problem was solved for the main library.
2653
2654 9. There were some crude static tests to avoid integer overflow when computing
2655 the size of patterns that contain repeated groups with explicit upper
2656 limits. As the maximum quantifier is 65535, the maximum group length was
2657 set at 30,000 so that the product of these two numbers did not overflow a
2658 32-bit integer. However, it turns out that people want to use groups that
2659 are longer than 30,000 bytes (though not repeat them that many times).
2660 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2661 made it possible to implement the integer overflow checks in a much more
2662 dynamic way, which I have now done. The artificial limitation on group
2663 length has been removed - we now have only the limit on the total length of
2664 the compiled pattern, which depends on the LINK_SIZE setting.
2665
2666 10. Fixed a bug in the documentation for get/copy named substring when
2667 duplicate names are permitted. If none of the named substrings are set, the
2668 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2669 empty string.
2670
2671 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2672 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2673 because the ] is interpreted as the first data character and the
2674 terminating ] is not found. PCRE has been made compatible with Perl in this
2675 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2676 cause memory overwriting.
2677
2678 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2679 string has been matched (to stop an infinite loop). It was not recognizing
2680 a conditional subpattern that could match an empty string if that
2681 subpattern was within another subpattern. For example, it looped when
2682 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2683 condition was not nested. This bug has been fixed.
2684
2685 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2686 past the start of the subject in the presence of bytes with the top bit
2687 set, for example "\x8aBCD".
2688
2689 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2690 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2691
2692 14. Optimized (?!) to (*FAIL).
2693
2694 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2695 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2696 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2697 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2698 does: it's just the validity check that is more restrictive.
2699
2700 16. Inserted checks for integer overflows during escape sequence (backslash)
2701 processing, and also fixed erroneous offset values for syntax errors during
2702 backslash processing.
2703
2704 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2705 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2706
2707 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2708 caused an overrun.
2709
2710 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2711 something other than just ASCII characters) inside a group that had an
2712 unlimited repeat caused a loop at compile time (while checking to see
2713 whether the group could match an empty string).
2714
2715 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2716 [\P{Any}] did so. (Error in the code for printing property names.)
2717
2718 21. An orphan \E inside a character class could cause a crash.
2719
2720 22. A repeated capturing bracket such as (A)? could cause a wild memory
2721 reference during compilation.
2722
2723 23. There are several functions in pcre_compile() that scan along a compiled
2724 expression for various reasons (e.g. to see if it's fixed length for look
2725 behind). There were bugs in these functions when a repeated \p or \P was
2726 present in the pattern. These operators have additional parameters compared
2727 with \d, etc, and these were not being taken into account when moving along
2728 the compiled data. Specifically:
2729
2730 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2731 length.
2732
2733 (b) An item such as \pL+ within a repeated group could cause crashes or
2734 loops.
2735
2736 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2737 "reference to non-existent subpattern" error.
2738
2739 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2740
2741 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2742 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2743
2744 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2745
2746 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2747 character were causing crashes (broken optimization).
2748
2749 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2750 \p or \P) caused a compile-time loop.
2751
2752 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2753 break. For example, the unstudied pattern [\r\n]A does not match the string
2754 "\r\nA" because change 7.0/46 below moves the current point on by two
2755 characters after failing to match at the start. However, the pattern \nA
2756 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2757 the same is true. There doesn't seem any very clean way out of this, but
2758 what I have chosen to do makes the common cases work: PCRE now takes note
2759 of whether there can be an explicit match for \r or \n anywhere in the
2760 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2761 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2762 pattern has explicit CR or LF references.
2763
2764 29. Added (*CR) etc for changing newline setting at start of pattern.
2765
2766
2767 Version 7.2 19-Jun-07
2768 ---------------------
2769
2770 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2771 which is apparently normally available under Windows.
2772
2773 2. Re-jig the pcregrep tests with different newline settings in an attempt
2774 to make them independent of the local environment's newline setting.
2775
2776 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2777
2778 4. Some of the "internals" tests were previously cut out when the link size
2779 was not 2, because the output contained actual offsets. The recent new
2780 "Z" feature of pcretest means that these can be cut out, making the tests
2781 usable with all link sizes.
2782
2783 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2784 stack recursion. This gives a massive performance boost under BSD, but just
2785 a small improvement under Linux. However, it saves one field in the frame
2786 in all cases.
2787
2788 6. Added more features from the forthcoming Perl 5.10:
2789
2790 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2791 recursion call. It refers to the nth most recently opened parentheses.
2792
2793 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2794 to be opened parentheses.
2795
2796 (c) Conditions that refer to capturing parentheses can be specified
2797 relatively, for example, (?(-2)... or (?(+3)...
2798
2799 (d) \K resets the start of the current match so that everything before
2800 is not part of it.
2801
2802 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2803
2804 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2805 reference syntax.
2806
2807 (g) (?| introduces a group in which the numbering of parentheses in each
2808 alternative starts with the same number.
2809
2810 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2811
2812 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2813 PCRE_INFO_JCHANGED.
2814
2815 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2816 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2817 for detecting groups that can match an empty string.
2818
2819 9. A pattern with a very large number of alternatives (more than several
2820 hundred) was running out of internal workspace during the pre-compile
2821 phase, where pcre_compile() figures out how much memory will be needed. A
2822 bit of new cunning has reduced the workspace needed for groups with
2823 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2824 workspace instead of running out of the 4096 that are available.
2825
2826 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2827
2828 11. Applied patch from Google to remove an optimization that didn't quite work.
2829 The report of the bug said:
2830
2831 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2832 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2833 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2834
2835 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2836 it matched the wrong number of bytes.
2837
2838
2839 Version 7.1 24-Apr-07
2840 ---------------------
2841
2842 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2843 that is more "standard", making use of automake and other Autotools. There
2844 is some re-arrangement of the files and adjustment of comments consequent
2845 on this.
2846
2847 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2848 for recursive directory scanning broke on some systems because the files
2849 are not scanned in any specific order and on different systems the order
2850 was different. A call to "sort" has been inserted into RunGrepTest for the
2851 approprate test as a short-term fix. In the longer term there may be an
2852 alternative.
2853
2854 3. I had an email from Eric Raymond about problems translating some of PCRE's
2855 man pages to HTML (despite the fact that I distribute HTML pages, some
2856 people do their own conversions for various reasons). The problems
2857 concerned the use of low-level troff macros .br and .in. I have therefore
2858 removed all such uses from the man pages (some were redundant, some could
2859 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2860 HTML has been updated to handle .nf/.fi and to complain if it encounters
2861 .br or .in.
2862
2863 4. Updated comments in configure.ac that get placed in config.h.in and also
2864 arranged for config.h to be included in the distribution, with the name
2865 config.h.generic, for the benefit of those who have to compile without
2866 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2867
2868 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2869 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2870 makevp.bat for the current PCRE, using the additional files
2871 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2872
2873 6. A Windows user reported a minor discrepancy with test 2, which turned out
2874 to be caused by a trailing space on an input line that had got lost in his
2875 copy. The trailing space was an accident, so I've just removed it.
2876
2877 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2878 that is needed.
2879
2880 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2881 as "const" (a) because they are and (b) because it helps the PHP
2882 maintainers who have recently made a script to detect big data structures
2883 in the php code that should be moved to the .rodata section. I remembered
2884 to update Builducptable as well, so it won't revert if ucptable.h is ever
2885 re-created.
2886
2887 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2888 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2889 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2890 support is not required. This saves 1.5-2K of code, which is important in
2891 some applications.
2892
2893 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2894 so as not to refer to the tables, even though these functions will never be
2895 called when UTF-8 support is disabled. Otherwise there are problems with a
2896 shared library.
2897
2898 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2899
2900 (a) It was defining its arguments as char * instead of void *.
2901
2902 (b) It was assuming that all moves were upwards in memory; this was true
2903 a long time ago when I wrote it, but is no longer the case.
2904
2905 The emulated memove() is provided for those environments that have neither
2906 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2907 is clearly not the case, as these two bugs were recently reported.
2908
2909 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2910 and Detrail to create the HTML documentation, the .txt form of the man
2911 pages, and it removes trailing spaces from listed files. It also creates
2912 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2913 case, it wraps all the #defines with #ifndefs. This script should be run
2914 before "make dist".
2915
2916 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2917 with Unicode property support.
2918
2919 (a) For a maximizing quantifier, if the two different cases of the
2920 character were of different lengths in their UTF-8 codings (there are
2921 some cases like this - I found 11), and the matching function had to
2922 back up over a mixture of the two cases, it incorrectly assumed they
2923 were both the same length.
2924
2925 (b) When PCRE was configured to use the heap rather than the stack for
2926 recursion during matching, it was not correctly preserving the data for
2927 the other case of a UTF-8 character when checking ahead for a match
2928 while processing a minimizing repeat. If the check also involved
2929 matching a wide character, but failed, corruption could cause an
2930 erroneous result when trying to check for a repeat of the original
2931 character.
2932
2933 13. Some tidying changes to the testing mechanism:
2934
2935 (a) The RunTest script now detects the internal link size and whether there
2936 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2937 values substituted by "configure". (The RunGrepTest script already did
2938 this for UTF-8.) The configure.ac script no longer substitutes the
2939 relevant variables.
2940
2941 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2942 with length and offset values. This means that the output is different
2943 for different internal link sizes. Test 2 is skipped for link sizes
2944 other than 2 because of this, bypassing the problem. Unfortunately,
2945 there was also a test in test 3 (the locale tests) that used /B and
2946 failed for link sizes other than 2. Rather than cut the whole test out,
2947 I have added a new /Z option to pcretest that replaces the length and
2948 offset values with spaces. This is now used to make test 3 independent
2949 of link size. (Test 2 will be tidied up later.)
2950
2951 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2952 segmentation fault instead of returning the appropriate error message.
2953
2954 15. In multiline mode when the newline sequence was set to "any", the pattern
2955 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2956 This doesn't seem right; it now treats the CRLF combination as the line
2957 ending, and so does not match in that case. It's only a pattern such as ^$
2958 that would hit this one: something like ^ABC$ would have failed after \r
2959 and then tried again after \r\n.
2960
2961 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2962 in an attempt to make files that differ only in their line terminators
2963 compare equal. This works on Linux.
2964
2965 17. Under certain error circumstances pcregrep might try to free random memory
2966 as it exited. This is now fixed, thanks to valgrind.
2967
2968 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2969 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2970 was because its rules for how to advance for /g after matching an empty
2971 string at the end of a line did not allow for this case. They now check for
2972 it specially.
2973
2974 20. pcretest is supposed to handle patterns and data of any length, by
2975 extending its buffers when necessary. It was getting this wrong when the
2976 buffer for a data line had to be extended.
2977
2978 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2979 CRLF as a newline sequence.
2980
2981 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2982 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2983 I have nevertheless tidied it up.
2984
2985 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2986
2987 24. Added a man page for pcre-config.
2988
2989
2990 Version 7.0 19-Dec-06
2991 ---------------------
2992
2993 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2994 moving to gcc 4.1.1.
2995
2996 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2997 sys/time.h, which is documented as needed for this function. It doesn't
2998 seem to matter on Linux, but it showed up on some releases of OS X.
2999
3000 3. It seems that there are systems where bytes whose values are greater than
3001 127 match isprint() in the "C" locale. The "C" locale should be the
3002 default when a C program starts up. In most systems, only ASCII printing
3003 characters match isprint(). This difference caused the output from pcretest
3004 to vary, making some of the tests fail. I have changed pcretest so that:
3005
3006 (a) When it is outputting text in the compiled version of a pattern, bytes
3007 other than 32-126 are always shown as hex escapes.
3008
3009 (b) When it is outputting text that is a matched part of a subject string,
3010 it does the same, unless a different locale has been set for the match
3011 (using the /L modifier). In this case, it uses isprint() to decide.
3012
3013 4. Fixed a major bug that caused incorrect computation of the amount of memory
3014 required for a compiled pattern when options that changed within the
3015 pattern affected the logic of the preliminary scan that determines the
3016 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3017 that the computed length was too small. The symptoms of this bug were
3018 either the PCRE error "internal error: code overflow" from pcre_compile(),
3019 or a glibc crash with a message such as "pcretest: free(): invalid next
3020 size (fast)". Examples of patterns that provoked this bug (shown in
3021 pcretest format) are:
3022
3023 /(?-x: )/x
3024 /(?x)(?-x: \s*#\s*)/
3025 /((?i)[\x{c0}])/8
3026 /(?i:[\x{c0}])/8
3027
3028 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3029 is now done differently.
