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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1477 - (show annotations)
Wed May 21 17:53:49 2014 UTC (5 years, 9 months ago) by ph10
File size: 265141 byte(s)
Error occurred while calculating annotation data.
Casts and type changes for compiler warnings.
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
30 Version 8.35 04-April-2014
31 --------------------------
32
33 1. A new flag is set, when property checks are present in an XCLASS.
34 When this flag is not set, PCRE can perform certain optimizations
35 such as studying these XCLASS-es.
36
37 2. The auto-possessification of character sets were improved: a normal
38 and an extended character set can be compared now. Furthermore
39 the JIT compiler optimizes more character set checks.
40
41 3. Got rid of some compiler warnings for potentially uninitialized variables
42 that show up only when compiled with -O2.
43
44 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
45 of a match later then the end of the match. The pcretest program was not
46 handling the case sensibly - it was outputting from the start to the next
47 binary zero. It now reports this situation in a message, and outputs the
48 text from the end to the start.
49
50 5. Fast forward search is improved in JIT. Instead of the first three
51 characters, any three characters with fixed position can be searched.
52 Search order: first, last, middle.
53
54 6. Improve character range checks in JIT. Characters are read by an inprecise
55 function now, which returns with an unknown value if the character code is
56 above a certain treshold (e.g: 256). The only limitation is that the value
57 must be bigger than the treshold as well. This function is useful, when
58 the characters above the treshold are handled in the same way.
59
60 7. The macros whose names start with RAWUCHAR are placeholders for a future
61 mode in which only the bottom 21 bits of 32-bit data items are used. To
62 make this more memorable for those maintaining the code, the names have
63 been changed to start with UCHAR21, and an extensive comment has been added
64 to their definition.
65
66 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
67 to the export list in Makefile.am (they were accidentally omitted from the
68 8.34 tarball).
69
70 9. The informational output from pcretest used the phrase "starting byte set"
71 which is inappropriate for the 16-bit and 32-bit libraries. As the output
72 for "first char" and "need char" really means "non-UTF-char", I've changed
73 "byte" to "char", and slightly reworded the output. The documentation about
74 these values has also been (I hope) clarified.
75
76 10. Another JIT related optimization: use table jumps for selecting the correct
77 backtracking path, when more than four alternatives are present inside a
78 bracket.
79
80 11. Empty match is not possible, when the minimum length is greater than zero,
81 and there is no \K in the pattern. JIT should avoid empty match checks in
82 such cases.
83
84 12. In a caseless character class with UCP support, when a character with more
85 than one alternative case was not the first character of a range, not all
86 the alternative cases were added to the class. For example, s and \x{17f}
87 are both alternative cases for S: the class [RST] was handled correctly,
88 but [R-T] was not.
89
90 13. The configure.ac file always checked for pthread support when JIT was
91 enabled. This is not used in Windows, so I have put this test inside a
92 check for the presence of windows.h (which was already tested for).
93
94 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
95 The algorithm provides a way to skip certain starting offsets, and usually
96 faster than linear prefix searches.
97
98 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
99 as for 'fr_FR' and 'french'. For some reason, however, it then used the
100 Windows-specific input and output files, which have 'french' screwed in.
101 So this could never have worked. One of the problems with locales is that
102 they aren't always the same. I have now updated RunTest so that it checks
103 the output of the locale test (test 3) against three different output
104 files, and it allows the test to pass if any one of them matches. With luck
105 this should make the test pass on some versions of Solaris where it was
106 failing. Because of the uncertainty, the script did not used to stop if
107 test 3 failed; it now does. If further versions of a French locale ever
108 come to light, they can now easily be added.
109
110 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
111 there was a message during ./configure, but it did not stop. This now
112 provokes an error. The invalid example in README has been corrected.
113 If a value less than the minimum is given, the minimum value has always
114 been used, but now a warning is given.
115
116 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
117 was a bug in the test system, which is now fixed. Also, the list of various
118 configurations that are tested for each release did not have one with both
119 16/32 bits and --enable-bar-anycrlf. It now does.
120
121 18. pcretest was missing "-C bsr" for displaying the \R default setting.
122
123 19. Little endian PowerPC systems are supported now by the JIT compiler.
124
125 20. The fast forward newline mechanism could enter to an infinite loop on
126 certain invalid UTF-8 input. Although we don't support these cases
127 this issue can be fixed by a performance optimization.
128
129 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
130 not take account if existing stack usage. There is now a new global
131 variable called pcre_stack_guard that can be set to point to an external
132 function to check stack availability. It is called at the start of
133 processing every parenthesized group.
134
135 22. A typo in the code meant that in ungreedy mode the max/min qualifier
136 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
137 not match "ab".
138
139 23. When UTF was disabled, the JIT program reported some incorrect compile
140 errors. These messages are silenced now.
141
142 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
143 compiler.
144
145 25. Change all the temporary files used in RunGrepTest to be different to those
146 used by RunTest so that the tests can be run simultaneously, for example by
147 "make -j check".
148
149
150 Version 8.34 15-December-2013
151 -----------------------------
152
153 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
154 executable memory. Patch inspired by Carsten Klein.
155
156 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
157 this macro is never tested and has no effect, because the work to support
158 coverage involves only compiling and linking options and special targets in
159 the Makefile. The comment in config.h implied that defining the macro would
160 enable coverage support, which is totally false. There was also support for
161 setting this macro in the CMake files (my fault, I just copied it from
162 configure). SUPPORT_GCOV has now been removed.
163
164 3. Make a small performance improvement in strlen16() and strlen32() in
165 pcretest.
166
167 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
168 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
169 these cases). There was also one in pcretest.c.
170
171 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
172
173 6. In UTF mode, the code for checking whether a group could match an empty
174 string (which is used for indefinitely repeated groups to allow for
175 breaking an infinite loop) was broken when the group contained a repeated
176 negated single-character class with a character that occupied more than one
177 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
178 UTF-8 mode). The effect was undefined: the group might or might not be
179 deemed as matching an empty string, or the program might have crashed.
180
181 7. The code for checking whether a group could match an empty string was not
182 recognizing that \h, \H, \v, \V, and \R must match a character.
183
184 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
185 an empty string. If it can, pcretest shows this in its information output.
186
187 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
188 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
189 matched by pcre_exec() without using JIT:
190
191 (a) If the rest of the pattern did not match after a maximal run of
192 grapheme clusters, the code for backing up to try with fewer of them
193 did not always back up over a full grapheme when characters that do not
194 have the modifier quality were involved, e.g. Hangul syllables.
195
196 (b) If the match point in a subject started with modifier character, and
197 there was no match, the code could incorrectly back up beyond the match
198 point, and potentially beyond the first character in the subject,
199 leading to a segfault or an incorrect match result.
200
201 10. A conditional group with an assertion condition could lead to PCRE
202 recording an incorrect first data item for a match if no other first data
203 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
204 first data item, and therefore matched "ca" after "c" instead of at the
205 start.
206
207 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
208 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
209
210 12. The source of pcregrep now includes z/OS-specific code so that it can be
211 compiled for z/OS as part of the special z/OS distribution.
212
213 13. Added the -T and -TM options to pcretest.
214
215 14. The code in pcre_compile.c for creating the table of named capturing groups
216 has been refactored. Instead of creating the table dynamically during the
217 actual compiling pass, the information is remembered during the pre-compile
218 pass (on the stack unless there are more than 20 named groups, in which
219 case malloc() is used) and the whole table is created before the actual
220 compile happens. This has simplified the code (it is now nearly 150 lines
221 shorter) and prepared the way for better handling of references to groups
222 with duplicate names.
223
224 15. A back reference to a named subpattern when there is more than one of the
225 same name now checks them in the order in which they appear in the pattern.
226 The first one that is set is used for the reference. Previously only the
227 first one was inspected. This change makes PCRE more compatible with Perl.
228
229 16. Unicode character properties were updated from Unicode 6.3.0.
230
231 17. The compile-time code for auto-possessification has been refactored, based
232 on a patch by Zoltan Herczeg. It now happens after instead of during
233 compilation. The code is cleaner, and more cases are handled. The option
234 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
235 options in pcretest are provided to set it. It can also be set by
236 (*NO_AUTO_POSSESS) at the start of a pattern.
237
238 18. The character VT has been added to the default ("C" locale) set of
239 characters that match \s and are generally treated as white space,
240 following this same change in Perl 5.18. There is now no difference between
241 "Perl space" and "POSIX space". Whether VT is treated as white space in
242 other locales depends on the locale.
243
244 19. The code for checking named groups as conditions, either for being set or
245 for being recursed, has been refactored (this is related to 14 and 15
246 above). Processing unduplicated named groups should now be as fast at
247 numerical groups, and processing duplicated groups should be faster than
248 before.
249
250 20. Two patches to the CMake build system, by Alexander Barkov:
251
252 (1) Replace the "source" command by "." in CMakeLists.txt because
253 "source" is a bash-ism.
254
255 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
256 without these the CMake build does not work on Solaris.
257
258 21. Perl has changed its handling of \8 and \9. If there is no previously
259 encountered capturing group of those numbers, they are treated as the
260 literal characters 8 and 9 instead of a binary zero followed by the
261 literals. PCRE now does the same.
262
263 22. Following Perl, added \o{} to specify codepoints in octal, making it
264 possible to specify values greater than 0777 and also making them
265 unambiguous.
266
267 23. Perl now gives an error for missing closing braces after \x{... instead of
268 treating the string as literal. PCRE now does the same.
269
270 24. RunTest used to grumble if an inappropriate test was selected explicitly,
271 but just skip it when running all tests. This make it awkward to run ranges
272 of tests when one of them was inappropriate. Now it just skips any
273 inappropriate tests, as it always did when running all tests.
274
275 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
276 character types such as \d or \w, too many callouts were inserted, and the
277 data that they returned was rubbish.
278
279 26. In UCP mode, \s was not matching two of the characters that Perl matches,
280 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
281 were matched by \h. The code has now been refactored so that the lists of
282 the horizontal and vertical whitespace characters used for \h and \v (which
283 are defined only in one place) are now also used for \s.
284
285 27. Add JIT support for the 64 bit TileGX architecture.
286 Patch by Jiong Wang (Tilera Corporation).
287
288 28. Possessive quantifiers for classes (both explicit and automatically
289 generated) now use special opcodes instead of wrapping in ONCE brackets.
290
291 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
292 (because it's meaningless), this was not happening when PCRE_CASELESS was
293 set. Not wrong, but inefficient.
294
295 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
296 properties for \w, \d, etc) is present in a test regex. Otherwise if the
297 test contains no characters greater than 255, Perl doesn't realise it
298 should be using Unicode semantics.
299
300 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
301 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
302 does in Unicode mode.