3030
3031 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3032 wrapper classes; (b) implement a new function in the C++ scanner that is
3033 more efficient than the old way of doing things because it avoids levels of
3034 recursion in the regex matching; (c) add a paragraph to the documentation
3035 for the FullMatch() function.
3036
3037 6. The escape sequence \n was being treated as whatever was defined as
3038 "newline". Not only was this contrary to the documentation, which states
3039 that \n is character 10 (hex 0A), but it also went horribly wrong when
3040 "newline" was defined as CRLF. This has been fixed.
3041
3042 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3043 was being set to -1 for the "end of line" case (supposedly a value that no
3044 character can have). Though this value is never used (the check for end of
3045 line is "zero bytes in current character"), it caused compiler complaints.
3046 I've changed it to 0xffffffff.
3047
3048 8. In pcre_version.c, the version string was being built by a sequence of
3049 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3050 string (as it is for production releases) called a macro with an empty
3051 argument. The C standard says the result of this is undefined. The gcc
3052 compiler treats it as an empty string (which was what was wanted) but it is
3053 reported that Visual C gives an error. The source has been hacked around to
3054 avoid this problem.
3055
3056 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3057 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3058 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3059 of them did).
3060
3061 10. Originally, pcretest opened its input and output without "b"; then I was
3062 told that "b" was needed in some environments, so it was added for release
3063 5.0 to both the input and output. (It makes no difference on Unix-like
3064 systems.) Later I was told that it is wrong for the input on Windows. I've
3065 now abstracted the modes into two macros, to make it easier to fiddle with
3066 them, and removed "b" from the input mode under Windows.
3067
3068 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3069
3070 12. Added -help and --help to pcretest as an official way of being reminded
3071 of the options.
3072
3073 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3074 and pcrecpp.cc because they annoy compilers at high warning levels.
3075
3076 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3077
3078 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3079 scripts are not C programs :-) and which was not noticed because it works
3080 on Linux.
3081
3082 16. pcretest is supposed to handle any length of pattern and data line (as one
3083 line or as a continued sequence of lines) by extending its input buffer if
3084 necessary. This feature was broken for very long pattern lines, leading to
3085 a string of junk being passed to pcre_compile() if the pattern was longer
3086 than about 50K.
3087
3088 17. I have done a major re-factoring of the way pcre_compile() computes the
3089 amount of memory needed for a compiled pattern. Previously, there was code
3090 that made a preliminary scan of the pattern in order to do this. That was
3091 OK when PCRE was new, but as the facilities have expanded, it has become
3092 harder and harder to keep it in step with the real compile phase, and there
3093 have been a number of bugs (see for example, 4 above). I have now found a
3094 cunning way of running the real compile function in a "fake" mode that
3095 enables it to compute how much memory it would need, while actually only
3096 ever using a few hundred bytes of working memory and without too many
3097 tests of the mode. This should make future maintenance and development
3098 easier. A side effect of this work is that the limit of 200 on the nesting
3099 depth of parentheses has been removed (though this was never a serious
3100 limitation, I suspect). However, there is a downside: pcre_compile() now
3101 runs more slowly than before (30% or more, depending on the pattern). I
3102 hope this isn't a big issue. There is no effect on runtime performance.
3103
3104 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3105 newline (only possible for the last line of a file) and it was a
3106 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3107
3108 19. Added additional timing features to pcretest. (1) The -tm option now times
3109 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3110 separate command line item, by a number that specifies the number of
3111 repeats to use when timing. The default is 50000; this gives better
3112 precision, but takes uncomfortably long for very large patterns.
3113
3114 20. Extended pcre_study() to be more clever in cases where a branch of a
3115 subpattern has no definite first character. For example, (a*|b*)[cd] would
3116 previously give no result from pcre_study(). Now it recognizes that the
3117 first character must be a, b, c, or d.
3118
3119 21. There was an incorrect error "recursive call could loop indefinitely" if
3120 a subpattern (or the entire pattern) that was being tested for matching an
3121 empty string contained only one non-empty item after a nested subpattern.
3122 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3123 incorrectly, because the \d was being skipped in the check.
3124
3125 22. The pcretest program now has a new pattern option /B and a command line
3126 option -b, which is equivalent to adding /B to every pattern. This causes
3127 it to show the compiled bytecode, without the additional information that
3128 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3129 is the same as /B/I).
3130
3131 23. A new optimization is now able automatically to treat some sequences such
3132 as a*b as a*+b. More specifically, if something simple (such as a character
3133 or a simple class like \d) has an unlimited quantifier, and is followed by
3134 something that cannot possibly match the quantified thing, the quantifier
3135 is automatically "possessified".
3136
3137 24. A recursive reference to a subpattern whose number was greater than 39
3138 went wrong under certain circumstances in UTF-8 mode. This bug could also
3139 have affected the operation of pcre_study().
3140
3141 25. Realized that a little bit of performance could be had by replacing
3142 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3143
3144 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3145
3146 27. Possessive quantifiers such as a++ were previously implemented by turning
3147 them into atomic groups such as ($>a+). Now they have their own opcodes,
3148 which improves performance. This includes the automatically created ones
3149 from 23 above.
3150
3151 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3152 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3153 the first matched character to be a colon. This applied both to named and
3154 numbered groups.
3155
3156 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3157
3158 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3159 building PCRE on Windows easier, so I have included it in the distribution.
3160
3161 31. There is now a check in pcretest against a ridiculously large number being
3162 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3163 loop, the loop is abandoned.
3164
3165 32. Forward references to subpatterns in conditions such as (?(2)...) where
3166 subpattern 2 is defined later cause pcre_compile() to search forwards in
3167 the pattern for the relevant set of parentheses. This search went wrong
3168 when there were unescaped parentheses in a character class, parentheses
3169 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3170
3171 33. "Subroutine" calls and backreferences were previously restricted to
3172 referencing subpatterns earlier in the regex. This restriction has now
3173 been removed.
3174
3175 34. Added a number of extra features that are going to be in Perl 5.10. On the
3176 whole, these are just syntactic alternatives for features that PCRE had
3177 previously implemented using the Python syntax or my own invention. The
3178 other formats are all retained for compatibility.
3179
3180 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3181 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3182 also .NET compatible.
3183
3184 (b) A recursion or subroutine call to a named group can now be defined as
3185 (?&name) as well as (?P>name).
3186
3187 (c) A backreference to a named group can now be defined as \k<name> or
3188 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3189 5.10, are also .NET compatible.
3190
3191 (d) A conditional reference to a named group can now use the syntax
3192 (?(<name>) or (?('name') as well as (?(name).
3193
3194 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3195 groups (named and numbered) that are never evaluated inline, but can be
3196 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3197 is always false. There may be only one alternative in such a group.
3198
3199 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3200 as the simple (?(R). The condition is true only if the most recent
3201 recursion is that of the given number or name. It does not search out
3202 through the entire recursion stack.
3203
3204 (g) The escape \gN or \g{N} has been added, where N is a positive or
3205 negative number, specifying an absolute or relative reference.
3206
3207 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3208 some "unreachable code" warnings.
3209
3210 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3211 things, this adds five new scripts.
3212
3213 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3214 There were also incompatibilities regarding the handling of \Q..\E inside
3215 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3216 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3217
3218 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3219 matches an empty string, and forcibly breaks the loop. There were bugs in
3220 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3221 against aaaa the result was just "a" rather than "aaaa", for example. Two
3222 separate and independent bugs (that affected different cases) have been
3223 fixed.
3224
3225 39. Refactored the code to abolish the use of different opcodes for small
3226 capturing bracket numbers. This is a tidy that I avoided doing when I
3227 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3228 The new approach is not only tidier, it makes it possible to reduce the
3229 memory needed to fix the previous bug (38).
3230
3231 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3232 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3233 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3234 mode.
3235
3236 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3237 report.
3238
3239 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3240 copy construction and assignment in the C++ wrapper.
3241
3242 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3243 couple of bugs that could have given wrong results in the "--newline=crlf"
3244 case.
3245
3246 44. Added a number of casts and did some reorganization of signed/unsigned int
3247 variables following suggestions from Dair Grant. Also renamed the variable
3248 "this" as "item" because it is a C++ keyword.
3249
3250 45. Arranged for dftables to add
3251
3252 #include "pcre_internal.h"
3253
3254 to pcre_chartables.c because without it, gcc 4.x may remove the array
3255 definition from the final binary if PCRE is built into a static library and
3256 dead code stripping is activated.
3257
3258 46. For an unanchored pattern, if a match attempt fails at the start of a
3259 newline sequence, and the newline setting is CRLF or ANY, and the next two
3260 characters are CRLF, advance by two characters instead of one.
3261
3262
3263 Version 6.7 04-Jul-06
3264 ---------------------
3265
3266 1. In order to handle tests when input lines are enormously long, pcretest has
3267 been re-factored so that it automatically extends its buffers when
3268 necessary. The code is crude, but this _is_ just a test program. The
3269 default size has been increased from 32K to 50K.
3270
3271 2. The code in pcre_study() was using the value of the re argument before
3272 testing it for NULL. (Of course, in any sensible call of the function, it
3273 won't be NULL.)
3274
3275 3. The memmove() emulation function in pcre_internal.h, which is used on
3276 systems that lack both memmove() and bcopy() - that is, hardly ever -
3277 was missing a "static" storage class specifier.
3278
3279 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3280 containing an extended class (one that cannot be represented by a bitmap
3281 because it contains high-valued characters or Unicode property items, e.g.
3282 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3283 pattern, but PCRE should not loop if you do not (it no longer does).
3284 [Detail: two cases were found: (a) a repeated subpattern containing an
3285 extended class; (b) a recursive reference to a subpattern that followed a
3286 previous extended class. It wasn't skipping over the extended class
3287 correctly when UTF-8 mode was not set.]
3288
3289 5. A negated single-character class was not being recognized as fixed-length
3290 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3291 compile error "lookbehind assertion is not fixed length".
3292
3293 6. The RunPerlTest auxiliary script was showing an unexpected difference
3294 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3295 write a Perl script that can interpret lines of an input file either as
3296 byte characters or as UTF-8, which is what "perltest" was being required to
3297 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3298 can't do is switch easily at run time between having the "use utf8;" pragma
3299 or not. In the end, I fudged it by using the RunPerlTest script to insert
3300 "use utf8;" explicitly for the UTF-8 tests.
3301
3302 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3303 the end of the subject string, contrary to the documentation and to what
3304 Perl does. This was true of both matching functions. Now it matches only at
3305 the start of the subject and immediately after *internal* newlines.
3306
3307 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3308 a pointer to an int instead of a pointer to an unsigned long int. This
3309 caused problems on 64-bit systems.
3310
3311 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3312 instance of the 'standard' template library not being so standard".
3313
3314 10. There was no check on the number of named subpatterns nor the maximum
3315 length of a subpattern name. The product of these values is used to compute
3316 the size of the memory block for a compiled pattern. By supplying a very
3317 long subpattern name and a large number of named subpatterns, the size
3318 computation could be caused to overflow. This is now prevented by limiting
3319 the length of names to 32 characters, and the number of named subpatterns
3320 to 10,000.
3321
3322 11. Subpatterns that are repeated with specific counts have to be replicated in
3323 the compiled pattern. The size of memory for this was computed from the
3324 length of the subpattern and the repeat count. The latter is limited to
3325 65535, but there was no limit on the former, meaning that integer overflow
3326 could in principle occur. The compiled length of a repeated subpattern is
3327 now limited to 30,000 bytes in order to prevent this.
3328
3329 12. Added the optional facility to have named substrings with the same name.
3330
3331 13. Added the ability to use a named substring as a condition, using the
3332 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3333 are numbers (not recommended). Forward references are permitted.
3334
3335 14. Added forward references in named backreferences (if you see what I mean).
3336
3337 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3338 pattern could run off the end of the subject. For example, the pattern
3339 "(?s)(.{1,5})"8 did this with the subject "ab".
3340
3341 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3342 PCRE_CASELESS was set when matching characters that were quantified with ?
3343 or *.
3344
3345 17. A character class other than a single negated character that had a minimum
3346 but no maximum quantifier - for example [ab]{6,} - was not handled
3347 correctly by pce_dfa_exec(). It would match only one character.
3348
3349 18. A valid (though odd) pattern that looked like a POSIX character
3350 class but used an invalid character after [ (for example [[,abc,]]) caused
3351 pcre_compile() to give the error "Failed: internal error: code overflow" or
3352 in some cases to crash with a glibc free() error. This could even happen if
3353 the pattern terminated after [[ but there just happened to be a sequence of
3354 letters, a binary zero, and a closing ] in the memory that followed.
3355
3356 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3357 over the years. Originally (before any Unicode support), just the bottom 8
3358 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3359 output from "man perlunicode" includes this:
3360
3361 The regular expression compiler produces polymorphic opcodes. That
3362 is, the pattern adapts to the data and automatically switches to
3363 the Unicode character scheme when presented with Unicode data--or
3364 instead uses a traditional byte scheme when presented with byte
3365 data.
3366
3367 Sadly, a wide octal escape does not cause a switch, and in a string with
3368 no other multibyte characters, these octal escapes are treated as before.
3369 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3370 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3371 Unicode string.
3372
3373 I have not perpetrated such confusion in PCRE. Up till now, it took just
3374 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3375 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3376 translate to the appropriate multibyte character.
3377
3378 29. Applied some refactoring to reduce the number of warnings from Microsoft
3379 and Borland compilers. This has included removing the fudge introduced
3380 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3381 a warning about an unused variable.
3382
3383 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3384 characters since release 4.0, because Perl (from release 5.004) does not.
3385 [Or at least, is documented not to: some releases seem to be in conflict
3386 with the documentation.] However, when a pattern was studied with
3387 pcre_study() and all its branches started with \s, PCRE still included VT
3388 as a possible starting character. Of course, this did no harm; it just
3389 caused an unnecessary match attempt.
3390
3391 22. Removed a now-redundant internal flag bit that recorded the fact that case
3392 dependency changed within the pattern. This was once needed for "required
3393 byte" processing, but is no longer used. This recovers a now-scarce options
3394 bit. Also moved the least significant internal flag bit to the most-
3395 significant bit of the word, which was not previously used (hangover from
3396 the days when it was an int rather than a uint) to free up another bit for
3397 the future.
3398
3399 23. Added support for CRLF line endings as well as CR and LF. As well as the
3400 default being selectable at build time, it can now be changed at runtime
3401 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3402 specify that it is scanning data with non-default line endings.
3403
3404 24. Changed the definition of CXXLINK to make it agree with the definition of
3405 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3406
3407 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3408 recursions. This makes a big different to stack usage for some patterns.
3409
3410 26. If a subpattern containing a named recursion or subroutine reference such
3411 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3412 the space required for the compiled pattern went wrong and gave too small a
3413 value. Depending on the environment, this could lead to "Failed: internal
3414 error: code overflow at offset 49" or "glibc detected double free or
3415 corruption" errors.
3416
3417 27. Applied patches from Google (a) to support the new newline modes and (b) to
3418 advance over multibyte UTF-8 characters in GlobalReplace.
3419
3420 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3421 difference for some implementation of PCRE in some Windows version.
3422
3423 29. Added some extra testing facilities to pcretest:
3424
3425 \q<number> in a data line sets the "match limit" value
3426 \Q<number> in a data line sets the "match recursion limt" value
3427 -S <number> sets the stack size, where <number> is in megabytes
3428
3429 The -S option isn't available for Windows.
3430
3431
3432 Version 6.6 06-Feb-06
3433 ---------------------
3434
3435 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3436 in pcreposix.h. I have copied the definition from pcre.h.
3437
3438 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3439 because pcre.h is no longer a built file.
3440
3441 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3442 not normally included in the compiled code.
3443
3444
3445 Version 6.5 01-Feb-06
3446 ---------------------
3447
3448 1. When using the partial match feature with pcre_dfa_exec(), it was not
3449 anchoring the second and subsequent partial matches at the new starting
3450 point. This could lead to incorrect results. For example, with the pattern
3451 /1234/, partially matching against "123" and then "a4" gave a match.
3452
3453 2. Changes to pcregrep:
3454
3455 (a) All non-match returns from pcre_exec() were being treated as failures
3456 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3457 error message is output. Some extra information is given for the
3458 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3459 probably the only errors that are likely to be caused by users (by
3460 specifying a regex that has nested indefinite repeats, for instance).
3461 If there are more than 20 of these errors, pcregrep is abandoned.
3462
3463 (b) A binary zero was treated as data while matching, but terminated the
3464 output line if it was written out. This has been fixed: binary zeroes
3465 are now no different to any other data bytes.
3466
3467 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3468 used to set a locale for matching. The --locale=xxxx long option has
3469 been added (no short equivalent) to specify a locale explicitly on the
3470 pcregrep command, overriding the environment variables.
3471
3472 (d) When -B was used with -n, some line numbers in the output were one less
3473 than they should have been.
3474
3475 (e) Added the -o (--only-matching) option.
3476
3477 (f) If -A or -C was used with -c (count only), some lines of context were
3478 accidentally printed for the final match.
3479
3480 (g) Added the -H (--with-filename) option.
3481
3482 (h) The combination of options -rh failed to suppress file names for files
3483 that were found from directory arguments.
3484
3485 (i) Added the -D (--devices) and -d (--directories) options.
3486
3487 (j) Added the -F (--fixed-strings) option.
3488
3489 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3490
3491 (l) Added the --colo(u)r option.
3492
3493 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3494 is not present by default.
3495
3496 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3497 items such as (?R) or (?1), when the recursion could match a number of
3498 alternatives. If it matched one of the alternatives, but subsequently,
3499 outside the recursion, there was a failure, the code tried to back up into
3500 the recursion. However, because of the way PCRE is implemented, this is not
3501 possible, and the result was an incorrect result from the match.
3502
3503 In order to prevent this happening, the specification of recursion has
3504 been changed so that all such subpatterns are automatically treated as
3505 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3506
3507 4. I had overlooked the fact that, in some locales, there are characters for
3508 which isalpha() is true but neither isupper() nor islower() are true. In
3509 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3510 and ordfeminine) are like this. This affected the treatment of \w and \W
3511 when they appeared in character classes, but not when they appeared outside
3512 a character class. The bit map for "word" characters is now created
3513 separately from the results of isalnum() instead of just taking it from the
3514 upper, lower, and digit maps. (Plus the underscore character, of course.)
3515
3516 5. The above bug also affected the handling of POSIX character classes such as
3517 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3518 permanent tables. Instead, the bit maps for such a class were previously
3519 created as the appropriate unions of the upper, lower, and digit bitmaps.
3520 Now they are created by subtraction from the [[:word:]] class, which has
3521 its own bitmap.
3522
3523 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3524 It is created by subtracting the vertical space characters (\x09, \x0a,
3525 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3526 subtraction was done in the overall bitmap for a character class, meaning
3527 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3528 be recognized. This bug has been fixed.
3529
3530 7. Patches from the folks at Google:
3531
3532 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3533 real life, but is still worth protecting against".
3534
3535 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3536 regular expressions".
3537
3538 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3539 have it.
3540
3541 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3542 "configure" and the latter not, in order to fix a problem somebody had
3543 with compiling the Arg class on HP-UX.
3544
3545 (e) Improve the error-handling of the C++ wrapper a little bit.
3546
3547 (f) New tests for checking recursion limiting.
3548
3549 8. The pcre_memmove() function, which is used only if the environment does not
3550 have a standard memmove() function (and is therefore rarely compiled),
3551 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3552 returning a result (though PCRE never actually uses the result).
3553
3554 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3555 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3556 returned instead of calling malloc() with an overflowing number that would
3557 most likely cause subsequent chaos.
3558
3559 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3560
3561 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3562 with this option is matched, the nmatch and pmatch options of regexec() are
3563 ignored.
3564
3565 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3566 provided in case anyone wants to the the POSIX interface with UTF-8
3567 strings.
3568
3569 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3570 C++ linking (needed for some HP-UX environments).
3571
3572 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3573 (unused parameter) and in the pcre_printint() function (omitted "default"
3574 switch label when the default is to do nothing).
3575
3576 15. Added some code to make it possible, when PCRE is compiled as a C++
3577 library, to replace subject pointers for pcre_exec() with a smart pointer
3578 class, thus making it possible to process discontinuous strings.
3579
3580 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3581 much the same function. They were added by different people who were trying
3582 to make PCRE easy to compile on non-Unix systems. It has been suggested
3583 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3584 for compiling on Windows systems. I have therefore replaced it with
3585 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3586 defaults to "extern" for C or "extern C" for C++, which works fine on
3587 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3588 SCOPE with something explicit in config.h. In addition:
3589
3590 (a) pcreposix.h still had just "extern" instead of either of these macros;
3591 I have replaced it with PCRE_DATA_SCOPE.
3592
3593 (b) Functions such as _pcre_xclass(), which are internal to the library,
3594 but external in the C sense, all had PCRE_EXPORT in their definitions.
3595 This is apparently wrong for the Windows case, so I have removed it.
3596 (It makes no difference on Unix-like systems.)
3597
3598 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3599 of recursive calls to match(). This is different to MATCH_LIMIT because
3600 that limits the total number of calls to match(), not all of which increase
3601 the depth of recursion. Limiting the recursion depth limits the amount of
3602 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3603 when PCRE is compiled, and changed at run time. A patch from Google adds
3604 this functionality to the C++ interface.
3605
3606 18. Changes to the handling of Unicode character properties:
3607
3608 (a) Updated the table to Unicode 4.1.0.
3609
3610 (b) Recognize characters that are not in the table as "Cn" (undefined).
3611
3612 (c) I revised the way the table is implemented to a much improved format
3613 which includes recognition of ranges. It now supports the ranges that
3614 are defined in UnicodeData.txt, and it also amalgamates other
3615 characters into ranges. This has reduced the number of entries in the
3616 table from around 16,000 to around 3,000, thus reducing its size
3617 considerably. I realized I did not need to use a tree structure after
3618 all - a binary chop search is just as efficient. Having reduced the
3619 number of entries, I extended their size from 6 bytes to 8 bytes to
3620 allow for more data.
3621
3622 (d) Added support for Unicode script names via properties such as \p{Han}.
3623
3624 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3625 matching that character.
3626
3627 20. When matching a repeated Unicode property with a minimum greater than zero,
3628 (for example \pL{2,}), PCRE could look past the end of the subject if it
3629 reached it while seeking the minimum number of characters. This could
3630 happen only if some of the characters were more than one byte long, because
3631 there is a check for at least the minimum number of bytes.
3632
3633 21. Refactored the implementation of \p and \P so as to be more general, to
3634 allow for more different types of property in future. This has changed the
3635 compiled form incompatibly. Anybody with saved compiled patterns that use
3636 \p or \P will have to recompile them.
3637
3638 22. Added "Any" and "L&" to the supported property types.
3639
3640 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3641 but give a compile time error if the value is greater than 0xff.
3642
3643 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3644 accidentally not being installed or uninstalled.
3645
3646 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3647 made were to insert the current release number. This seemed silly, because
3648 it made things harder for people building PCRE on systems that don't run
3649 "configure". I have turned pcre.h into a distributed file, no longer built
3650 by "configure", with the version identification directly included. There is
3651 no longer a pcre.h.in file.
3652
3653 However, this change necessitated a change to the pcre-config script as
3654 well. It is built from pcre-config.in, and one of the substitutions was the
3655 release number. I have updated configure.ac so that ./configure now finds
3656 the release number by grepping pcre.h.
3657
3658 26. Added the ability to run the tests under valgrind.
3659
3660
3661 Version 6.4 05-Sep-05
3662 ---------------------
3663
3664 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3665 "--" to be printed when multiple files were scanned, even when none of the
3666 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3667 consider it to be a bug, and have restored the previous behaviour.
3668
3669 2. A couple of code tidies to get rid of compiler warnings.
3670
3671 3. The pcretest program used to cheat by referring to symbols in the library
3672 whose names begin with _pcre_. These are internal symbols that are not
3673 really supposed to be visible externally, and in some environments it is
3674 possible to suppress them. The cheating is now confined to including
3675 certain files from the library's source, which is a bit cleaner.
3676
3677 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3678 file's purpose clearer.
3679
3680 5. Reorganized pcre_ucp_findchar().