303
304 32. Added the "forbid" facility to pcretest so that putting tests into the
305 wrong test files can sometimes be quickly detected.
306
307 33. There is now a limit (default 250) on the depth of nesting of parentheses.
308 This limit is imposed to control the amount of system stack used at compile
309 time. It can be changed at build time by --with-parens-nest-limit=xxx or
310 the equivalent in CMake.
311
312 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
313 errors. Perl warns for these when in warning mode, but PCRE has no facility
314 for giving warnings.
315
316 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
317 However, this was not working for (?!) because it is optimized to (*FAIL),
318 for which PCRE does not allow quantifiers. The optimization is now disabled
319 when a quantifier follows (?!). I can't see any use for this, but it makes
320 things uniform.
321
322 36. Perl no longer allows group names to start with digits, so I have made this
323 change also in PCRE. It simplifies the code a bit.
324
325 37. In extended mode, Perl ignores spaces before a + that indicates a
326 possessive quantifier. PCRE allowed a space before the quantifier, but not
327 before the possessive +. It now does.
328
329 38. The use of \K (reset reported match start) within a repeated possessive
330 group such as (a\Kb)*+ was not working.
331
332 40. Document that the same character tables must be used at compile time and
333 run time, and that the facility to pass tables to pcre_exec() and
334 pcre_dfa_exec() is for use only with saved/restored patterns.
335
336 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
337 features for Builds with MSVC:
338
339 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
340 builds). The .rc files can be used to set FileDescription and many other
341 attributes.
342
343 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
344 This allows higher-level build scripts which want .pdb files to avoid
345 hard-coding the exact files needed."
346
347 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
348 mean "start of word" and "end of word", respectively, as a transition aid.
349
350 43. A minimizing repeat of a class containing codepoints greater than 255 in
351 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
352 compiled to use the heap for recursion.
353
354 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
355 is configured.
356
357
358 Version 8.33 28-May-2013
359 ------------------------
360
361 1. Added 'U' to some constants that are compared to unsigned integers, to
362 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
363 variables that are added to signed variables, to ensure the result is
364 signed and can be negated.
365
366 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
367 CMake config files.
368
369 3. Revise the creation of config.h.generic so that all boolean macros are
370 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
371 overriding via -D on the command line possible.
372
373 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
374 to unsigned int is reported to make a quite noticeable speed difference in
375 a specific Windows environment. Testing on Linux did also appear to show
376 some benefit (and it is clearly not harmful). Also fixed the definition of
377 Xop which should be unsigned.
378
379 5. Related to (4), changing the definition of the intermediate variable cc
380 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
381 improvements.
382
383 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
384 unnecessary spaces.
385
386 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
387 and later.
388
389 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
390
391 9. Optimizing fast_forward_start_bits in JIT.
392
393 10. Adding support for callouts in JIT, and fixing some issues revealed
394 during this work. Namely:
395
396 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
397
398 (b) Minimum length was not checked before the matching is started.
399
400 11. The value of capture_last that is passed to callouts was incorrect in some
401 cases when there was a capture on one path that was subsequently abandoned
402 after a backtrack. Also, the capture_last value is now reset after a
403 recursion, since all captures are also reset in this case.
404
405 12. The interpreter no longer returns the "too many substrings" error in the
406 case when an overflowing capture is in a branch that is subsequently
407 abandoned after a backtrack.
408
409 13. In the pathological case when an offset vector of size 2 is used, pcretest
410 now prints out the matched string after a yield of 0 or 1.
411
412 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
413 Only supported by the JIT compiler at the moment.
414
415 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
416
417 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
418 offset of the starting point of the matching process, provided the offsets
419 vector is large enough.
420
421 17. The \A escape now records a lookbehind value of 1, though its execution
422 does not actually inspect the previous character. This is to ensure that,
423 in partial multi-segment matching, at least one character from the old
424 segment is retained when a new segment is processed. Otherwise, if there
425 are no lookbehinds in the pattern, \A might match incorrectly at the start
426 of a new segment.
427
428 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
429
430 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
431 gives some modest performance improvement in 8-bit mode.
432
433 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
434 be expressed in certain programming languages using Universal Character
435 Names.
436
437 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
438 which points out that "non characters" are not "characters that may not
439 appear in Unicode strings" but rather "characters that are reserved for
440 internal use and have only local meaning".
441
442 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
443 there was a conditional group that depended on an assertion, if the
444 assertion was false, the callout that immediately followed the alternation
445 in the condition was skipped when pcre_exec() was used for matching.
446
447 23. Allow an explicit callout to be inserted before an assertion that is the
448 condition for a conditional group, for compatibility with automatic
449 callouts, which always insert a callout at this point.
450
451 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
452 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
453
454 25. (*PRUNE) is now supported by the JIT compiler.
455
456 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
457
458 27. Fix the case where there are two or more SKIPs with arguments that may be
459 ignored.
460
461 28. (*SKIP) is now supported by the JIT compiler.
462
463 29. (*THEN) is now supported by the JIT compiler.
464
465 30. Update RunTest with additional test selector options.
466
467 31. The way PCRE handles backtracking verbs has been changed in two ways.
468
469 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
470 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
471 backtracking. In some cases this makes it more Perl-compatible, but Perl's
472 rather obscure rules do not always do the same thing.
473
474 (2) Previously, backtracking verbs were confined within assertions. This is
475 no longer the case for positive assertions, except for (*ACCEPT). Again,
476 this sometimes improves Perl compatibility, and sometimes does not.
477
478 32. A number of tests that were in test 2 because Perl did things differently
479 have been moved to test 1, because either Perl or PCRE has changed, and
480 these tests are now compatible.
481
482 32. Backtracking control verbs are now handled in the same way in JIT and
483 interpreter.
484
485 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
486 contained a forward subroutine reference caused a compile error.
487
488 34. Auto-detect and optimize limited repetitions in JIT.
489
490 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
491 blocking (*UTF) etc.
492
493 36. In the interpreter, maximizing pattern repetitions for characters and
494 character types now use tail recursion, which reduces stack usage.
495
496 37. The value of the max lookbehind was not correctly preserved if a compiled
497 and saved regex was reloaded on a host of different endianness.
498
499 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
500 of the compiled pattern block, expand the flags field from 16 to 32 bits
501 because it was almost full.
502
503 39. Try madvise first before posix_madvise.
504
505 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
506 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
507 original change locked out all matches of empty strings. This has been
508 changed so that one match of an empty string per line is recognized.
509 Subsequent searches on the same line (for colouring or for --only-matching,
510 for example) do not recognize empty strings.
511
512 41. Applied a user patch to fix a number of spelling mistakes in comments.
513
514 42. Data lines longer than 65536 caused pcretest to crash.
515
516 43. Clarified the data type for length and startoffset arguments for pcre_exec
517 and pcre_dfa_exec in the function-specific man pages, where they were
518 explicitly stated to be in bytes, never having been updated. I also added
519 some clarification to the pcreapi man page.
520
521 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
522 a segmentation fault.
523
524
525 Version 8.32 30-November-2012
526 -----------------------------
527
528 1. Improved JIT compiler optimizations for first character search and single
529 character iterators.
530
531 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
532 Patch by Daniel Richard G.
533
534 3. Single character iterator optimizations in the JIT compiler.
535
536 4. Improved JIT compiler optimizations for character ranges.
537
538 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
539 Reported by Giuseppe D'Angelo.
540
541 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
542 of a line, was being set incorrectly in cases where .* appeared inside
543 atomic brackets at the start of a pattern, or where there was a subsequent
544 *PRUNE or *SKIP.
545
546 7. Improved instruction cache flush for POWER/PowerPC.
547 Patch by Daniel Richard G.
548
549 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
550 grep:
551
552 (a) There is now no limit to the number of patterns to be matched.
553
554 (b) An error is given if a pattern is too long.
555
556 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
557 are now supported.
558
559 (d) --exclude-from and --include-from (multiple use) have been added.
560
561 (e) Exclusions and inclusions now apply to all files and directories, not
562 just to those obtained from scanning a directory recursively.
563
564 (f) Multiple uses of -f and --file-list are now supported.
565
566 (g) In a Windows environment, the default for -d has been changed from
567 "read" (the GNU grep default) to "skip", because otherwise the presence
568 of a directory in the file list provokes an error.
569
570 (h) The documentation has been revised and clarified in places.
571
572 9. Improve the matching speed of capturing brackets.
573
574 10. Changed the meaning of \X so that it now matches a Unicode extended
575 grapheme cluster.
576
577 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
578 out POSIX threads when JIT support is configured.
579
580 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
581
582 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
583 structure could go wrong in environments where size_t is not the same size
584 as int.
585
586 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
587 set.
588
589 15. The EBCDIC support had decayed; later updates to the code had included
590 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
591 general tidy up of EBCDIC-related issues, and the documentation was also
592 not quite right. There is now a test that can be run on ASCII systems to
593 check some of the EBCDIC-related things (but is it not a full test).
594
595 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
596 in a small tidy to the code.
597
598 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
599
600 18. If the --only-matching (-o) option in pcregrep is specified multiple
601 times, each one causes appropriate output. For example, -o1 -o2 outputs the
602 substrings matched by the 1st and 2nd capturing parentheses. A separating
603 string can be specified by --om-separator (default empty).
604
605 19. Improving the first n character searches.
606
607 20. Turn case lists for horizontal and vertical white space into macros so that
608 they are defined only once.
609
610 21. This set of changes together give more compatible Unicode case-folding
611 behaviour for characters that have more than one other case when UCP
612 support is available.
613
614 (a) The Unicode property table now has offsets into a new table of sets of
615 three or more characters that are case-equivalent. The MultiStage2.py
616 script that generates these tables (the pcre_ucd.c file) now scans
617 CaseFolding.txt instead of UnicodeData.txt for character case
618 information.
619
620 (b) The code for adding characters or ranges of characters to a character
621 class has been abstracted into a generalized function that also handles
622 case-independence. In UTF-mode with UCP support, this uses the new data
623 to handle characters with more than one other case.
624
625 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
626 whose other case is greater than 256 are now correctly matched
627 caselessly. Previously, the high codepoint matched the low one, but not
628 vice versa.
629
630 (d) The processing of \h, \H, \v, and \ in character classes now makes use
631 of the new class addition function, using character lists defined as
632 macros alongside the case definitions of 20 above.
633
634 (e) Caseless back references now work with characters that have more than
635 one other case.
636
637 (f) General caseless matching of characters with more than one other case
638 is supported.
639
640 22. Unicode character properties were updated from Unicode 6.2.0
641
642 23. Improved CMake support under Windows. Patch by Daniel Richard G.
643
644 24. Add support for 32-bit character strings, and UTF-32
645
646 25. Major JIT compiler update (code refactoring and bugfixing).
647 Experimental Sparc 32 support is added.
648
649 26. Applied a modified version of Daniel Richard G's patch to create
650 pcre.h.generic and config.h.generic by "make" instead of in the
651 PrepareRelease script.
652
653 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
654 pcre_compile.c when checking for a zero character.
655
656 28. Introducing a native interface for JIT. Through this interface, the compiled
657 machine code can be directly executed. The purpose of this interface is to
658 provide fast pattern matching, so several sanity checks are not performed.
659 However, feature tests are still performed. The new interface provides
660 1.4x speedup compared to the old one.
661
662 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
663 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
664 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
665
666 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
667 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
668 the "old" RFC 2279). Instead, it ended up passing a negative length to
669 pcre_exec().
670
671 31. Add support for GCC's visibility feature to hide internal functions.
672
673 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
674 "unknown -C option" after outputting 0 or 1.
675
676 33. There is now support for generating a code coverage report for the test
677 suite in environments where gcc is the compiler and lcov is installed. This
678 is mainly for the benefit of the developers.
679
680 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
681 unaddressable using valgrind annotations, allowing valgrind to detect
682 invalid memory accesses. This is mainly for the benefit of the developers.
683
684 25. (*UTF) can now be used to start a pattern in any of the three libraries.
685
686 26. Give configure error if --enable-cpp but no C++ compiler found.
687
688
689 Version 8.31 06-July-2012
690 -------------------------
691
692 1. Fixing a wrong JIT test case and some compiler warnings.
693
694 2. Removed a bashism from the RunTest script.
695
696 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
697 to unsigned type, result still unsigned" that was given by an MS compiler
698 on encountering the code "-sizeof(xxx)".
699
700 4. Partial matching support is added to the JIT compiler.
701
702 5. Fixed several bugs concerned with partial matching of items that consist
703 of more than one character:
704
705 (a) /^(..)\1/ did not partially match "aba" because checking references was
706 done on an "all or nothing" basis. This also applied to repeated
707 references.
708
709 (b) \R did not give a hard partial match if \r was found at the end of the
710 subject.
711
712 (c) \X did not give a hard partial match after matching one or more
713 characters at the end of the subject.
714
715 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
716 a partial match for the string "\r".
717
718 (e) When newline was set to CRLF, the metacharacter "." did not recognize
719 a partial match for a CR character at the end of the subject string.
720
721 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
722 running pcretest, the text "(JIT)" added to the output whenever JIT is
723 actually used to run the match.
724
725 7. Individual JIT compile options can be set in pcretest by following -s+[+]
726 or /S+[+] with a digit between 1 and 7.
727
728 8. OP_NOT now supports any UTF character not just single-byte ones.
729
730 9. (*MARK) control verb is now supported by the JIT compiler.
731
732 10. The command "./RunTest list" lists the available tests without actually
733 running any of them. (Because I keep forgetting what they all are.)
734
735 11. Add PCRE_INFO_MAXLOOKBEHIND.
736
737 12. Applied a (slightly modified) user-supplied patch that improves performance
738 when the heap is used for recursion (compiled with --disable-stack-for-
739 recursion). Instead of malloc and free for each heap frame each time a
740 logical recursion happens, frames are retained on a chain and re-used where
741 possible. This sometimes gives as much as 30% improvement.
742
743 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
744 call.
745
746 14. As documented, (*COMMIT) is now confined to within a positive assertion.
747
748 15. It is now possible to link pcretest with libedit as an alternative to
749 libreadline.
750
751 16. (*COMMIT) control verb is now supported by the JIT compiler.
752
753 17. The Unicode data tables have been updated to Unicode 6.1.0.
754
755 18. Added --file-list option to pcregrep.
756
757 19. Added binary file support to pcregrep, including the -a, --binary-files,
758 -I, and --text options.
759
760 20. The madvise function is renamed for posix_madvise for QNX compatibility
761 reasons. Fixed by Giuseppe D'Angelo.
762
763 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
764
765 22. Changed the option for creating symbolic links for 16-bit man pages from
766 -s to -sf so that re-installing does not cause issues.
767
768 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
769
770 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
771 matches in certain environments (the workspace was not being correctly
772 retained). Also added to pcre_dfa_exec() a simple plausibility check on
773 some of the workspace data at the beginning of a restart.
774
775 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
776 was not doing so when it should - probably a typo introduced by SVN 528
777 (change 8.10/14).
778
779 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
780 \w+ when the character tables indicated that \x{c4} was a word character.
781 There were several related cases, all because the tests for doing a table
782 lookup were testing for characters less than 127 instead of 255.
783
784 27. If a pattern contains capturing parentheses that are not used in a match,
785 their slots in the ovector are set to -1. For those that are higher than
786 any matched groups, this happens at the end of processing. In the case when
787 there were back references that the ovector was too small to contain
788 (causing temporary malloc'd memory to be used during matching), and the
789 highest capturing number was not used, memory off the end of the ovector
790 was incorrectly being set to -1. (It was using the size of the temporary
791 memory instead of the true size.)
792
793 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
794 ovector size, it uses memory at the end of the block that it has got.
795
796 29. Check for an overlong MARK name and give an error at compile time. The
797 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
798
799 30. JIT compiler update.
800
801 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
802 Rill for the patch.
803
804 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
805
806 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
807
808 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
809 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
810 was enabled.
811
812 35. Improve JIT code generation for greedy plus quantifier.
813
814 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
815 1 to "aa" instead of to an empty string. The bug affected repeated groups
816 that could potentially match an empty string.
817
818 37. Optimizing single character iterators in JIT.
819
820 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
821 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
822 codepoints that are too big for the mode are faulted, and in a UTF mode,
823 disallowed codepoints are also faulted.
824
825 39. If PCRE was compiled with UTF support, in three places in the DFA
826 matcher there was code that should only have been obeyed in UTF mode, but
827 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
828 processing when bytes with values greater than 127 were present. In 16-bit
829 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
830 both cases the values are those that cannot be the first data item in a UTF
831 character. The three items that might have provoked this were recursions,
832 possessively repeated groups, and atomic groups.
833
834 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
835 and pcregrep, because some OS require shared objects to be explicitly
836 passed to ld, causing the link step to fail if they are not.
837
838 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
839 mode, patterns that started with \h* or \R* might be incorrectly matched.
840
841
842 Version 8.30 04-February-2012
843 -----------------------------
844
845 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
846 name is defined in ctype.h.
847
848 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
849 only in quite long subpatterns.
850
851 3. Removed the function pcre_info(), which has been obsolete and deprecated
852 since it was replaced by pcre_fullinfo() in February 2000.
853
854 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
855 match a (*MARK), and the match failed at the start of the subject, a
856 reference to memory before the start of the subject could occur. This bug
857 was introduced by fix 17 of release 8.21.
858
859 5. A reference to an unset group with zero minimum repetition was giving
860 totally wrong answers (in non-JavaScript-compatibility mode). For example,
861 /(another)?(\1?)test/ matched against "hello world test". This bug was
862 introduced in release 8.13.
863
864 6. Add support for 16-bit character strings (a large amount of work involving
865 many changes and refactorings).
866
867 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
868 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
869 from a file.
870
871 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
872 rounding is not applied in this particular case).
873
874 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
875 if they appear, or are escaped, in patterns.
876
877 10. Get rid of a number of -Wunused-but-set-variable warnings.
878
879 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
880 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
881 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
882 also returns the mark "x". This bug applied to capturing parentheses,
883 non-capturing parentheses, and atomic parentheses. It also applied to some
884 assertions.
885
886 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
887 information out of configure.ac instead of relying on pcre.h.generic, which
888 is not stored in the repository.
889
890 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
891 -lreadline.
892
893 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
894
895 15. Applied Graycode's patch to put the top-level frame on the stack rather
896 than the heap when not using the stack for recursion. This gives a
897 performance improvement in many cases when recursion is not deep.
898
899 16. Experimental code added to "pcretest -C" to output the stack frame size.
900
901
902 Version 8.21 12-Dec-2011
903 ------------------------
904
905 1. Updating the JIT compiler.
906
907 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
908 are added as well.
909
910 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
911 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
912 calling _pcre_jit_exec. Some extra comments are added.
913
914 4. (*MARK) settings inside atomic groups that do not contain any capturing
915 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
916 was introduced by change 18 for 8.20.
917
918 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
919 ECMA-262 standard.
920
921 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
922 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
923 This bug was probably introduced by change 9 of 8.13.
924
925 7. While fixing 6 above, I noticed that a number of other items were being
926 incorrectly rejected as "not fixed length". This arose partly because newer
927 opcodes had not been added to the fixed-length checking code. I have (a)
928 corrected the bug and added tests for these items, and (b) arranged for an
929 error to occur if an unknown opcode is encountered while checking for fixed
930 length instead of just assuming "not fixed length". The items that were
931 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
932 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
933 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
934
935 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
936 being incorrectly compiled and would have given unpredicatble results.
937
938 9. A possessively repeated subpattern with minimum repeat count greater than
939 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
940 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
941 the first (A) could occur when it should not.
942
943 10. Add a cast and remove a redundant test from the code.
944
945 11. JIT should use pcre_malloc/pcre_free for allocation.
946
947 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
948 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
949 is anything other than /usr, -L is still shown).
950
951 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
952
953 14. Perl does not support \N without a following name in a [] class; PCRE now
954 also gives an error.
955
956 15. If a forward reference was repeated with an upper limit of around 2000,
957 it caused the error "internal error: overran compiling workspace". The
958 maximum number of forward references (including repeats) was limited by the
959 internal workspace, and dependent on the LINK_SIZE. The code has been
960 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
961 the default depends on LINK_SIZE. There is a new upper limit (for safety)
962 of around 200,000 forward references. While doing this, I also speeded up
963 the filling in of repeated forward references.
964
965 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
966 incorrectly expecting the subject to contain another "a" after the start.
967
968 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
969 in the match, the SKIP should be ignored. This was not happening; instead
970 the SKIP was being treated as NOMATCH. For patterns such as
971 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
972 tested.
973
974 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
975 now much more compatible with Perl, in particular in cases where the result
976 is a non-match for a non-anchored pattern. For example, if
977 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
978 "m", where previously it did not return a name. A side effect of this
979 change is that for partial matches, the last encountered mark name is
980 returned, as for non matches. A number of tests that were previously not
981 Perl-compatible have been moved into the Perl-compatible test files. The
982 refactoring has had the pleasing side effect of removing one argument from
983 the match() function, thus reducing its stack requirements.
984
985 19. If the /S+ option was used in pcretest to study a pattern using JIT,
986 subsequent uses of /S (without +) incorrectly behaved like /S+.
987
988 21. Retrieve executable code size support for the JIT compiler and fixing
989 some warnings.