3681
3682
3683 Version 6.3 15-Aug-05
3684 ---------------------
3685
3686 1. The file libpcre.pc.in did not have general read permission in the tarball.
3687
3688 2. There were some problems when building without C++ support:
3689
3690 (a) If C++ support was not built, "make install" and "make test" still
3691 tried to test it.
3692
3693 (b) There were problems when the value of CXX was explicitly set. Some
3694 changes have been made to try to fix these, and ...
3695
3696 (c) --disable-cpp can now be used to explicitly disable C++ support.
3697
3698 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3699 backslash in a target when C++ was disabled. This confuses some
3700 versions of "make", apparently. Using an intermediate variable solves
3701 this. (Same for CPP_LOBJ.)
3702
3703 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3704 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3705 necessary on certain architectures.
3706
3707 4. Added a setting of -export-symbols-regex to the link command to remove
3708 those symbols that are exported in the C sense, but actually are local
3709 within the library, and not documented. Their names all begin with
3710 "_pcre_". This is not a perfect job, because (a) we have to except some
3711 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3712 available (and never for static libraries). I have made a note to try to
3713 find a way round (a) in the future.
3714
3715
3716 Version 6.2 01-Aug-05
3717 ---------------------
3718
3719 1. There was no test for integer overflow of quantifier values. A construction
3720 such as {1111111111111111} would give undefined results. What is worse, if
3721 a minimum quantifier for a parenthesized subpattern overflowed and became
3722 negative, the calculation of the memory size went wrong. This could have
3723 led to memory overwriting.
3724
3725 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3726
3727 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3728 operating environments where this matters.
3729
3730 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3731 PCRE options from within the C++ wrapper.
3732
3733 5. Named capturing subpatterns were not being correctly counted when a pattern
3734 was compiled. This caused two problems: (a) If there were more than 100
3735 such subpatterns, the calculation of the memory needed for the whole
3736 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3737 back references of the form \12, where the number was greater than 9, were
3738 not recognized as back references, even though there were sufficient
3739 previous subpatterns.
3740
3741 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3742 versions of gcc, e.g. 2.95.4.
3743
3744
3745 Version 6.1 21-Jun-05
3746 ---------------------
3747
3748 1. There was one reference to the variable "posix" in pcretest.c that was not
3749 surrounded by "#if !defined NOPOSIX".
3750
3751 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3752 the cross-check on the old pcre_info() function, for the benefit of the
3753 cut-down version of PCRE that is currently imported into Exim.
3754
3755 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3756 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3757 patterns that start (?i) but I don't think that matters. The use of (?i) is
3758 just an example; this all applies to the other options as well.
3759
3760 4. Since libtool seems to echo the compile commands it is issuing, the output
3761 from "make" can be reduced a bit by putting "@" in front of each libtool
3762 compile command.
3763
3764 5. Patch from the folks at Google for configure.in to be a bit more thorough
3765 in checking for a suitable C++ installation before trying to compile the
3766 C++ stuff. This should fix a reported problem when a compiler was present,
3767 but no suitable headers.
3768
3769 6. The man pages all had just "PCRE" as their title. I have changed them to
3770 be the relevant file name. I have also arranged that these names are
3771 retained in the file doc/pcre.txt, which is a concatenation in text format
3772 of all the man pages except the little individual ones for each function.
3773
3774 7. The NON-UNIX-USE file had not been updated for the different set of source
3775 files that come with release 6. I also added a few comments about the C++
3776 wrapper.
3777
3778
3779 Version 6.0 07-Jun-05
3780 ---------------------
3781
3782 1. Some minor internal re-organization to help with my DFA experiments.
3783
3784 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3785 didn't matter for the library itself when fully configured, but did matter
3786 when compiling without UCP support, or within Exim, where the ucp files are
3787 not imported.
3788
3789 3. Refactoring of the library code to split up the various functions into
3790 different source modules. The addition of the new DFA matching code (see
3791 below) to a single monolithic source would have made it really too
3792 unwieldy, quite apart from causing all the code to be include in a
3793 statically linked application, when only some functions are used. This is
3794 relevant even without the DFA addition now that patterns can be compiled in
3795 one application and matched in another.
3796
3797 The downside of splitting up is that there have to be some external
3798 functions and data tables that are used internally in different modules of
3799 the library but which are not part of the API. These have all had their
3800 names changed to start with "_pcre_" so that they are unlikely to clash
3801 with other external names.
3802
3803 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3804 a different (DFA) algorithm. Although it is slower than the original
3805 function, it does have some advantages for certain types of matching
3806 problem.
3807
3808 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3809 including restarting after a partial match.
3810
3811 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3812 defined when compiling for Windows was sent to me. I have put it into the
3813 code, though I have no means of testing or verifying it.
3814
3815 7. Added the pcre_refcount() auxiliary function.
3816
3817 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3818 match before or at the first newline in the subject string. In pcretest,
3819 the /f option on a pattern can be used to set this.
3820
3821 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3822 would behave wrongly. This has been present in PCRE since release 4.0.
3823
3824 10. A number of changes to the pcregrep command:
3825
3826 (a) Refactored how -x works; insert ^(...)$ instead of setting
3827 PCRE_ANCHORED and checking the length, in preparation for adding
3828 something similar for -w.
3829
3830 (b) Added the -w (match as a word) option.
3831
3832 (c) Refactored the way lines are read and buffered so as to have more
3833 than one at a time available.
3834
3835 (d) Implemented a pcregrep test script.
3836
3837 (e) Added the -M (multiline match) option. This allows patterns to match
3838 over several lines of the subject. The buffering ensures that at least
3839 8K, or the rest of the document (whichever is the shorter) is available
3840 for matching (and similarly the previous 8K for lookbehind assertions).
3841
3842 (f) Changed the --help output so that it now says
3843
3844 -w, --word-regex(p)
3845
3846 instead of two lines, one with "regex" and the other with "regexp"
3847 because that confused at least one person since the short forms are the
3848 same. (This required a bit of code, as the output is generated
3849 automatically from a table. It wasn't just a text change.)
3850
3851 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3852 option but starts with a hyphen. Could be a pattern or a path name
3853 starting with a hyphen, for instance.
3854
3855 (h) "-" can be given as a file name to represent stdin.
3856
3857 (i) When file names are being printed, "(standard input)" is used for
3858 the standard input, for compatibility with GNU grep. Previously
3859 "<stdin>" was used.
3860
3861 (j) The option --label=xxx can be used to supply a name to be used for
3862 stdin when file names are being printed. There is no short form.
3863
3864 (k) Re-factored the options decoding logic because we are going to add
3865 two more options that take data. Such options can now be given in four
3866 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3867
3868 (l) Added the -A, -B, and -C options for requesting that lines of context
3869 around matches be printed.
3870
3871 (m) Added the -L option to print the names of files that do not contain
3872 any matching lines, that is, the complement of -l.
3873
3874 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3875 continue to scan other files.
3876
3877 (o) The -s option was incorrectly implemented. For compatibility with other
3878 greps, it now suppresses the error message for a non-existent or non-
3879 accessible file (but not the return code). There is a new option called
3880 -q that suppresses the output of matching lines, which was what -s was
3881 previously doing.
3882
3883 (p) Added --include and --exclude options to specify files for inclusion
3884 and exclusion when recursing.
3885
3886 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3887 Hopefully, it now does.
3888
3889 12. Missing cast in pcre_study().
3890
3891 13. Added an "uninstall" target to the makefile.
3892
3893 14. Replaced "extern" in the function prototypes in Makefile.in with
3894 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3895 world, but is set differently for Windows.
3896
3897 15. Added a second compiling function called pcre_compile2(). The only
3898 difference is that it has an extra argument, which is a pointer to an
3899 integer error code. When there is a compile-time failure, this is set
3900 non-zero, in addition to the error test pointer being set to point to an
3901 error message. The new argument may be NULL if no error number is required
3902 (but then you may as well call pcre_compile(), which is now just a
3903 wrapper). This facility is provided because some applications need a
3904 numeric error indication, but it has also enabled me to tidy up the way
3905 compile-time errors are handled in the POSIX wrapper.
3906
3907 16. Added VPATH=.libs to the makefile; this should help when building with one
3908 prefix path and installing with another. (Or so I'm told by someone who
3909 knows more about this stuff than I do.)
3910
3911 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3912 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3913 match everything, including newlines. This is not POSIX-compatible, but
3914 somebody wanted the feature. From pcretest it can be activated by using
3915 both the P and the s flags.
3916
3917 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3918
3919 19. libpcre.pc was being incorrectly installed as executable.
3920
3921 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3922 it now also looks for '\r' so that it will work unmodified on Windows.
3923
3924 21. Added Google's contributed C++ wrapper to the distribution.
3925
3926 22. Added some untidy missing memory free() calls in pcretest, to keep
3927 Electric Fence happy when testing.
3928
3929
3930
3931 Version 5.0 13-Sep-04
3932 ---------------------
3933
3934 1. Internal change: literal characters are no longer packed up into items
3935 containing multiple characters in a single byte-string. Each character
3936 is now matched using a separate opcode. However, there may be more than one
3937 byte in the character in UTF-8 mode.
3938
3939 2. The pcre_callout_block structure has two new fields: pattern_position and
3940 next_item_length. These contain the offset in the pattern to the next match
3941 item, and its length, respectively.
3942
3943 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3944 insertion of callouts before each pattern item. Added the /C option to
3945 pcretest to make use of this.
3946
3947 4. On the advice of a Windows user, the lines
3948
3949 #if defined(_WIN32) || defined(WIN32)
3950 _setmode( _fileno( stdout ), 0x8000 );
3951 #endif /* defined(_WIN32) || defined(WIN32) */
3952
3953 have been added to the source of pcretest. This apparently does useful
3954 magic in relation to line terminators.
3955
3956 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3957 for the benefit of those environments where the "b" makes a difference.
3958
3959 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3960 to know about it. I have put a hack into configure.in that adds in code
3961 to set GCC=yes if CC=icc. This seems to end up at a point in the
3962 generated configure script that is early enough to affect the setting of
3963 compiler options, which is what is needed, but I have no means of testing
3964 whether it really works. (The user who reported this had patched the
3965 generated configure script, which of course I cannot do.)
3966
3967 LATER: After change 22 below (new libtool files), the configure script
3968 seems to know about icc (and also ecc). Therefore, I have commented out
3969 this hack in configure.in.
3970
3971 7. Added support for pkg-config (2 patches were sent in).
3972
3973 8. Negated POSIX character classes that used a combination of internal tables
3974 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3975 [[:^ascii]]. Typically, they would match almost any characters. The other
3976 POSIX classes were not broken in this way.
3977
3978 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3979 to find the match, as PCRE was deluded into thinking that the match had to
3980 start at the start point or following a newline. The same bug applied to
3981 patterns with negative forward assertions or any backward assertions
3982 preceding ".*" at the start, unless the pattern required a fixed first
3983 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3984
3985 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3986 starting at the last subject character, bytes beyond the end of the subject
3987 string were read.
3988
3989 11. Renamed the variable "class" as "classbits" to make life easier for C++
3990 users. (Previously there was a macro definition, but it apparently wasn't
3991 enough.)
3992
3993 12. Added the new field "tables" to the extra data so that tables can be passed
3994 in at exec time, or the internal tables can be re-selected. This allows
3995 a compiled regex to be saved and re-used at a later time by a different
3996 program that might have everything at different addresses.
3997
3998 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3999 -R library as well as a -L library.
4000
4001 14. The debugging options of pcretest (-d on the command line or D on a
4002 pattern) showed incorrect output for anything following an extended class
4003 that contained multibyte characters and which was followed by a quantifier.
4004
4005 15. Added optional support for general category Unicode character properties
4006 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4007 support. It adds about 90K to the size of the library. The meanings of the
4008 inbuilt class escapes such as \d and \s have NOT been changed.
4009
4010 16. Updated pcredemo.c to include calls to free() to release the memory for the
4011 compiled pattern.
4012
4013 17. The generated file chartables.c was being created in the source directory
4014 instead of in the building directory. This caused the build to fail if the
4015 source directory was different from the building directory, and was
4016 read-only.
4017
4018 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4019 file. No doubt somebody will tell me if they don't make sense... Also added
4020 Dan Mooney's comments about building on OpenVMS.
4021
4022 19. Added support for partial matching via the PCRE_PARTIAL option for
4023 pcre_exec() and the \P data escape in pcretest.
4024
4025 20. Extended pcretest with 3 new pattern features:
4026
4027 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4028 write the compiled pattern to the file whose name is "rest-of-line".
4029 This is a straight binary dump of the data, with the saved pointer to
4030 the character tables forced to be NULL. The study data, if any, is
4031 written too. After writing, pcretest reads a new pattern.
4032
4033 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4034 compiled pattern from the given file. There must not be any
4035 occurrences of "<" in the file name (pretty unlikely); if there are,
4036 pcretest will instead treat the initial "<" as a pattern delimiter.
4037 After reading in the pattern, pcretest goes on to read data lines as
4038 usual.
4039
4040 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4041 and 16-bit fields in a compiled pattern, to simulate a pattern that
4042 was compiled on a host of opposite endianness.
4043
4044 21. The pcre-exec() function can now cope with patterns that were compiled on
4045 hosts of opposite endianness, with this restriction:
4046
4047 As for any compiled expression that is saved and used later, the tables
4048 pointer field cannot be preserved; the extra_data field in the arguments
4049 to pcre_exec() should be used to pass in a tables address if a value
4050 other than the default internal tables were used at compile time.
4051
4052 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4053 now diagnosed as an error. Previously, most of the time, a negative number
4054 would have been treated as zero, but if in addition "ovector" was passed as
4055 NULL, a crash could occur.
4056
4057 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4058 new versions from the libtool 1.5 distribution (the last one is a copy of
4059 a file called libtool.m4). This seems to have fixed the need to patch
4060 "configure" to support Darwin 1.3 (which I used to do). However, I still
4061 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4062 workstation).
4063
4064 24. Changed the PCRE licence to be the more standard "BSD" licence.
4065
4066
4067 Version 4.5 01-Dec-03
4068 ---------------------
4069
4070 1. There has been some re-arrangement of the code for the match() function so
4071 that it can be compiled in a version that does not call itself recursively.
4072 Instead, it keeps those local variables that need separate instances for
4073 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4074 needs to "recurse". Keeping track of where control must go is done by means
4075 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4076 hide most of the details from the main code, and operates only if
4077 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4078 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4079 operating.
4080
4081 To make it easier for callers to provide specially tailored get/free
4082 functions for this usage, two new functions, pcre_stack_malloc, and
4083 pcre_stack_free, are used. They are always called in strict stacking order,
4084 and the size of block requested is always the same.
4085
4086 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4087 PCRE has been compiled to use the stack or the heap for recursion. The
4088 -C option of pcretest uses this to show which version is compiled.
4089
4090 A new data escape \S, is added to pcretest; it causes the amounts of store
4091 obtained and freed by both kinds of malloc/free at match time to be added
4092 to the output.
4093
4094 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4095 what's available on my current Linux desktop machine.
4096
4097 3. When matching a UTF-8 string, the test for a valid string at the start has
4098 been extended. If start_offset is not zero, PCRE now checks that it points
4099 to a byte that is the start of a UTF-8 character. If not, it returns
4100 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4101 this is necessary because there may be backward assertions in the pattern.
4102 When matching the same subject several times, it may save resources to use
4103 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4104
4105 4. The code for checking the validity of UTF-8 strings has been tightened so
4106 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4107 containing "overlong sequences".
4108
4109 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4110 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4111 should have been "&", but it just so happened that all the cases this let
4112 through by mistake were picked up later in the function.
4113
4114 6. I had used a variable called "isblank" - this is a C99 function, causing
4115 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4116
4117 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4118 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4119 so the version is shown; (c) stop "make test" echoing "./RunTest".
4120
4121 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4122
4123 9. The prototype for memmove() for systems that don't have it was using
4124 size_t, but the inclusion of the header that defines size_t was later. I've
4125 moved the #includes for the C headers earlier to avoid this.
4126
4127 10. Added some adjustments to the code to make it easier to compiler on certain
4128 special systems:
4129
4130 (a) Some "const" qualifiers were missing.
4131 (b) Added the macro EXPORT before all exported functions; by default this
4132 is defined to be empty.
4133 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4134 that it reads its output file name as an argument instead of writing
4135 to the standard output and assuming this can be redirected.
4136
4137 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4138 class containing characters with values greater than 255, PCRE compilation
4139 went into a loop.
4140
4141 12. A recursive reference to a subpattern that was within another subpattern
4142 that had a minimum quantifier of zero caused PCRE to crash. For example,
4143 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4144 recursion. If the recursively-called subpattern itself had a zero repeat,
4145 that was OK.
4146
4147 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4148 buffer into which it was copied (for escape processing) was still set at
4149 1024, so long lines caused crashes.
4150
4151 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4152 "internal error: code overflow...". This applied to any character class
4153 that was followed by a possessive quantifier.
4154
4155 15. Modified the Makefile to add libpcre.la as a prerequisite for
4156 libpcreposix.la because I was told this is needed for a parallel build to
4157 work.
4158
4159 16. If a pattern that contained .* following optional items at the start was
4160 studied, the wrong optimizing data was generated, leading to matching
4161 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4162 matching string must start with a or b or c. The correct conclusion for
4163 this pattern is that a match can start with any character.
4164
4165
4166 Version 4.4 13-Aug-03
4167 ---------------------
4168
4169 1. In UTF-8 mode, a character class containing characters with values between
4170 127 and 255 was not handled correctly if the compiled pattern was studied.
4171 In fixing this, I have also improved the studying algorithm for such
4172 classes (slightly).
4173
4174 2. Three internal functions had redundant arguments passed to them. Removal
4175 might give a very teeny performance improvement.
4176
4177 3. Documentation bug: the value of the capture_top field in a callout is *one
4178 more than* the number of the hightest numbered captured substring.
4179
4180 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4181 in incorrectly linking with a previously installed version. They now link
4182 explicitly with libpcre.la.
4183
4184 5. configure.in no longer needs to recognize Cygwin specially.
4185
4186 6. A problem in pcre.in for Windows platforms is fixed.
4187
4188 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4189 pcretest, it used to include the size of the study block as part of its
4190 output. Unfortunately, the structure contains a field that has a different
4191 size on different hardware architectures. This meant that the tests that
4192 showed this size failed. As the block is currently always of a fixed size,
4193 this information isn't actually particularly useful in pcretest output, so
4194 I have just removed it.
4195
4196 8. Three pre-processor statements accidentally did not start in column 1.
4197 Sadly, there are *still* compilers around that complain, even though
4198 standard C has not required this for well over a decade. Sigh.
4199
4200 9. In pcretest, the code for checking callouts passed small integers in the
4201 callout_data field, which is a void * field. However, some picky compilers
4202 complained about the casts involved for this on 64-bit systems. Now
4203 pcretest passes the address of the small integer instead, which should get
4204 rid of the warnings.
4205
4206 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4207 both compile and run time, and gives an error if an invalid UTF-8 sequence
4208 is found. There is a option for disabling this check in cases where the
4209 string is known to be correct and/or the maximum performance is wanted.
4210
4211 11. In response to a bug report, I changed one line in Makefile.in from
4212
4213 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4214 to
4215 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4216
4217 to look similar to other lines, but I have no way of telling whether this
4218 is the right thing to do, as I do not use Windows. No doubt I'll get told
4219 if it's wrong...
4220
4221
4222 Version 4.3 21-May-03
4223 ---------------------
4224
4225 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4226 Makefile.
4227
4228 2. Some refactoring to improve the quality of the code:
4229
4230 (i) The utf8_table... variables are now declared "const".
4231
4232 (ii) The code for \cx, which used the "case flipping" table to upper case
4233 lower case letters, now just substracts 32. This is ASCII-specific,
4234 but the whole concept of \cx is ASCII-specific, so it seems
4235 reasonable.
4236
4237 (iii) PCRE was using its character types table to recognize decimal and
4238 hexadecimal digits in the pattern. This is silly, because it handles
4239 only 0-9, a-f, and A-F, but the character types table is locale-
4240 specific, which means strange things might happen. A private
4241 table is now used for this - though it costs 256 bytes, a table is
4242 much faster than multiple explicit tests. Of course, the standard
4243 character types table is still used for matching digits in subject
4244 strings against \d.
4245
4246 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4247 ending in _t are). So I've renamed it as ESC_tee.
4248
4249 3. The first argument for regexec() in the POSIX wrapper should have been
4250 defined as "const".
4251
4252 4. Changed pcretest to use malloc() for its buffers so that they can be
4253 Electric Fenced for debugging.
4254
4255 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4256 to read one or more bytes before the start of the subject string. Often this
4257 had no effect on PCRE's behaviour, but in some circumstances it could
4258 provoke a segmentation fault.
4259
4260 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4261 to try to read one or more bytes before the start of the subject string.
4262
4263 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4264 UTF-8 support could misbehave in various ways if the subject string
4265 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4266 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4267 back over UTF-8 characters.)
4268
4269
4270 Version 4.2 14-Apr-03
4271 ---------------------
4272
4273 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4274
4275 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4276 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4277 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4278 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4279 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4280 and BUILD_EXEEXT
4281 Note: automatic setting of the BUILD variables is not yet working
4282 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4283 compile-time but not at link-time
4284 [LINK]: use for linking executables only
4285 make different versions for Windows and non-Windows
4286 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4287 libraries
4288 [LINK_FOR_BUILD]: new variable
4289 [OBJEXT]: use throughout
4290 [EXEEXT]: use throughout
4291 <winshared>: new target
4292 <wininstall>: new target
4293 <dftables.o>: use native compiler
4294 <dftables>: use native linker
4295 <install>: handle Windows platform correctly
4296 <clean>: ditto
4297 <check>: ditto
4298 copy DLL to top builddir before testing
4299
4300 As part of these changes, -no-undefined was removed again. This was reported
4301 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4302 in any case.
4303
4304 3. Some tidies to get rid of compiler warnings:
4305
4306 . In the match_data structure, match_limit was an unsigned long int, whereas
4307 match_call_count was an int. I've made them both unsigned long ints.
4308
4309 . In pcretest the fact that a const uschar * doesn't automatically cast to
4310 a void * provoked a warning.
4311
4312 . Turning on some more compiler warnings threw up some "shadow" variables
4313 and a few more missing casts.
4314
4315 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4316 option, a class that contained a single character with a value between 128
4317 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4318
4319 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4320 option, a class that contained several characters, but with at least one
4321 whose value was between 128 and 255 caused PCRE to crash.
4322
4323
4324 Version 4.1 12-Mar-03
4325 ---------------------
4326
4327 1. Compiling with gcc -pedantic found a couple of places where casts were
4328 needed, and a string in dftables.c that was longer than standard compilers are
4329 required to support.
4330
4331 2. Compiling with Sun's compiler found a few more places where the code could
4332 be tidied up in order to avoid warnings.
4333
4334 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4335 first of these names is deprecated in the latest Autoconf in favour of the name
4336 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4337 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4338 analogy I have changed it to CFLAGS_FOR_BUILD.
4339
4340 4. Added -no-undefined to the linking command in the Makefile, because this is
4341 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4342 linking step for the pcreposix library.
4343
4344 5. PCRE was failing to diagnose the case of two named groups with the same
4345 name.
4346
4347 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4348 literal character that is needed in the subject for a match, and scans along to
4349 ensure that it is present before embarking on the full matching process. This
4350 saves time in cases of nested unlimited repeats that are never going to match.
4351 Problem: the scan can take a lot of time if the subject is very long (e.g.
4352 megabytes), thus penalizing straightforward matches. It is now done only if the
4353 amount of subject to be scanned is less than 1000 bytes.
4354
4355 7. A lesser problem with the same optimization is that it was recording the
4356 first character of an anchored pattern as "needed", thus provoking a search
4357 right along the subject, even when the first match of the pattern was going to
4358 fail. The "needed" character is now not set for anchored patterns, unless it
4359 follows something in the pattern that is of non-fixed length. Thus, it still
4360 fulfils its original purpose of finding quick non-matches in cases of nested
4361 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4362
4363
4364 Version 4.0 17-Feb-03
4365 ---------------------
4366
4367 1. If a comment in an extended regex that started immediately after a meta-item
4368 extended to the end of string, PCRE compiled incorrect data. This could lead to
4369 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4370
4371 2. Moved to autoconf 2.53 and libtool 1.4.2.