990
991 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
992 not work when the shorter character appeared right at the end of the
993 subject string.
994
995 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
996 systems.
997
998 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
999 output it when the /M option is used in pcretest.
1000
1001 25. The CheckMan script was not being included in the distribution. Also, added
1002 an explicit "perl" to run Perl scripts from the PrepareRelease script
1003 because this is reportedly needed in Windows.
1004
1005 26. If study data was being save in a file and studying had not found a set of
1006 "starts with" bytes for the pattern, the data written to the file (though
1007 never used) was taken from uninitialized memory and so caused valgrind to
1008 complain.
1009
1010 27. Updated RunTest.bat as provided by Sheri Pierce.
1011
1012 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1013
1014 29. Computation of memory usage for the table of capturing group names was
1015 giving an unnecessarily large value.
1016
1017
1018 Version 8.20 21-Oct-2011
1019 ------------------------
1020
1021 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1022 a POSIX class. After further experiments with Perl, which convinced me that
1023 Perl has bugs and confusions, a closing square bracket is no longer allowed
1024 in a POSIX name. This bug also affected patterns with classes that started
1025 with full stops.
1026
1027 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1028 captured substring, but while checking the failing first alternative,
1029 substring 1 is temporarily captured. If the output vector supplied to
1030 pcre_exec() was not big enough for this capture, the yield of the function
1031 was still zero ("insufficient space for captured substrings"). This cannot
1032 be totally fixed without adding another stack variable, which seems a lot
1033 of expense for a edge case. However, I have improved the situation in cases
1034 such as /(a)(b)x|abc/ matched against "abc", where the return code
1035 indicates that fewer than the maximum number of slots in the ovector have
1036 been set.
1037
1038 3. Related to (2) above: when there are more back references in a pattern than
1039 slots in the output vector, pcre_exec() uses temporary memory during
1040 matching, and copies in the captures as far as possible afterwards. It was
1041 using the entire output vector, but this conflicts with the specification
1042 that only 2/3 is used for passing back captured substrings. Now it uses
1043 only the first 2/3, for compatibility. This is, of course, another edge
1044 case.
1045
1046 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1047 main code base, and can be used by building with --enable-jit. When this is
1048 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1049 runtime --no-jit option is given.
1050
1051 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1052 ovector, the return from the function was zero, implying that there were
1053 other matches that did not fit. The correct "exactly full" value is now
1054 returned.
1055
1056 6. If a subpattern that was called recursively or as a subroutine contained
1057 (*PRUNE) or any other control that caused it to give a non-standard return,
1058 invalid errors such as "Error -26 (nested recursion at the same subject
1059 position)" or even infinite loops could occur.
1060
1061 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1062 computing the minimum length on reaching *ACCEPT, and so ended up with the
1063 wrong value of 1 rather than 0. Further investigation indicates that
1064 computing a minimum subject length in the presence of *ACCEPT is difficult
1065 (think back references, subroutine calls), and so I have changed the code
1066 so that no minimum is registered for a pattern that contains *ACCEPT.
1067
1068 8. If (*THEN) was present in the first (true) branch of a conditional group,
1069 it was not handled as intended. [But see 16 below.]
1070
1071 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1072 Sheri Pierce.
1073
1074 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1075 the first byte in a match must be "a".
1076
1077 11. Change 17 for 8.13 increased the recursion depth for patterns like
1078 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1079 pattern contains any instances of (*THEN). If it does not, the old
1080 optimizations are restored. It would be nice to do this on a per-group
1081 basis, but at the moment that is not feasible.
1082
1083 12. In some environments, the output of pcretest -C is CRLF terminated. This
1084 broke RunTest's code that checks for the link size. A single white space
1085 character after the value is now allowed for.
1086
1087 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1088 For "fr", it uses the Windows-specific input and output files.
1089
1090 14. If (*THEN) appeared in a group that was called recursively or as a
1091 subroutine, it did not work as intended. [But see next item.]
1092
1093 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1094 pattern fragments (but not containing any | characters). If A and B are
1095 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1096 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1097 D. In other words, Perl considers parentheses that do not contain any |
1098 characters to be part of a surrounding alternative, whereas PCRE was
1099 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1100 differently. PCRE now behaves in the same way as Perl, except in the case
1101 of subroutine/recursion calls such as (?1) which have in any case always
1102 been different (but PCRE had them first :-).
1103
1104 16. Related to 15 above: Perl does not treat the | in a conditional group as
1105 creating alternatives. Such a group is treated in the same way as an
1106 ordinary group without any | characters when processing (*THEN). PCRE has
1107 been changed to match Perl's behaviour.
1108
1109 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1110 RunGrepTest script failed.
1111
1112 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1113 inevitable for groups that contain captures, but it can lead to a lot of
1114 stack use in large patterns. The old behaviour has been restored for atomic
1115 groups that do not contain any capturing parentheses.
1116
1117 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1118 suppress the check for a minimum subject length at run time. (If it was
1119 given to pcre_exec() or pcre_dfa_exec() it did work.)
1120
1121 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1122 fail to work when decoding hex characters in data strings in EBCDIC
1123 environments.
1124
1125 21. It appears that in at least one Mac OS environment, the isxdigit() function
1126 is implemented as a macro that evaluates to its argument more than once,
1127 contravening the C 90 Standard (I haven't checked a later standard). There
1128 was an instance in pcretest which caused it to go wrong when processing
1129 \x{...} escapes in subject strings. The has been rewritten to avoid using
1130 things like p++ in the argument of isxdigit().
1131
1132
1133 Version 8.13 16-Aug-2011
1134 ------------------------
1135
1136 1. The Unicode data tables have been updated to Unicode 6.0.0.
1137
1138 2. Two minor typos in pcre_internal.h have been fixed.
1139
1140 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1141 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1142 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1143
1144 4. There were a number of related bugs in the code for matching backrefences
1145 caselessly in UTF-8 mode when codes for the characters concerned were
1146 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1147 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1148 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1149 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1150 2-byte code at the end of the subject (it thought there wasn't enough data
1151 left).
1152
1153 5. Comprehensive information about what went wrong is now returned by
1154 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1155 as the output vector has at least 2 elements. The offset of the start of
1156 the failing character and a reason code are placed in the vector.
1157
1158 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1159 now returned is for the first byte of the failing character, instead of the
1160 last byte inspected. This is an incompatible change, but I hope it is small
1161 enough not to be a problem. It makes the returned offset consistent with
1162 pcre_exec() and pcre_dfa_exec().
1163
1164 7. pcretest now gives a text phrase as well as the error number when
1165 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1166 failure, the offset and reason code are output.
1167
1168 8. When \R was used with a maximizing quantifier it failed to skip backwards
1169 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1170 back over a single character (\n). This seems wrong (because it treated the
1171 two characters as a single entity when going forwards), conflicts with the
1172 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1173 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1174 has been changed.
1175
1176 9. Some internal refactoring has changed the processing so that the handling
1177 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1178 time (the PCRE_DOTALL option was changed this way some time ago: version
1179 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1180 which was always a bit of a fudge. It also means that there is one less
1181 argument for the match() function, which reduces its stack requirements
1182 slightly. This change also fixes an incompatibility with Perl: the pattern
1183 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1184
1185 10. More internal refactoring has drastically reduced the number of recursive
1186 calls to match() for possessively repeated groups such as (abc)++ when
1187 using pcre_exec().
1188
1189 11. While implementing 10, a number of bugs in the handling of groups were
1190 discovered and fixed:
1191
1192 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1193 (a|)*(?1) gave a compile-time internal error.
1194 ((a|)+)+ did not notice that the outer group could match an empty string.
1195 (^a|^)+ was not marked as anchored.
1196 (.*a|.*)+ was not marked as matching at start or after a newline.
1197
1198 12. Yet more internal refactoring has removed another argument from the match()
1199 function. Special calls to this function are now indicated by setting a
1200 value in a variable in the "match data" data block.
1201
1202 13. Be more explicit in pcre_study() instead of relying on "default" for
1203 opcodes that mean there is no starting character; this means that when new
1204 ones are added and accidentally left out of pcre_study(), testing should
1205 pick them up.
1206
1207 14. The -s option of pcretest has been documented for ages as being an old
1208 synonym of -m (show memory usage). I have changed it to mean "force study
1209 for every regex", that is, assume /S for every regex. This is similar to -i
1210 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1211 using it. It makes it easier to run collections of tests with and without
1212 study enabled, and thereby test pcre_study() more easily. All the standard
1213 tests are now run with and without -s (but some patterns can be marked as
1214 "never study" - see 20 below).
1215
1216 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1217 restoration of the capturing data to the outer values was not happening
1218 correctly.
1219
1220 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1221 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1222 pattern had matched an empty string, and so incorrectly returned a no
1223 match.
1224
1225 17. There was optimizing code for the last branch of non-capturing parentheses,
1226 and also for the obeyed branch of a conditional subexpression, which used
1227 tail recursion to cut down on stack usage. Unfortunately, now that there is
1228 the possibility of (*THEN) occurring in these branches, tail recursion is
1229 no longer possible because the return has to be checked for (*THEN). These
1230 two optimizations have therefore been removed. [But see 8.20/11 above.]
1231
1232 18. If a pattern containing \R was studied, it was assumed that \R always
1233 matched two bytes, thus causing the minimum subject length to be
1234 incorrectly computed because \R can also match just one byte.
1235
1236 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1237 was incorrectly computed.
1238
1239 20. If /S is present twice on a test pattern in pcretest input, it now
1240 *disables* studying, thereby overriding the use of -s on the command line
1241 (see 14 above). This is necessary for one or two tests to keep the output
1242 identical in both cases.
1243
1244 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1245 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1246
1247 22. When an atomic group that contained a capturing parenthesis was
1248 successfully matched, but the branch in which it appeared failed, the
1249 capturing was not being forgotten if a higher numbered group was later
1250 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1251 group 1 to "a", when in fact it should be unset. This applied to multi-
1252 branched capturing and non-capturing groups, repeated or not, and also to
1253 positive assertions (capturing in negative assertions does not happen
1254 in PCRE) and also to nested atomic groups.
1255
1256 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1257 subject after a captured substring, to make it easier to tell which of a
1258 number of identical substrings has been captured.
1259
1260 24. The way atomic groups are processed by pcre_exec() has been changed so that
1261 if they are repeated, backtracking one repetition now resets captured
1262 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1263 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1264 "aaa". Previously, it would have been "a". As part of this code
1265 refactoring, the way recursive calls are handled has also been changed.
1266
1267 25. If an assertion condition captured any substrings, they were not passed
1268 back unless some other capturing happened later. For example, if
1269 (?(?=(a))a) was matched against "a", no capturing was returned.