4372
4373 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4374 the special perltest8 script is no longer needed - all the tests can be run
4375 from a single perltest script.
4376
4377 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4378 by \s. It has now been removed in PCRE. This means it isn't recognized as
4379 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4380 class [:space:] *does* include VT, thereby creating a mess.
4381
4382 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4383 space and tab.
4384
4385 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4386 its new features into the main test script, reducing the number of scripts.
4387
4388 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4389 were backward compatible, and made the (?i) apply to the whole pattern, as if
4390 /i were given. Now it behaves more logically, and applies the option setting
4391 only to what follows. PCRE has been changed to follow suit. However, if it
4392 finds options settings right at the start of the pattern, it extracts them into
4393 the global options, as before. Thus, they show up in the info data.
4394
4395 8. Added support for the \Q...\E escape sequence. Characters in between are
4396 treated as literals. This is slightly different from Perl in that $ and @ are
4397 also handled as literals inside the quotes. In Perl, they will cause variable
4398 interpolation. Note the following examples:
4399
4400 Pattern PCRE matches Perl matches
4401
4402 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4403 \Qabc\$xyz\E abc\$xyz abc\$xyz
4404 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4405
4406 For compatibility with Perl, \Q...\E sequences are recognized inside character
4407 classes as well as outside them.
4408
4409 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4410 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4411 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4412 signed/unsigned warnings.
4413
4414 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4415 option for pcretest, so I've replaced it by a simple function that does just
4416 that job.
4417
4418 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4419 "pcregrep -".
4420
4421 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4422 Java package. This provides some syntactic sugar for simple cases of what my
4423 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4424 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4425 item, you can use this simplified notation. Note that only makes sense with
4426 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4427 greediness, whatever the setting of the PCRE_UNGREEDY option.
4428
4429 13. A change of greediness default within a pattern was not taking effect at
4430 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4431 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4432 was abstracted outside.
4433
4434 14. PCRE now supports the \G assertion. It is true when the current matching
4435 position is at the start point of the match. This differs from \A when the
4436 starting offset is non-zero. Used with the /g option of pcretest (or similar
4437 code), it works in the same way as it does for Perl's /g option. If all
4438 alternatives of a regex begin with \G, the expression is anchored to the start
4439 match position, and the "anchored" flag is set in the compiled expression.
4440
4441 15. Some bugs concerning the handling of certain option changes within patterns
4442 have been fixed. These applied to options other than (?ims). For example,
4443 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4444 been the other way round. Some of this was related to change 7 above.
4445
4446 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4447 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4448 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4449 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4450
4451 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4452 mode. Unlike ".", it always matches newline, whatever the setting of
4453 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4454 assertions. Perl allows it, but it doesn't (in general) work because it can't
4455 calculate the length of the lookbehind. At least, that's the case for Perl
4456 5.8.0 - I've been told they are going to document that it doesn't work in
4457 future.
4458
4459 18. Added an error diagnosis for escapes that PCRE does not support: these are
4460 \L, \l, \N, \P, \p, \U, \u, and \X.
4461
4462 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4463 reading past the end of the pattern in cases such as /[abcd/.
4464
4465 20. PCRE was getting more memory than necessary for patterns with classes that
4466 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4467
4468 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4469 compiling PCRE for use with Virtual Pascal.
4470
4471 22. Small fix to the Makefile to make it work properly if the build is done
4472 outside the source tree.
4473
4474 23. Added a new extension: a condition to go with recursion. If a conditional
4475 subpattern starts with (?(R) the "true" branch is used if recursion has
4476 happened, whereas the "false" branch is used only at the top level.
4477
4478 24. When there was a very long string of literal characters (over 255 bytes
4479 without UTF support, over 250 bytes with UTF support), the computation of how
4480 much memory was required could be incorrect, leading to segfaults or other
4481 strange effects.
4482
4483 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4484 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4485 there was a subsequent back reference to those brackets. This meant that, for
4486 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4487 possible to check for precisely this case. All we can do is abandon the
4488 optimization if .* occurs inside capturing brackets when there are any back
4489 references whatsoever. (See below for a better fix that came later.)
4490
4491 26. The handling of the optimization for finding the first character of a
4492 non-anchored pattern, and for finding a character that is required later in the
4493 match were failing in some cases. This didn't break the matching; it just
4494 failed to optimize when it could. The way this is done has been re-implemented.
4495
4496 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4497
4498 28. Added a new feature that provides some of the functionality that Perl
4499 provides with (?{...}). The facility is termed a "callout". The way it is done
4500 in PCRE is for the caller to provide an optional function, by setting
4501 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4502 global variable. By default it is unset, which disables all calling out. To get
4503 the function called, the regex must include (?C) at appropriate points. This
4504 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4505 This provides a means of identifying different callout points. When PCRE
4506 reaches such a point in the regex, if pcre_callout has been set, the external
4507 function is called. It is provided with data in a structure called
4508 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4509 matching continues; if it returns a non-zero value, the match at the current
4510 point fails. However, backtracking will occur if possible. [This was changed
4511 later and other features added - see item 49 below.]
4512
4513 29. pcretest is upgraded to test the callout functionality. It provides a
4514 callout function that displays information. By default, it shows the start of
4515 the match and the current position in the text. There are some new data escapes
4516 to vary what happens:
4517
4518 \C+ in addition, show current contents of captured substrings
4519 \C- do not supply a callout function
4520 \C!n return 1 when callout number n is reached
4521 \C!n!m return 1 when callout number n is reached for the mth time
4522
4523 30. If pcregrep was called with the -l option and just a single file name, it
4524 output "<stdin>" if a match was found, instead of the file name.
4525
4526 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4527 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4528 pcre_exec(). This saves a malloc/free per call. The default value of
4529 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4530 when configuring.
4531
4532 32. The default maximum size of a compiled pattern is 64K. There have been a
4533 few cases of people hitting this limit. The code now uses macros to handle the
4534 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4535 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4536 configuring. Tests 2 and 5 work only with 2-byte links because they output
4537 debugging information about compiled patterns.
4538
4539 33. Internal code re-arrangements:
4540
4541 (a) Moved the debugging function for printing out a compiled regex into
4542 its own source file (printint.c) and used #include to pull it into
4543 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4544 separate copies.
4545
4546 (b) Defined the list of op-code names for debugging as a macro in
4547 internal.h so that it is next to the definition of the opcodes.
4548
4549 (c) Defined a table of op-code lengths for simpler skipping along compiled
4550 code. This is again a macro in internal.h so that it is next to the
4551 definition of the opcodes.
4552
4553 34. Added support for recursive calls to individual subpatterns, along the
4554 lines of Robin Houston's patch (but implemented somewhat differently).
4555
4556 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4557 allow it to read and process whole directories in Win32. This code was
4558 contributed by Lionel Fourquaux; it has not been tested by me.
4559
4560 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4561 used to name a group. Names consist of alphanumerics and underscores, and must
4562 be unique. Back references use the syntax (?P=name) and recursive calls use
4563 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4564 numbers. The function pcre_fullinfo() can be used after compilation to extract
4565 a name/number map. There are three relevant calls:
4566
4567 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4568 PCRE_INFO_NAMECOUNT yields the number of entries
4569 PCRE_INFO_NAMETABLE yields a pointer to the map.
4570
4571 The map is a vector of fixed-size entries. The size of each entry depends on
4572 the length of the longest name used. The first two bytes of each entry are the
4573 group number, most significant byte first. There follows the corresponding
4574 name, zero terminated. The names are in alphabetical order.
4575
4576 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4577 case instead of 255. Making it the same both with and without UTF-8 support
4578 means that the same test output works with both.
4579
4580 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4581 calling malloc() with a zero argument.
4582
4583 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4584 optimization. I've improved things by keeping a bitmap of backreferences with
4585 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4586 fact referenced, the optimization can be applied. It is unlikely that a
4587 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4588 the match to follow \n) will appear inside brackets with a number greater than
4589 31, but if it does, any back reference > 31 suppresses the optimization.
4590
4591 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4592 of disabling numbered capturing parentheses. Any opening parenthesis that is
4593 not followed by ? behaves as if it were followed by ?: but named parentheses
4594 can still be used for capturing (and they will acquire numbers in the usual
4595 way).
4596
4597 41. Redesigned the return codes from the match() function into yes/no/error so
4598 that errors can be passed back from deep inside the nested calls. A malloc
4599 failure while inside a recursive subpattern call now causes the
4600 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4601
4602 42. It is now possible to set a limit on the number of times the match()
4603 function is called in a call to pcre_exec(). This facility makes it possible to
4604 limit the amount of recursion and backtracking, though not in a directly
4605 obvious way, because the match() function is used in a number of different
4606 circumstances. The count starts from zero for each position in the subject
4607 string (for non-anchored patterns). The default limit is, for compatibility, a
4608 large number, namely 10 000 000. You can change this in two ways:
4609
4610 (a) When configuring PCRE before making, you can use --with-match-limit=n
4611 to set a default value for the compiled library.
4612
4613 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4614 a different value is set. See 45 below.
4615
4616 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4617
4618 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4619 of things that can be changed at compile time. The first argument specifies
4620 what is wanted and the second points to where the information is to be placed.
4621 The current list of available information is:
4622
4623 PCRE_CONFIG_UTF8
4624
4625 The output is an integer that is set to one if UTF-8 support is available;
4626 otherwise it is set to zero.
4627
4628 PCRE_CONFIG_NEWLINE
4629
4630 The output is an integer that it set to the value of the code that is used for
4631 newline. It is either LF (10) or CR (13).
4632
4633 PCRE_CONFIG_LINK_SIZE
4634
4635 The output is an integer that contains the number of bytes used for internal
4636 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4637
4638 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4639
4640 The output is an integer that contains the threshold above which the POSIX
4641 interface uses malloc() for output vectors. See item 31 above.
4642
4643 PCRE_CONFIG_MATCH_LIMIT
4644
4645 The output is an unsigned integer that contains the default limit of the number
4646 of match() calls in a pcre_exec() execution. See 42 above.
4647
4648 44. pcretest has been upgraded by the addition of the -C option. This causes it
4649 to extract all the available output from the new pcre_config() function, and to
4650 output it. The program then exits immediately.
4651
4652 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4653 order to support additional features. One way would have been to define
4654 pcre_exec2() (for example) with extra arguments, but this would not have been
4655 extensible, and would also have required all calls to the original function to
4656 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4657 is used for passing in "extra" data from pcre_study().
4658
4659 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4660 contains the following fields:
4661
4662 flags a bitmap indicating which of the following fields are set
4663 study_data opaque data from pcre_study()
4664 match_limit a way of specifying a limit on match() calls for a specific
4665 call to pcre_exec()
4666 callout_data data for callouts (see 49 below)
4667
4668 The flag bits are also defined in pcre.h, and are
4669
4670 PCRE_EXTRA_STUDY_DATA
4671 PCRE_EXTRA_MATCH_LIMIT
4672 PCRE_EXTRA_CALLOUT_DATA
4673
4674 The pcre_study() function now returns one of these new pcre_extra blocks, with
4675 the actual study data pointed to by the study_data field, and the
4676 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4677 before. That is, this change is entirely upwards-compatible and requires no
4678 change to existing code.
4679
4680 If you want to pass in additional data to pcre_exec(), you can either place it
4681 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4682 block.
4683
4684 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4685 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4686 times with different match limits, until it finds the minimum value needed for
4687 pcre_exec() to complete. The value is then output. This can be instructive; for
4688 most simple matches the number is quite small, but for pathological cases it
4689 gets very large very quickly.
4690
4691 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4692 returns the size of the data block pointed to by the study_data field in a
4693 pcre_extra block, that is, the value that was passed as the argument to
4694 pcre_malloc() when PCRE was getting memory in which to place the information
4695 created by pcre_study(). The fourth argument should point to a size_t variable.
4696 pcretest has been extended so that this information is shown after a successful
4697 pcre_study() call when information about the compiled regex is being displayed.
4698
4699 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4700 because what follows is always an absolute path. (Later: it turns out that this
4701 is more than cosmetic for MinGW, because it doesn't like empty path
4702 components.)
4703
4704 49. Some changes have been made to the callout feature (see 28 above):
4705
4706 (i) A callout function now has three choices for what it returns:
4707
4708 0 => success, carry on matching
4709 > 0 => failure at this point, but backtrack if possible
4710 < 0 => serious error, return this value from pcre_exec()
4711
4712 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4713 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4714 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4715 use by callout functions. It will never be used by PCRE itself.
4716
4717 (ii) The pcre_extra structure (see 45 above) has a void * field called
4718 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4719 pcre_callout_block structure has a field of the same name. The contents of
4720 the field passed in the pcre_extra structure are passed to the callout
4721 function in the corresponding field in the callout block. This makes it
4722 easier to use the same callout-containing regex from multiple threads. For
4723 testing, the pcretest program has a new data escape
4724
4725 \C*n pass the number n (may be negative) as callout_data
4726
4727 If the callout function in pcretest receives a non-zero value as
4728 callout_data, it returns that value.
4729
4730 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4731 there were some redundant $(CFLAGS) in commands that are now specified as
4732 $(LINK), which already includes $(CFLAGS).
4733
4734 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4735 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4736 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4737 one-byte characters throughout. Note that case-insensitive matching applies
4738 only to characters whose values are less than 256. PCRE doesn't support the
4739 notion of cases for higher-valued characters.
4740
4741 (i) A character class whose characters are all within 0-255 is handled as
4742 a bit map, and the map is inverted for negative classes. Previously, a
4743 character > 255 always failed to match such a class; however it should
4744 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4745
4746 (ii) A negated character class with a single character < 255 is coded as
4747 "not this character" (OP_NOT). This wasn't working properly when the test
4748 character was multibyte, either singly or repeated.
4749
4750 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4751 mode, for example: \x{100}{2,3}.
4752
4753 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4754 singly or repeated) now correctly test multibyte characters. However,
4755 PCRE doesn't recognize any characters with values greater than 255 as
4756 digits, spaces, or word characters. Such characters always match \D, \S,
4757 and \W, and never match \d, \s, or \w.
4758
4759 (v) Classes may now contain characters and character ranges with values
4760 greater than 255. For example: [ab\x{100}-\x{400}].
4761
4762 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4763 PCRE in UTF-8 mode.
4764
4765 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4766 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4767 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4768 value.)
4769
4770 53. The single man page has become too large. I have therefore split it up into
4771 a number of separate man pages. These also give rise to individual HTML pages;
4772 these are now put in a separate directory, and there is an index.html page that
4773 lists them all. Some hyperlinking between the pages has been installed.
4774
4775 54. Added convenience functions for handling named capturing parentheses.
4776
4777 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4778 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4779 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4780 are faulted.
4781
4782 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4783 calling configure. These values are used when compiling the dftables.c program
4784 which is run to generate the source of the default character tables. They
4785 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4786 you will need to set these values.
4787
4788 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4789
4790
4791 Version 3.9 02-Jan-02
4792 ---------------------
4793
4794 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4795
4796 2. If --disable-static was given, the building process failed when trying to
4797 build pcretest and pcregrep. (For some reason it was using libtool to compile
4798 them, which is not right, as they aren't part of the library.)
4799
4800
4801 Version 3.8 18-Dec-01
4802 ---------------------
4803
4804 1. The experimental UTF-8 code was completely screwed up. It was packing the
4805 bytes in the wrong order. How dumb can you get?
4806
4807
4808 Version 3.7 29-Oct-01
4809 ---------------------
4810
4811 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4812 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4813 this didn't happen under Solaris 8, where I normally test things.
4814
4815 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4816 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4817 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4818 this fix an hour or so after the initial 3.7 release.)
4819
4820
4821 Version 3.6 23-Oct-01
4822 ---------------------
4823
4824 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4825 offsets passed as NULL with zero offset count.
4826
4827 2. The config.guess and config.sub files had not been updated when I moved to
4828 the latest autoconf.
4829
4830
4831 Version 3.5 15-Aug-01
4832 ---------------------
4833
4834 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4835 had been forgotten.
4836
4837 2. By using declared but undefined structures, we can avoid using "void"
4838 definitions in pcre.h while keeping the internal definitions of the structures
4839 private.
4840
4841 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4842 user point of view, this means that both static and shared libraries are built
4843 by default, but this can be individually controlled. More of the work of
4844 handling this static/shared cases is now inside libtool instead of PCRE's make
4845 file.
4846
4847 4. The pcretest utility is now installed along with pcregrep because it is
4848 useful for users (to test regexs) and by doing this, it automatically gets
4849 relinked by libtool. The documentation has been turned into a man page, so
4850 there are now .1, .txt, and .html versions in /doc.
4851
4852 5. Upgrades to pcregrep:
4853 (i) Added long-form option names like gnu grep.
4854 (ii) Added --help to list all options with an explanatory phrase.
4855 (iii) Added -r, --recursive to recurse into sub-directories.
4856 (iv) Added -f, --file to read patterns from a file.
4857
4858 6. pcre_exec() was referring to its "code" argument before testing that
4859 argument for NULL (and giving an error if it was NULL).
4860
4861 7. Upgraded Makefile.in to allow for compiling in a different directory from
4862 the source directory.
4863
4864 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4865 options bits, the pointer it was passed was to an int instead of to an unsigned
4866 long int. This mattered only on 64-bit systems.
4867
4868 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4869 generated) instead of pcre.in, which it its source. Also made the same change
4870 in several of the .c files.
4871
4872 10. A new release of gcc defines printf() as a macro, which broke pcretest
4873 because it had an ifdef in the middle of a string argument for printf(). Fixed
4874 by using separate calls to printf().
4875
4876 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4877 script, to force use of CR or LF instead of \n in the source. On non-Unix
4878 systems, the value can be set in config.h.
4879
4880 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4881 absolute limit. Changed the text of the error message to make this clear, and
4882 likewise updated the man page.
4883
4884 13. The limit of 99 on the number of capturing subpatterns has been removed.
4885 The new limit is 65535, which I hope will not be a "real" limit.
4886
4887
4888 Version 3.4 22-Aug-00
4889 ---------------------
4890
4891 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4892
4893 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4894
4895
4896 Version 3.3 01-Aug-00
4897 ---------------------
4898
4899 1. If an octal character was given, but the value was greater than \377, it
4900 was not getting masked to the least significant bits, as documented. This could
4901 lead to crashes in some systems.
4902
4903 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4904 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4905
4906 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4907 These just pass their arguments on to (pcre_free)(), but they are provided
4908 because some uses of PCRE bind it to non-C systems that can call its functions,
4909 but cannot call free() or pcre_free() directly.
4910
4911 4. Add "make test" as a synonym for "make check". Corrected some comments in
4912 the Makefile.
4913
4914 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4915 Makefile.
4916
4917 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4918 command called pgrep for grepping around the active processes.
4919
4920 7. Added the beginnings of support for UTF-8 character strings.
4921
4922 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4923 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4924 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4925 out for the ar command.)
4926
4927
4928 Version 3.2 12-May-00
4929 ---------------------
4930
4931 This is purely a bug fixing release.
4932
4933 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4934 of ZA. This was just one example of several cases that could provoke this bug,
4935 which was introduced by change 9 of version 2.00. The code for breaking
4936 infinite loops after an iteration that matches an empty string was't working
4937 correctly.
4938
4939 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4940 when matched against abbab (for example). After matching an empty string, it
4941 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4942 caused it to match further down the string than it should.
4943
4944 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4945 was there because it doesn't seem to be needed, and it causes trouble on some
4946 systems, as it is not a Standard C header. It has been removed.
4947
4948 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4949 were reported on the Macintosh. The changes were from
4950
4951 while ((c = *(++ptr)) != 0 && c != '\n');
4952 to
4953 while ((c = *(++ptr)) != 0 && c != '\n') ;
4954
4955 Totally extraordinary, but if that's what it takes...
4956
4957 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4958 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4959 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4960 assumes the way PCRE uses memmove() (always moving upwards).
4961
4962 6. PCRE is being used in one environment where strchr() is not available. There
4963 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4964 faster code anyway.
4965
4966
4967 Version 3.1 09-Feb-00
4968 ---------------------
4969
4970 The only change in this release is the fixing of some bugs in Makefile.in for
4971 the "install" target:
4972
4973 (1) It was failing to install pcreposix.h.
4974
4975 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4976
4977
4978 Version 3.0 01-Feb-00
4979 ---------------------
4980
4981 1. Add support for the /+ modifier to perltest (to output $` like it does in
4982 pcretest).
4983
4984 2. Add support for the /g modifier to perltest.
4985
4986 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4987 matches null strings.
4988
4989 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4990 pattern. Perl treats empty patterns specially - it reuses the most recent
4991 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4992 effect.
4993
4994 5. The POSIX interface was broken in that it was just handing over the POSIX
4995 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4996 required a bigger vector, with some working space on the end. This means that
4997 the POSIX wrapper now has to get and free some memory, and copy the results.
4998
4999 6. Added some simple autoconf support, placing the test data and the
5000 documentation in separate directories, re-organizing some of the
5001 information files, and making it build pcre-config (a GNU standard). Also added
5002 libtool support for building PCRE as a shared library, which is now the
5003 default.
5004
5005 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5006 09 are not valid octal constants. Single digits will be used for minor values
5007 less than 10.
5008
5009 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5010 existing programs that set these in the POSIX interface can use PCRE without
5011 modification.
5012
5013 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5014 return all that pcre_info() returns, plus additional data. The pcre_info()
5015 function is retained for compatibility, but is considered to be obsolete.
5016
5017 10. Added experimental recursion feature (?R) to handle one common case that
5018 Perl 5.6 will be able to do with (?p{...}).
5019
5020 11. Added support for POSIX character classes like [:alpha:], which Perl is
5021 adopting.
5022
5023
5024 Version 2.08 31-Aug-99
5025 ----------------------
5026
5027 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5028 trying to match at the startoffset position, but instead was moving forward to
5029 the next newline as if a previous match had failed.
5030
5031 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5032 and could get into a loop if a null string was matched other than at the start
5033 of the subject.
5034
5035 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5036 be distinguished at compile time, and for completeness also added PCRE_DATE.
5037
5038 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5039 in GnuWin32 environments.
5040
5041
5042 Version 2.07 29-Jul-99
5043 ----------------------
5044
5045 1. The documentation is now supplied in plain text form and HTML as well as in
5046 the form of man page sources.
5047
5048 2. C++ compilers don't like assigning (void *) values to other pointer types.
5049 In particular this affects malloc(). Although there is no problem in Standard
5050 C, I've put in casts to keep C++ compilers happy.
5051
5052 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5053 should be (const char *).
5054
5055 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5056 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5057 However, I haven't made this a standard facility. The documentation doesn't
5058 mention it, and the Makefile doesn't support it.
5059
5060 5. The Makefile now contains an "install" target, with editable destinations at
5061 the top of the file. The pcretest program is not installed.
5062
5063 6. pgrep -V now gives the PCRE version number and date.
5064
5065 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5066 causing the entire string to be ignored, instead of just the last character.
5067
5068 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5069 non-matching string, it can take a very, very long time, even for strings of
5070 quite modest length, because of the nested recursion. PCRE now does better in
5071 some of these cases. It does this by remembering the last required literal
5072 character in the pattern, and pre-searching the subject to ensure it is present
5073 before running the real match. In other words, it applies a heuristic to detect
5074 some types of certain failure quickly, and in the above example, if presented
5075 with a string that has no trailing " it gives "no match" very quickly.
5076
5077 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5078 other alternatives are tried instead.
5079
5080
5081 Version 2.06 09-Jun-99
5082 ----------------------
5083
5084 1. Change pcretest's output for amount of store used to show just the code
5085 space, because the remainder (the data block) varies in size between 32-bit and
5086 64-bit systems.
5087
5088 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5089 start matching at. This allows lookbehinds to work when searching for multiple
5090 occurrences in a string.
5091
5092 3. Added additional options to pcretest for testing multiple occurrences:
5093
5094 /+ outputs the rest of the string that follows a match
5095 /g loops for multiple occurrences, using the new startoffset argument
5096 /G loops for multiple occurrences by passing an incremented pointer
5097
5098 4. PCRE wasn't doing the "first character" optimization for patterns starting
5099 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5100 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5101 the letter 'x'. On long subject strings, this gives a significant speed-up.
5102
5103
5104 Version 2.05 21-Apr-99
5105 ----------------------
5106
5107 1. Changed the type of magic_number from int to long int so that it works
5108 properly on 16-bit systems.