1270
1271 26. When studying a pattern that contained subroutine calls or assertions,
1272 the code for finding the minimum length of a possible match was handling
1273 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1274 group 1 called group 2 while simultaneously a separate group 2 called group
1275 1). A stack overflow occurred in this case. I have fixed this by limiting
1276 the recursion depth to 10.
1277
1278 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1279 Fortmann. This supports explicit test numbers on the command line, and has
1280 argument validation and error reporting.
1281
1282 28. An instance of \X with an unlimited repeat could fail if at any point the
1283 first character it looked at was a mark character.
1284
1285 29. Some minor code refactoring concerning Unicode properties and scripts
1286 should reduce the stack requirement of match() slightly.
1287
1288 30. Added the '=' option to pcretest to check the setting of unused capturing
1289 slots at the end of the pattern, which are documented as being -1, but are
1290 not included in the return count.
1291
1292 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1293 compiled something random. Now it gives a compile-time error (as does
1294 Perl).
1295
1296 32. A *MARK encountered during the processing of a positive assertion is now
1297 recorded and passed back (compatible with Perl).
1298
1299 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1300 had alternative anchored branches, the search for a second match in a line
1301 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1302 matched the line "0102" twice. The same bug affected patterns that started
1303 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1304 twice.
1305
1306 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1307 does, and because of capturing effects, quantifying parenthesized
1308 assertions may at times be useful. Quantifiers are now allowed for
1309 parenthesized assertions.
1310
1311 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1312
1313 36. \g was being checked for fancy things in a character class, when it should
1314 just be a literal "g".
1315
1316 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1317 appearance of a nested POSIX class supersedes an apparent external class.
1318 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1319 unescaped square brackets may also appear as part of class names. For
1320 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1321 more like Perl. (But see 8.20/1 above.)
1322
1323 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1324 was because it thought it was \N{name}, which is not supported).
1325
1326 39. Add minix to OS list not supporting the -S option in pcretest.
1327
1328 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1329 cannot analyze patterns in sufficient detail to catch mutual recursions
1330 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1331 subgroup is called recursively as a subpattern for a second time at the
1332 same position in the subject string. In previous releases this might have
1333 been caught by the recursion limit, or it might have run out of stack.
1334
1335 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1336 happen only once. PCRE was, however incorrectly giving a compile time error
1337 "recursive call could loop indefinitely" because it cannot analyze the
1338 pattern in sufficient detail. The compile time test no longer happens when
1339 PCRE is compiling a conditional subpattern, but actual runaway loops are
1340 now caught at runtime (see 40 above).
1341
1342 42. It seems that Perl allows any characters other than a closing parenthesis
1343 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1344 has been changed to be the same.
1345
1346 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1347 as not to get warnings when autogen.sh is called. Also changed
1348 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1349
1350 44. To help people who use pcregrep to scan files containing exceedingly long
1351 lines, the following changes have been made:
1352
1353 (a) The default value of the buffer size parameter has been increased from
1354 8K to 20K. (The actual buffer used is three times this size.)
1355
1356 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1357 PCRE is built.
1358
1359 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1360 to be set at run time.
1361
1362 (d) Numerical values in pcregrep options can be followed by K or M, for
1363 example --buffer-size=50K.
1364
1365 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1366 given and the return code is set to 2.
1367
1368 45. Add a pointer to the latest mark to the callout data block.
1369
1370 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1371 partial match of an empty string instead of no match. This was specific to
1372 the use of ".".
1373
1374 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1375 complete match instead of a partial match. This bug was dependent on both
1376 the PCRE_UTF8 and PCRE_DOTALL options being set.
1377
1378 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1379 starting byte set, because \b was not being ignored.
1380
1381
1382 Version 8.12 15-Jan-2011
1383 ------------------------
1384
1385 1. Fixed some typos in the markup of the man pages, and wrote a script that
1386 checks for such things as part of the documentation building process.
1387
1388 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1389 --match-limit and --recursion-limit options (added for 8.11). In
1390 particular, this made one of the standard tests fail. (The integer value
1391 went into the wrong half of a long int.)
1392
1393 3. If the --colour option was given to pcregrep with -v (invert match), it
1394 did strange things, either producing crazy output, or crashing. It should,
1395 of course, ignore a request for colour when reporting lines that do not
1396 match.
1397
1398 4. Another pcregrep bug caused similar problems if --colour was specified with
1399 -M (multiline) and the pattern match finished with a line ending.
1400
1401 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1402 matched in multiline mode, the following line was shown as part of the
1403 match. This seems wrong, so I have changed it.
1404
1405 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1406 the check for further matches in the same line (so they could be coloured)
1407 to overrun the end of the current line. If another match was found, it was
1408 incorrectly shown (and then shown again when found in the next line).
1409
1410 7. If pcregrep was compiled under Windows, there was a reference to the
1411 function pcregrep_exit() before it was defined. I am assuming this was
1412 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1413 reported by a user. I've moved the definition above the reference.
1414
1415
1416 Version 8.11 10-Dec-2010
1417 ------------------------
1418
1419 1. (*THEN) was not working properly if there were untried alternatives prior
1420 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1421 backtracked to try for "b" instead of moving to the next alternative branch
1422 at the same level (in this case, to look for "c"). The Perl documentation
1423 is clear that when (*THEN) is backtracked onto, it goes to the "next
1424 alternative in the innermost enclosing group".
1425
1426 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1427 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1428 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1429 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1430 (*THEN).
1431
1432 3. If \s appeared in a character class, it removed the VT character from
1433 the class, even if it had been included by some previous item, for example
1434 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1435 of \s, but is part of the POSIX "space" class.)
1436
1437 4. A partial match never returns an empty string (because you can always
1438 match an empty string at the end of the subject); however the checking for
1439 an empty string was starting at the "start of match" point. This has been
1440 changed to the "earliest inspected character" point, because the returned
1441 data for a partial match starts at this character. This means that, for
1442 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1443 (previously it gave "no match").
1444
1445 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1446 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1447 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1448 has an implication that the given string is incomplete (because a partial
1449 match is preferred over a full match). For this reason, these items now
1450 give a partial match in this situation. [Aside: previously, the one case
1451 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1452 match rather than a full match, which was wrong by the old rules, but is
1453 now correct.]
1454
1455 6. There was a bug in the handling of #-introduced comments, recognized when
1456 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1457 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1458 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1459 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1460 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1461 places in pcre_compile().
1462
1463 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1464 comments when looking ahead for named forward references to subpatterns,
1465 the only newline sequence it recognized was NL. It now handles newlines
1466 according to the set newline convention.
1467
1468 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1469 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1470 cater for a lack of strerror(). These oversights have been fixed.
1471
1472 9. Added --match-limit and --recursion-limit to pcregrep.
1473
1474 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1475
1476 11. When the -o option was used, pcregrep was setting a return code of 1, even
1477 when matches were found, and --line-buffered was not being honoured.
1478
1479 12. Added an optional parentheses number to the -o and --only-matching options
1480 of pcregrep.
1481
1482 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1483 can match an empty string. The code to do it in pcretest and pcredemo
1484 needed fixing:
1485
1486 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1487 only one byte after an empty string match just before CRLF (this case
1488 just got forgotten; "any" and "anycrlf" were OK).
1489
1490 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1491 mode when an empty string match preceded an ASCII character followed by
1492 a non-ASCII character. (The code for advancing by one character rather
1493 than one byte was nonsense.)
1494
1495 (c) The pcredemo.c sample program did not have any code at all to handle
1496 the cases when CRLF is a valid newline sequence.
1497
1498 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1499 as a starting offset was within the subject string. There is now a new
1500 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1501 negative or greater than the length of the string. In order to test this,
1502 pcretest is extended to allow the setting of negative starting offsets.
1503
1504 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1505 starting offset points to the beginning of a UTF-8 character was
1506 unnecessarily clumsy. I tidied it up.
1507
1508 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1509 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1510
1511 17. Nobody had reported that the --include_dir option, which was added in
1512 release 7.7 should have been called --include-dir (hyphen, not underscore)
1513 for compatibility with GNU grep. I have changed it to --include-dir, but
1514 left --include_dir as an undocumented synonym, and the same for
1515 --exclude-dir, though that is not available in GNU grep, at least as of
1516 release 2.5.4.
1517
1518 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1519 characters from a string of bytes) have been redefined so as not to use
1520 loops, in order to improve performance in some environments. At the same
1521 time, I abstracted some of the common code into auxiliary macros to save
1522 repetition (this should not affect the compiled code).
1523
1524 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1525 compile-time error is now given if \c is not followed by an ASCII
1526 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1527 different, and any byte value is allowed.)
1528
1529 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1530 START_OPTIMIZE option, which is now allowed at compile time - but just
1531 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1532 to pcregrep and other applications that have no direct access to PCRE
1533 options. The new /Y option in pcretest sets this option when calling
1534 pcre_compile().
1535
1536 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1537 back references. Groups containing recursive back references were forced to
1538 be atomic by that change, but in the case of named groups, the amount of
1539 memory required was incorrectly computed, leading to "Failed: internal
1540 error: code overflow". This has been fixed.
1541
1542 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1543 pcretest.c, to avoid build problems in some Borland environments.
1544
1545
1546 Version 8.10 25-Jun-2010
1547 ------------------------
1548
1549 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1550 THEN.
1551
1552 2. (*ACCEPT) was not working when inside an atomic group.
1553
1554 3. Inside a character class, \B is treated as a literal by default, but
1555 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1556 causes the error). The code is unchanged, but I tidied the documentation.
1557
1558 4. Inside a character class, PCRE always treated \R and \X as literals,
1559 whereas Perl faults them if its -w option is set. I have changed PCRE so
1560 that it faults them when PCRE_EXTRA is set.
1561
1562 5. Added support for \N, which always matches any character other than
1563 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1564
1565 6. When compiling pcregrep with newer versions of gcc which may have
1566 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1567 declared with attribute warn_unused_result" were given. Just casting the
1568 result to (void) does not stop the warnings; a more elaborate fudge is
1569 needed. I've used a macro to implement this.
1570
1571 7. Minor change to pcretest.c to avoid a compiler warning.
1572
1573 8. Added four artifical Unicode properties to help with an option to make
1574 \s etc use properties (see next item). The new properties are: Xan
1575 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1576
1577 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1578 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1579 this option. Modified pcretest to add /W to test this facility. Added
1580 REG_UCP to make it available via the POSIX interface.
1581
1582 10. Added --line-buffered to pcregrep.
1583
1584 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1585 studied, and the match started with a letter with a code point greater than
1586 127 whose first byte was different to the first byte of the other case of
1587 the letter, the other case of this starting letter was not recognized
1588 (#976).