5109
5110 2. Fixed a bug which caused patterns starting with .* not to work correctly
5111 when the subject string contained newline characters. PCRE was assuming
5112 anchoring for such patterns in all cases, which is not correct because .* will
5113 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5114 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5115 must be retried after every newline in the subject.
5116
5117
5118 Version 2.04 18-Feb-99
5119 ----------------------
5120
5121 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5122 computation of the store needed to hold the pattern was incorrect (too large).
5123 If such patterns were nested a few deep, this could multiply and become a real
5124 problem.
5125
5126 2. Added /M option to pcretest to show the memory requirement of a specific
5127 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5128
5129 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5130 compiled in such a way that the backtracking after subsequent failure was
5131 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5132 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5133
5134
5135 Version 2.03 02-Feb-99
5136 ----------------------
5137
5138 1. Fixed typo and small mistake in man page.
5139
5140 2. Added 4th condition (GPL supersedes if conflict) and created separate
5141 LICENCE file containing the conditions.
5142
5143 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5144 Perl, that is the internal \ allows the delimiter to be included in the
5145 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5146 the final delimiter, add \ to the end of the pattern (to test the error).
5147
5148 4. Added the convenience functions for extracting substrings after a successful
5149 match. Updated pcretest to make it able to test these functions.
5150
5151
5152 Version 2.02 14-Jan-99
5153 ----------------------
5154
5155 1. Initialized the working variables associated with each extraction so that
5156 their saving and restoring doesn't refer to uninitialized store.
5157
5158 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5159 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5160 fix the problem.
5161
5162 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5163 calls, and wasn't printing the correct value for compiling calls. Increased the
5164 default value of LOOPREPEAT, and the number of significant figures in the
5165 times.
5166
5167 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5168
5169 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5170 a building problem on Windows NT with a FAT file system.
5171
5172
5173 Version 2.01 21-Oct-98
5174 ----------------------
5175
5176 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5177 to character tables built by pcre_maketables() in the current locale. If NULL
5178 is passed, the default tables are used.
5179
5180
5181 Version 2.00 24-Sep-98
5182 ----------------------
5183
5184 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5185 it any more.
5186
5187 2. Allow quantification of (?>) groups, and make it work correctly.
5188
5189 3. The first character computation wasn't working for (?>) groups.
5190
5191 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5192 end of the subject) and add 5.005's \z, which really does match only at the
5193 very end of the subject.
5194
5195 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5196
5197 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5198 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5199 localized options. All options to pcre_study() were also removed.
5200
5201 7. Add other new features from 5.005:
5202
5203 $(?<= positive lookbehind
5204 $(?<! negative lookbehind
5205 (?imsx-imsx) added the unsetting capability
5206 such a setting is global if at outer level; local otherwise
5207 (?imsx-imsx:) non-capturing groups with option setting
5208 (?(cond)re|re) conditional pattern matching
5209
5210 A backreference to itself in a repeated group matches the previous
5211 captured string.
5212
5213 8. General tidying up of studying (both automatic and via "study")
5214 consequential on the addition of new assertions.
5215
5216 9. As in 5.005, unlimited repeated groups that could match an empty substring
5217 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5218 runtime if any iteration does actually match an empty substring.
5219
5220 10. Include the RunTest script in the distribution.
5221
5222 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5223 discrepancies, some of which were old and were also with respect to 5.004. They
5224 have now been fixed.
5225
5226
5227 Version 1.09 28-Apr-98
5228 ----------------------
5229
5230 1. A negated single character class followed by a quantifier with a minimum
5231 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5232 program crashes, or just wrong answers. This did not apply to negated classes
5233 containing more than one character, or to minima other than one.
5234
5235
5236 Version 1.08 27-Mar-98
5237 ----------------------
5238
5239 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5240
5241 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5242 latter must appear before anything that relies on it in the pattern.
5243
5244
5245 Version 1.07 16-Feb-98
5246 ----------------------
5247
5248 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5249 repeat of a potentially empty string).
5250
5251
5252 Version 1.06 23-Jan-98
5253 ----------------------
5254
5255 1. Added Markus Oberhumer's little patches for C++.
5256
5257 2. Literal strings longer than 255 characters were broken.
5258
5259
5260 Version 1.05 23-Dec-97
5261 ----------------------
5262
5263 1. Negated character classes containing more than one character were failing if
5264 PCRE_CASELESS was set at run time.
5265
5266
5267 Version 1.04 19-Dec-97
5268 ----------------------
5269
5270 1. Corrected the man page, where some "const" qualifiers had been omitted.
5271
5272 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5273 input syntax.
5274
5275 3. Fixed memory leak which occurred when a regex with back references was
5276 matched with an offsets vector that wasn't big enough. The temporary memory
5277 that is used in this case wasn't being freed if the match failed.
5278
5279 4. Tidied pcretest to ensure it frees memory that it gets.
5280
5281 5. Temporary memory was being obtained in the case where the passed offsets
5282 vector was exactly big enough.
5283
5284 6. Corrected definition of offsetof() from change 5 below.
5285
5286 7. I had screwed up change 6 below and broken the rules for the use of
5287 setjmp(). Now fixed.
5288
5289
5290 Version 1.03 18-Dec-97
5291 ----------------------
5292
5293 1. A erroneous regex with a missing opening parenthesis was correctly
5294 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5295 on some systems.
5296
5297 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5298 it was reported that one broken compiler failed on the former because "code" is
5299 also an independent variable.
5300
5301 3. The erroneous regex a[]b caused an array overrun reference.
5302
5303 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5304 fail on data ending with that character. (It was going on too far, and checking
5305 the next character, typically a binary zero.) This was specific to the
5306 optimized code for single-character negative classes.
5307
5308 5. Added a contributed patch from the TIN world which does the following:
5309
5310 + Add an undef for memmove, in case the the system defines a macro for it.
5311
5312 + Add a definition of offsetof(), in case there isn't one. (I don't know
5313 the reason behind this - offsetof() is part of the ANSI standard - but
5314 it does no harm).
5315
5316 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5317 most of the places where whitespace preceded '#'. I have given up and
5318 allowed the remaining 2 cases to be at the margin.
5319
5320 + Rename some variables in pcre to eliminate shadowing. This seems very
5321 pedantic, but does no harm, of course.
5322
5323 6. Moved the call to setjmp() into its own function, to get rid of warnings
5324 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5325
5326 7. Constructs such as \d{8,} were compiling into the equivalent of
5327 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5328 outcome, but in this particular case used more store than had been allocated,
5329 which caused the bug to be discovered because it threw up an internal error.
5330
5331 8. The debugging code in both pcre and pcretest for outputting the compiled
5332 form of a regex was going wrong in the case of back references followed by
5333 curly-bracketed repeats.
5334
5335
5336 Version 1.02 12-Dec-97
5337 ----------------------
5338
5339 1. Typos in pcre.3 and comments in the source fixed.
5340
5341 2. Applied a contributed patch to get rid of places where it used to remove
5342 'const' from variables, and fixed some signed/unsigned and uninitialized
5343 variable warnings.
5344
5345 3. Added the "runtest" target to Makefile.
5346
5347 4. Set default compiler flag to -O2 rather than just -O.
5348
5349
5350 Version 1.01 19-Nov-97
5351 ----------------------
5352
5353 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5354 like /([ab]*)*/, that is, for classes with more than one character in them.
5355
5356 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5357 as /((?>a*))*/ (a PCRE_EXTRA facility).
5358
5359
5360 Version 1.00 18-Nov-97
5361 ----------------------
5362
5363 1. Added compile-time macros to support systems such as SunOS4 which don't have
5364 memmove() or strerror() but have other things that can be used instead.
5365
5366 2. Arranged that "make clean" removes the executables.
5367
5368
5369 Version 0.99 27-Oct-97
5370 ----------------------
5371
5372 1. Fixed bug in code for optimizing classes with only one character. It was
5373 initializing a 32-byte map regardless, which could cause it to run off the end
5374 of the memory it had got.
5375
5376 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5377
5378
5379 Version 0.98 22-Oct-97
5380 ----------------------
5381
5382 1. Fixed bug in code for handling temporary memory usage when there are more
5383 back references than supplied space in the ovector. This could cause segfaults.
5384
5385
5386 Version 0.97 21-Oct-97
5387 ----------------------
5388
5389 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5390
5391 2. Optimized negated single characters not to use a bit map.
5392
5393 3. Brought error texts together as macro definitions; clarified some of them;
5394 fixed one that was wrong - it said "range out of order" when it meant "invalid
5395 escape sequence".
5396
5397 4. Changed some char * arguments to const char *.
5398
5399 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5400
5401 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5402 pcretest.
5403
5404
5405 Version 0.96 16-Oct-97
5406 ----------------------
5407
5408 1. Added a simple "pgrep" utility to the distribution.
5409
5410 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5411 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5412 where "ddd" means "one or more decimal digits".
5413
5414 3. Fixed serious bug. If a pattern had a back reference, but the call to
5415 pcre_exec() didn't supply a large enough ovector to record the related
5416 identifying subpattern, the match always failed. PCRE now remembers the number
5417 of the largest back reference, and gets some temporary memory in which to save
5418 the offsets during matching if necessary, in order to ensure that
5419 backreferences always work.
5420
5421 4. Increased the compatibility with Perl in a number of ways:
5422
5423 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5424 to request this handling. The option can be set at compile or exec time.
5425
5426 (b) $ matches before a terminating newline by default; an option
5427 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5428 mode). The option can be set at compile or exec time.
5429
5430 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5431 the same as Perl's. If the decimal number it represents is less than 10
5432 or there aren't that many previous left capturing parentheses, an octal
5433 escape is read. Inside a character class, it's always an octal escape,
5434 even if it is a single digit.
5435
5436 (d) An escaped but undefined alphabetic character is taken as a literal,
5437 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5438 escapes.
5439
5440 (e) {0} is now permitted. (The previous item is removed from the compiled
5441 pattern).
5442
5443 5. Changed all the names of code files so that the basic parts are no longer
5444 than 10 characters, and abolished the teeny "globals.c" file.
5445
5446 6. Changed the handling of character classes; they are now done with a 32-byte
5447 bit map always.
5448
5449 7. Added the -d and /D options to pcretest to make it possible to look at the
5450 internals of compilation without having to recompile pcre.
5451
5452
5453 Version 0.95 23-Sep-97
5454 ----------------------
5455
5456 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5457 \x20 at the start of a run of normal characters. These were being treated as
5458 real characters, instead of the source characters being re-checked.
5459
5460
5461 Version 0.94 18-Sep-97
5462 ----------------------
5463
5464 1. The functions are now thread-safe, with the caveat that the global variables
5465 containing pointers to malloc() and free() or alternative functions are the
5466 same for all threads.
5467
5468 2. Get pcre_study() to generate a bitmap of initial characters for non-
5469 anchored patterns when this is possible, and use it if passed to pcre_exec().
5470
5471
5472 Version 0.93 15-Sep-97
5473 ----------------------
5474
5475 1. /(b)|(:+)/ was computing an incorrect first character.
5476
5477 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5478 but not actually doing anything yet.
5479
5480 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5481 as Perl does (e.g. [-az] or [az-]).
5482
5483 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5484 all possible positions.
5485
5486 5. Split up into different modules to avoid including unneeded functions in a
5487 compiled binary. However, compile and exec are still in one module. The "study"
5488 function is split off.
5489
5490 6. The character tables are now in a separate module whose source is generated
5491 by an auxiliary program - but can then be edited by hand if required. There are
5492 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5493 toupper() in the code.
5494
5495 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5496 make them global. Abolish the function for setting them, as the caller can now
5497 set them directly.
5498
5499
5500 Version 0.92 11-Sep-97
5501 ----------------------
5502
5503 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5504 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5505
5506 2. Caseless matching was not working in character classes if the characters in
5507 the pattern were in upper case.
5508
5509 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5510
5511 4. Make PCRE_ANCHORED public and accept as a compile option.
5512
5513 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5514 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5515 pass them.
5516
5517 6. Give an error if bad option bits passed at compile or run time.
5518
5519 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5520 pcretest to cause it to pass that flag.
5521
5522 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5523 options, and the first character, if set.
5524
5525 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5526
5527
5528 Version 0.91 10-Sep-97
5529 ----------------------
5530
5531 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5532 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5533
5534 2. PCRE was looping on encountering an indefinitely repeated back reference to
5535 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5536 Perl does - treats the match as successful.
5537
5538 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5