1589
1590 12. If a pattern that was studied started with a repeated Unicode property
1591 test, for example, \p{Nd}+, there was the theoretical possibility of
1592 setting up an incorrect bitmap of starting bytes, but fortunately it could
1593 not have actually happened in practice until change 8 above was made (it
1594 added property types that matched character-matching opcodes).
1595
1596 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1597 possible starting bytes for non-anchored patterns.
1598
1599 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1600 \R, and also a number of cases that involve Unicode properties, both
1601 explicit and implicit when PCRE_UCP is set.
1602
1603 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1604 input, it could crash or give wrong results if characters with values
1605 greater than 0xc0 were present in the subject string. (Detail: it assumed
1606 UTF-8 input when processing these items.)
1607
1608 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1609 size_t is 64-bit (#991).
1610
1611 17. Added a check for running out of memory when PCRE is compiled with
1612 --disable-stack-for-recursion (#990).
1613
1614 18. If the last data line in a file for pcretest does not have a newline on
1615 the end, a newline was missing in the output.
1616
1617 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1618 less than 128) in its various bitmaps. However, there is a facility for
1619 generating tables according to the current locale when PCRE is compiled. It
1620 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1621 characters, are recognized by isspace() and therefore were getting set in
1622 these tables, and indeed these tables seem to approximate to ISO 8859. This
1623 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1624 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1625 which of course cannot start UTF-8 characters. I have changed the code so
1626 that only real ASCII characters (less than 128) and the correct starting
1627 bytes for UTF-8 encodings are set for characters greater than 127 when in
1628 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1629 altogether.)
1630
1631 20. Added the /T option to pcretest so as to be able to run tests with non-
1632 standard character tables, thus making it possible to include the tests
1633 used for 19 above in the standard set of tests.
1634
1635 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1636 reference to a subpattern the other side of a comment that contains an
1637 opening parenthesis caused either an internal compiling error, or a
1638 reference to the wrong subpattern.
1639
1640
1641 Version 8.02 19-Mar-2010
1642 ------------------------
1643
1644 1. The Unicode data tables have been updated to Unicode 5.2.0.
1645
1646 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1647 configured.
1648
1649 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1650 original author of that file, following a query about its status.
1651
1652 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1653 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1654
1655 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1656 quantifier applied to a forward-referencing subroutine call, could compile
1657 incorrect code or give the error "internal error: previously-checked
1658 referenced subpattern not found".
1659
1660 6. Both MS Visual Studio and Symbian OS have problems with initializing
1661 variables to point to external functions. For these systems, therefore,
1662 pcre_malloc etc. are now initialized to local functions that call the
1663 relevant global functions.
1664
1665 7. There were two entries missing in the vectors called coptable and poptable
1666 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1667 I've fixed the data, and added a kludgy way of testing at compile time that
1668 the lengths are correct (equal to the number of opcodes).
1669
1670 8. Following on from 7, I added a similar kludge to check the length of the
1671 eint vector in pcreposix.c.
1672
1673 9. Error texts for pcre_compile() are held as one long string to avoid too
1674 much relocation at load time. To find a text, the string is searched,
1675 counting zeros. There was no check for running off the end of the string,
1676 which could happen if a new error number was added without updating the
1677 string.
1678
1679 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1680
1681 11. \K was not working if it appeared in an atomic group or in a group that
1682 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1683 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1684 the assertion is positive, but not if it is negative.
1685
1686 12. Change 11 fortuitously reduced the size of the stack frame used in the
1687 "match()" function of pcre_exec.c by one pointer. Forthcoming
1688 implementation of support for (*MARK) will need an extra pointer on the
1689 stack; I have reserved it now, so that the stack frame size does not
1690 decrease.
1691
1692 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1693 item in branch that calls a recursion is a subroutine call - as in the
1694 second branch in the above example - was incorrectly given the compile-
1695 time error "recursive call could loop indefinitely" because pcre_compile()
1696 was not correctly checking the subroutine for matching a non-empty string.
1697
1698 14. The checks for overrunning compiling workspace could trigger after an
1699 overrun had occurred. This is a "should never occur" error, but it can be
1700 triggered by pathological patterns such as hundreds of nested parentheses.
1701 The checks now trigger 100 bytes before the end of the workspace.
1702
1703 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1704
1705
1706 Version 8.01 19-Jan-2010
1707 ------------------------
1708
1709 1. If a pattern contained a conditional subpattern with only one branch (in
1710 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1711 computed the wrong minimum data length (which is of course zero for such
1712 subpatterns). This could cause incorrect "no match" results.
1713
1714 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1715 the pattern is reset in the first branch, pcre_compile() failed with
1716 "internal error: code overflow at offset...". This happened only when
1717 the reset was to the original external option setting. (An optimization
1718 abstracts leading options settings into an external setting, which was the
1719 cause of this.)
1720
1721 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1722 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1723 assertion pattern did not match (meaning that the assertion was true), it
1724 was incorrectly treated as false if the SKIP had been reached during the
1725 matching. This also applied to assertions used as conditions.
1726
1727 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1728 assertion subpattern, including such a pattern used as a condition,
1729 unpredictable results occurred, instead of the error return
1730 PCRE_ERROR_DFA_UITEM.
1731
1732 5. The C++ GlobalReplace function was not working like Perl for the special
1733 situation when an empty string is matched. It now does the fancy magic
1734 stuff that is necessary.
1735
1736 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1737 removed. (These were left over from very, very early versions of PCRE.)
1738
1739 7. Some cosmetic changes to the code to make life easier when compiling it
1740 as part of something else:
1741
1742 (a) Change DEBUG to PCRE_DEBUG.
1743
1744 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1745 called "current" as "current_branch", to prevent a collision with the
1746 Linux macro when compiled as a kernel module.
1747
1748 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1749 prevent a collision with the Linux macro when compiled as a kernel
1750 module.
1751
1752 8. In pcre_compile() there are some checks for integer overflows that used to
1753 cast potentially large values to (double). This has been changed to that
1754 when building, a check for int64_t is made, and if it is found, it is used
1755 instead, thus avoiding the use of floating point arithmetic. (There is no
1756 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1757 double.
1758
1759 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1760 2005 (difference between two addresses compared to an unsigned value).
1761
1762 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1763 custom one, because of the following reported problem in Windows:
1764
1765 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1766 under Win32.
1767 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1768 therefore missing the function definition.
1769 - The compiler thus generates a "C" signature for the test function.
1770 - The linker fails to find the "C" function.
1771 - PCRE fails to configure if asked to do so against libbz2.
1772
1773 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1774 messages were output:
1775
1776 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1777 rerunning libtoolize, to keep the correct libtool macros in-tree.
1778 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1779
1780 I have done both of these things.
1781
1782 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1783 most of the time, it *can* run out if it is given a pattern that contains a
1784 runaway infinite recursion. I updated the discussion in the pcrestack man
1785 page.
1786
1787 13. Now that we have gone to the x.xx style of version numbers, the minor
1788 version may start with zero. Using 08 or 09 is a bad idea because users
1789 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1790 interpreted as invalid octal numbers. I've updated the previous comment in
1791 configure.ac, and also added a check that gives an error if 08 or 09 are
1792 used.
1793
1794 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1795 causing partial matching to fail when the end of the subject matched \W
1796 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1797
1798 15. There were some discrepancies between the declarations in pcre_internal.h
1799 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1800 their definitions. The declarations used "const uschar *" and the
1801 definitions used USPTR. Even though USPTR is normally defined as "const
1802 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1803 reported that: "This difference in casting confuses some C++ compilers, for
1804 example, SunCC recognizes above declarations as different functions and
1805 generates broken code for hbpcre." I have changed the declarations to use
1806 USPTR.
1807
1808 16. GNU libtool is named differently on some systems. The autogen.sh script now
1809 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1810 (FreeBSD).
1811
1812 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1813 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1814 comment: "Figure out how to create a longlong from a string: strtoll and
1815 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1816 instance, but it only takes 2 args instead of 3!"
1817
1818 18. A subtle bug concerned with back references has been fixed by a change of
1819 specification, with a corresponding code fix. A pattern such as
1820 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1821 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1822 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1823 same bug. Such groups have to be quantified to be useful, or contained
1824 inside another quantified group. (If there's no repetition, the reference
1825 can never match.) The problem arises because, having left the group and
1826 moved on to the rest of the pattern, a later failure that backtracks into
1827 the group uses the captured value from the final iteration of the group
1828 rather than the correct earlier one. I have fixed this in PCRE by forcing
1829 any group that contains a reference to itself to be an atomic group; that
1830 is, there cannot be any backtracking into it once it has completed. This is
1831 similar to recursive and subroutine calls.
1832
1833
1834 Version 8.00 19-Oct-09
1835 ----------------------
1836
1837 1. The table for translating pcre_compile() error codes into POSIX error codes
1838 was out-of-date, and there was no check on the pcre_compile() error code
1839 being within the table. This could lead to an OK return being given in
1840 error.
1841
1842 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1843 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1844 in a Windows environment.
1845
1846 3. The pcregrep --count option prints the count for each file even when it is
1847 zero, as does GNU grep. However, pcregrep was also printing all files when
1848 --files-with-matches was added. Now, when both options are given, it prints
1849 counts only for those files that have at least one match. (GNU grep just
1850 prints the file name in this circumstance, but including the count seems
1851 more useful - otherwise, why use --count?) Also ensured that the
1852 combination -clh just lists non-zero counts, with no names.
1853
1854 4. The long form of the pcregrep -F option was incorrectly implemented as
1855 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1856 but it seems right to fix it, and I didn't think it was worth preserving
1857 the old behaviour.
1858
1859 5. The command line items --regex=pattern and --regexp=pattern were not
1860 recognized by pcregrep, which required --regex pattern or --regexp pattern
1861 (with a space rather than an '='). The man page documented the '=' forms,
1862 which are compatible with GNU grep; these now work.
1863
1864 6. No libpcreposix.pc file was created for pkg-config; there was just
1865 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1866
1867 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1868 when UCP support is not needed, by modifying the Python script that
1869 generates it from Unicode data files. This should not matter if the module
1870 is correctly used as a library, but I received one complaint about 50K of
1871 unwanted data. My guess is that the person linked everything into his
1872 program rather than using a library. Anyway, it does no harm.
1873
1874 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1875 was a minimum greater than 1 for a wide character in a possessive
1876 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1877 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1878 character. Chaos in the form of incorrect output or a compiling loop could
1879 result.
1880
1881 9. The restrictions on what a pattern can contain when partial matching is
1882 requested for pcre_exec() have been removed. All patterns can now be
1883 partially matched by this function. In addition, if there are at least two
1884 slots in the offset vector, the offset of the earliest inspected character
1885 for the match and the offset of the end of the subject are set in them when
1886 PCRE_ERROR_PARTIAL is returned.
1887
1888 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1889 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1890 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1891 and may be more useful for multi-segment matching.
1892
1893 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1894 used to be given if ever the end of the subject was reached; now it is
1895 given only if matching could not proceed because another character was
1896 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1897 string "Z", which now yields "no match" instead of "partial match". In the
1898 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1899 final character ended with (*FAIL).
1900
1901 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1902 if the pattern had a "must contain" character that was already found in the
1903 earlier partial match, unless partial matching was again requested. For
1904 example, with the pattern /dog.(body)?/, the "must contain" character is
1905 "g". If the first part-match was for the string "dog", restarting with
1906 "sbody" failed. This bug has been fixed.
1907
1908 13. The string returned by pcre_dfa_exec() after a partial match has been
1909 changed so that it starts at the first inspected character rather than the
1910 first character of the match. This makes a difference only if the pattern
1911 starts with a lookbehind assertion or \b or \B (\K is not supported by
1912 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1913 matching functions compatible, and I think it's the right thing to do.
1914
1915 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1916 so that the demonstration program is easily available in environments where
1917 PCRE has not been installed from source.
1918
1919 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1920 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1921 library.
1922
1923 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1924 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1925 is not the first non-POSIX option to be added. Clearly some people find
1926 these options useful.
1927
1928 17. If a caller to the POSIX matching function regexec() passes a non-zero
1929 value for nmatch with a NULL value for pmatch, the value of
1930 nmatch is forced to zero.
1931
1932 18. RunGrepTest did not have a test for the availability of the -u option of
1933 the diff command, as RunTest does. It now checks in the same way as
1934 RunTest, and also checks for the -b option.
1935
1936 19. If an odd number of negated classes containing just a single character
1937 interposed, within parentheses, between a forward reference to a named
1938 subpattern and the definition of the subpattern, compilation crashed with
1939 an internal error, complaining that it could not find the referenced
1940 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1941 [The bug was that it was starting one character too far in when skipping
1942 over the character class, thus treating the ] as data rather than
1943 terminating the class. This meant it could skip too much.]
1944
1945 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1946 /g option in pcretest when the pattern contains \K, which makes it possible
1947 to have an empty string match not at the start, even when the pattern is
1948 anchored. Updated pcretest and pcredemo to use this option.
1949
1950 21. If the maximum number of capturing subpatterns in a recursion was greater
1951 than the maximum at the outer level, the higher number was returned, but
1952 with unset values at the outer level. The correct (outer level) value is
1953 now given.
1954
1955 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1956 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1957 make it do so. The string so far is captured, making this feature
1958 compatible with Perl.
1959
1960 23. The tests have been re-organized, adding tests 11 and 12, to make it
1961 possible to check the Perl 5.10 features against Perl 5.10.
1962
1963 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1964 pattern matches a fixed length string. PCRE did not allow this; now it
1965 does. Neither allows recursion.
1966
1967 25. I finally figured out how to implement a request to provide the minimum
1968 length of subject string that was needed in order to match a given pattern.
1969 (It was back references and recursion that I had previously got hung up
1970 on.) This code has now been added to pcre_study(); it finds a lower bound
1971 to the length of subject needed. It is not necessarily the greatest lower
1972 bound, but using it to avoid searching strings that are too short does give
1973 some useful speed-ups. The value is available to calling programs via
1974 pcre_fullinfo().
1975
1976 26. While implementing 25, I discovered to my embarrassment that pcretest had
1977 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1978 study optimizations had never been tested with that matching function.
1979 Oops. What is worse, even when it was passed study data, there was a bug in
1980 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1981 were also very few tests of studied patterns with pcre_dfa_exec().
1982
1983 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1984 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1985 on the other side of the coin, they are no longer allowed to have different
1986 names, because these cannot be distinguished in PCRE, and this has caused
1987 confusion. (This is a difference from Perl.)
1988
1989 28. When duplicate subpattern names are present (necessarily with different
1990 numbers, as required by 27 above), and a test is made by name in a
1991 conditional pattern, either for a subpattern having been matched, or for
1992 recursion in such a pattern, all the associated numbered subpatterns are
1993 tested, and the overall condition is true if the condition is true for any
1994 one of them. This is the way Perl works, and is also more like the way
1995 testing by number works.
1996
1997
1998 Version 7.9 11-Apr-09
1999 ---------------------
2000
2001 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2002 (pcretest), all targets were linked against these libraries. This included
2003 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2004 libraries. This caused unwanted dependencies to be created. This problem
2005 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2006 pcretest is linked with readline.
2007
2008 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2009 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2010 moved outside it again, because FALSE and TRUE are already defined in AIX,
2011 but BOOL is not.
2012
2013 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2014 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2015
2016 4. The pcregrep documentation said spaces were inserted as well as colons (or
2017 hyphens) following file names and line numbers when outputting matching
2018 lines. This is not true; no spaces are inserted. I have also clarified the
2019 wording for the --colour (or --color) option.
2020
2021 5. In pcregrep, when --colour was used with -o, the list of matching strings
2022 was not coloured; this is different to GNU grep, so I have changed it to be
2023 the same.
2024
2025 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2026 each matching line was coloured. Now it goes on to look for further matches
2027 of any of the test patterns, which is the same behaviour as GNU grep.
2028
2029 7. A pattern that could match an empty string could cause pcregrep to loop; it
2030 doesn't make sense to accept an empty string match in pcregrep, so I have
2031 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2032 seems to be how GNU grep behaves. [But see later change 40 for release
2033 8.33.]
2034
2035 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2036 start or after a newline", because the conditional assertion was not being
2037 correctly handled. The rule now is that both the assertion and what follows
2038 in the first alternative must satisfy the test.
2039
2040 9. If auto-callout was enabled in a pattern with a conditional group whose
2041 condition was an assertion, PCRE could crash during matching, both with
2042 pcre_exec() and pcre_dfa_exec().
2043
2044 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2045 used for matching.
2046
2047 11. Unicode property support in character classes was not working for
2048 characters (bytes) greater than 127 when not in UTF-8 mode.
2049
2050 12. Added the -M command line option to pcretest.
2051
2052 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2053
2054 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2055
2056 16. Added comments and documentation about mis-use of no_arg in the C++
2057 wrapper.
2058
2059 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2060 from Martin Jerabek that uses macro names for all relevant character and
2061 string constants.
2062
2063 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2064 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2065 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2066 these, but not everybody uses configure.
2067
2068 19. A conditional group that had only one branch was not being correctly
2069 recognized as an item that could match an empty string. This meant that an
2070 enclosing group might also not be so recognized, causing infinite looping
2071 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2072 with the subject "ab", where knowledge that the repeated group can match
2073 nothing is needed in order to break the loop.
2074
2075 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2076 exec(), but without supplying a callout function, matching went wrong.
2077
2078 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2079 leak if the size of the offset vector was greater than 30. When the vector
2080 is smaller, the saved offsets during recursion go onto a local stack
2081 vector, but for larger vectors malloc() is used. It was failing to free
2082 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2083 error, in fact).
2084
2085 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2086 heapframe that is used only when UTF-8 support is enabled. This caused no
2087 problem, but was untidy.
2088
2089 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2090 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2091 included within another project.
2092
2093 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2094 slightly modified by me:
2095
2096 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2097 not building pcregrep.
2098
2099 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2100 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2101
2102 25. Forward references, both numeric and by name, in patterns that made use of
2103 duplicate group numbers, could behave incorrectly or give incorrect errors,
2104 because when scanning forward to find the reference group, PCRE was not
2105 taking into account the duplicate group numbers. A pattern such as
2106 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2107
2108 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2109 the feature of a custom pointer more persuasive (as requested by a user).
2110
2111 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2112 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2113 pre-defined.
2114
2115 28. Added support for (*UTF8) at the start of a pattern.
2116
2117 29. Arrange for flags added by the "release type" setting in CMake to be shown
2118 in the configuration summary.
2119
2120
2121 Version 7.8 05-Sep-08
2122 ---------------------
2123
2124 1. Replaced UCP searching code with optimized version as implemented for Ad
2125 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2126 stage table and inline lookup instead of a function, giving speed ups of 2
2127 to 5 times on some simple patterns that I tested. Permission was given to
2128 distribute the MultiStage2.py script that generates the tables (it's not in
2129 the tarball, but is in the Subversion repository).
2130
2131 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2132 scripts.
2133
2134 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2135 a group with a zero qualifier. The result of the study could be incorrect,
2136 or the function might crash, depending on the pattern.
2137
2138 4. Caseless matching was not working for non-ASCII characters in back
2139 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2140 It now works when Unicode Property Support is available.
2141
2142 5. In pcretest, an escape such as \x{de} in the data was always generating
2143 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2144 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2145 truncation.
2146
2147 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2148
2149 7. Added two (int) casts to pcregrep when printing the difference of two
2150 pointers, in case they are 64-bit values.
2151
2152 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2153 test 2 if it fails.
2154
2155 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2156 and a #define of that name to empty if it is not externally set. This is to
2157 allow users of MSVC to set it if necessary.
2158
2159 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2160 the convenience functions in the pcre_get.c source file.
2161
2162 11. An option change at the start of a pattern that had top-level alternatives
2163 could cause overwriting and/or a crash. This command provoked a crash in
2164 some environments:
2165
2166 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2167
2168 This potential security problem was recorded as CVE-2008-2371.
2169
2170 12. For a pattern where the match had to start at the beginning or immediately
2171 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2172 pcre_dfa_exec() could read past the end of the passed subject if there was
2173 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2174 pcretest so that it places the subject at the end of its malloc-ed buffer.
2175
2176 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2177 exec() might read past the end of the data buffer in UTF-8 mode.
2178
2179 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2180 the data contained the byte 0x85 as part of a UTF-8 character within its
2181 first line. This applied both to normal and DFA matching.
2182
2183 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2184 /^[^d]*?$/8 failed to match "abc".
2185
2186 16. Added a missing copyright notice to pcrecpp_internal.h.
2187
2188 17. Make it more clear in the documentation that values returned from
2189 pcre_exec() in ovector are byte offsets, not character counts.
2190
2191 18. Tidied a few places to stop certain compilers from issuing warnings.
2192
2193 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2194 supplied by Stefan Weber. I made a further small update for 7.8 because
2195 there is a change of source arrangements: the pcre_searchfuncs.c module is
2196 replaced by pcre_ucd.c.
2197
2198
2199 Version 7.7 07-May-08
2200 ---------------------
2201
2202 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2203 a string to a long long, pretend we don't even have a long long." This is
2204 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2205
2206 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2207 pre-7.6 versions, which defined a global no_arg variable instead of putting
2208 it in the RE class. (See also #8 below.)
2209
2210 3. Remove a line of dead code, identified by coverity and reported by Nuno
2211 Lopes.
2212
2213 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2214
2215 (1) The include/exclude patterns were being applied to the whole pathnames
2216 of files, instead of just to the final components.
2217
2218 (2) If there was more than one level of directory, the subdirectories were
2219 skipped unless they satisfied the include/exclude conditions. This is
2220 inconsistent with GNU grep (and could even be seen as contrary to the
2221 pcregrep specification - which I improved to make it absolutely clear).
2222 The action now is always to scan all levels of directory, and just
2223 apply the include/exclude patterns to regular files.
2224
2225 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2226 --exclude_dir in the tests to avoid scanning .svn directories.
2227
2228 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2229 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2230 doesn't support NULs in patterns.
2231
2232 7. Added some missing "const"s to declarations of static tables in
2233 pcre_compile.c and pcre_dfa_exec.c.
2234
2235 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2236 caused by fix #2 above. (Subsequently also a second patch to fix the
2237 first patch. And a third patch - this was a messy problem.)
2238
2239 9. Applied Craig's patch to remove the use of push_back().
2240
2241 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2242 matching function regexec().
2243
2244 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2245 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2246 references. PCRE supports relative numbers with this syntax (I don't think
2247 Oniguruma does).
2248
2249 12. Previously, a group with a zero repeat such as (...){0} was completely
2250 omitted from the compiled regex. However, this means that if the group
2251 was called as a subroutine from elsewhere in the pattern, things went wrong
2252 (an internal error was given). Such groups are now left in the compiled
2253 pattern, with a new opcode that causes them to be skipped at execution
2254 time.
2255
2256 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2257 to the way PCRE behaves:
2258
2259 (a) A lone ] character is dis-allowed (Perl treats it as data).
2260
2261 (b) A back reference to an unmatched subpattern matches an empty string
2262 (Perl fails the current match path).
2263
2264 (c) A data ] in a character class must be notated as \] because if the
2265 first data character in a class is ], it defines an empty class. (In
2266 Perl it is not possible to have an empty class.) The empty class []
2267 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2268 The negative empty class [^] matches any one character, independently
2269 of the DOTALL setting.
2270
2271 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2272 non-existent subpattern following a character class starting with ']' and
2273 containing () gave an internal compiling error instead of "reference to
2274 non-existent subpattern". Fortunately, when the pattern did exist, the
2275 compiled code was correct. (When scanning forwards to check for the
2276 existence of the subpattern, it was treating the data ']' as terminating
2277 the class, so got the count wrong. When actually compiling, the reference
2278 was subsequently set up correctly.)
2279
2280 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2281 it was being rejected as not supported by pcre_dfa_exec(), even though
2282 other assertions are supported. I have made pcre_dfa_exec() support
2283 (*FAIL).
2284
2285 16. The implementation of 13c above involved the invention of a new opcode,
2286 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2287 cannot be changed at match time, I realized I could make a small
2288 improvement to matching performance by compiling OP_ALLANY instead of
2289 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2290 on the OP_ANY path.
2291
2292 17. Compiling pcretest on Windows with readline support failed without the
2293 following two fixes: (1) Make the unistd.h include conditional on
2294 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2295
2296 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2297 ncurses library to be included for pcretest when ReadLine support is
2298 requested, but also to allow for it to be overridden. This patch came from
2299 Daniel Bergström.
2300
2301 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2302 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2303 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2304 spotting this.
2305
2306
2307 Version 7.6 28-Jan-08
2308 ---------------------
2309
2310 1. A character class containing a very large number of characters with
2311 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2312 overflow.
2313
2314 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2315 HAVE_LONG_LONG is not defined.
2316
2317 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2318 bring them up to date and include new features. This patch includes:
2319
2320 - Fixed PH's badly added libz and libbz2 support.
2321 - Fixed a problem with static linking.
2322 - Added pcredemo. [But later removed - see 7 below.]
2323 - Fixed dftables problem and added an option.
2324 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2325 HAVE_LONG_LONG.
2326 - Added readline support for pcretest.
2327 - Added an listing of the option settings after cmake has run.
2328
2329 4. A user submitted a patch to Makefile that makes it easy to create
2330 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2331 Makefile.am that cause it to include this special target, without
2332 affecting anything else. Note that the same mingw target plus all
2333 the other distribution libraries and programs are now supported
2334 when configuring with CMake (see 6 below) instead of with
2335 Configure/Make.
2336
2337 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2338 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2339 exported in the Windows port". It has not yet been confirmed that the patch
2340 solves the problem, but it does no harm.
2341
2342 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2343 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2344 with CMake, and also correct the comment about stack recursion.
2345
2346 7. Remove the automatic building of pcredemo from the ./configure system and
2347 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2348 of a program that users should build themselves after PCRE is installed, so
2349 building it automatically is not really right. What is more, it gave
2350 trouble in some build environments.
2351
2352 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2353
2354
2355 Version 7.5 10-Jan-08
2356 ---------------------
2357
2358 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2359 values in parens when parsing an RE using the C++ wrapper."
2360
2361 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2362 Characters greater than 255 were excluded from the class instead of being
2363 included.
2364
2365 3. The same bug as (2) above applied to negated POSIX classes such as
2366 [:^space:].
2367
2368 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2369 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2370 I have changed it.
2371
2372 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2373 first named subpattern) and a construct such as (?&a) would reference the
2374 first named subpattern whose name started with "a" (in other words, the
2375 length check was missing). Both these problems are fixed. "Subpattern name
2376 expected" is now given for (?&) (a zero-length name), and this patch also
2377 makes it give the same error for \k'' (previously it complained that that
2378 was a reference to a non-existent subpattern).
2379
2380 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2381 this is right because (?- can be followed by option settings as well as by
2382 digits. I have, however, made the messages clearer.
2383
2384 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2385 than the number used in the conditional) now cause a compile-time error.
2386 This is actually not compatible with Perl, which accepts such patterns, but
2387 treats the conditional as always being FALSE (as PCRE used to), but it
2388 seems to me that giving a diagnostic is better.
2389
2390 8. Change "alphameric" to the more common word "alphanumeric" in comments
2391 and messages.
2392
2393 9. Fix two occurrences of "backslash" in comments that should have been
2394 "backspace".
2395
2396 10. Remove two redundant lines of code that can never be obeyed (their function
2397 was moved elsewhere).
2398
2399 11. The program that makes PCRE's Unicode character property table had a bug
2400 which caused it to generate incorrect table entries for sequences of
2401 characters that have the same character type, but are in different scripts.
2402 It amalgamated them into a single range, with the script of the first of
2403 them. In other words, some characters were in the wrong script. There were
2404 thirteen such cases, affecting characters in the following ranges:
2405
2406 U+002b0 - U+002c1
2407 U+0060c - U+0060d
2408 U+0061e - U+00612
2409 U+0064b - U+0065e
2410 U+0074d - U+0076d
2411 U+01800 - U+01805
2412 U+01d00 - U+01d77
2413 U+01d9b - U+01dbf
2414 U+0200b - U+0200f
2415 U+030fc - U+030fe
2416 U+03260 - U+0327f
2417 U+0fb46 - U+0fbb1
2418 U+10450 - U+1049d
2419
2420 12. The -o option (show only the matching part of a line) for pcregrep was not
2421 compatible with GNU grep in that, if there was more than one match in a
2422 line, it showed only the first of them. It now behaves in the same way as
2423 GNU grep.
2424
2425 13. If the -o and -v options were combined for pcregrep, it printed a blank
2426 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2427 does the same. The return code can be used to tell if there were any
2428 non-matching lines.
2429
2430 14. Added --file-offsets and --line-offsets to pcregrep.
2431
2432 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2433 infinitely looping recursion. The bug was that positive lookaheads were not
2434 being skipped when checking for a possible empty match (negative lookaheads
2435 and both kinds of lookbehind were skipped).
2436
2437 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2438 inclusion of <windows.h> to before rather than after the definition of
2439 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2440
2441 17. Specifying a possessive quantifier with a specific limit for a Unicode
2442 character property caused pcre_compile() to compile bad code, which led at
2443 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2444 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2445 caused the error; without that there was no problem.
2446
2447 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2448
2449 19. Added --enable-pcretest-libreadline.
2450
2451 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2452 RE::GlobalReplace(). As a result, the number of replacements returned was
2453 double what it should be. I removed one of the increments, but Craig sent a
2454 later patch that removed the other one (the right fix) and added unit tests
2455 that check the return values (which was not done before).
2456
2457 21. Several CMake things:
2458
2459 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2460 the names libpcre and libpcreposix, not just pcre and pcreposix.
2461
2462 (2) The above change means that pcretest and pcregrep are now correctly
2463 linked with the newly-built libraries, not previously installed ones.
2464
2465 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2466
2467 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2468 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2469 UTF-8 newline character). The key issue is that the pattern starts .*;
2470 this means that the match must be either at the beginning, or after a
2471 newline. The bug was in the code for advancing after a failed match and
2472 checking that the new position followed a newline. It was not taking
2473 account of UTF-8 characters correctly.
2474
2475 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2476 character classes. PCRE was not treating the sequence [:...:] as a
2477 character class unless the ... were all letters. Perl, however, seems to
2478 allow any characters between [: and :], though of course it rejects as
2479 unknown any "names" that contain non-letters, because all the known class
2480 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2481 for example, whereas PCRE did not - it did not recognize a POSIX character
2482 class. This seemed a bit dangerous, so the code has been changed to be
2483 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2484 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2485 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2486 Perl does, and where it didn't before.
2487
2488 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2489 Windows environments %n is disabled by default.
2490
2491
2492 Version 7.4 21-Sep-07
2493 ---------------------
2494
2495 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2496 means that a class such as [\s] counted as "explicit reference to CR or
2497 LF". That isn't really right - the whole point of the change was to try to
2498 help when there was an actual mention of one of the two characters. So now
2499 the change happens only if \r or \n (or a literal CR or LF) character is
2500 encountered.
2501
2502 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2503 of both had grown to the point where there were only 3 bits left.
2504 Fortunately, there was spare space in the data structure, and so I have
2505 moved the internal flags into a new 16-bit field to free up more option
2506 bits.
2507
2508 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2509 but did not set the internal JCHANGED flag - either of these is enough to
2510 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2511 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2512 start sets both bits.
2513
2514 4. Added options (at build time, compile time, exec time) to change \R from
2515 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2516
2517 5. doc/pcresyntax.html was missing from the distribution.
2518
2519 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2520 compatibility, even though it is no longer used.
2521
2522 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2523 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2524 windows.h file is present (where different names are used). [This was
2525 reversed later after testing - see 16 below.]
2526
2527 8. Changed all #include <config.h> to #include "config.h". There were also
2528 some further <pcre.h> cases that I changed to "pcre.h".
2529
2530 9. When pcregrep was used with the --colour option, it missed the line ending
2531 sequence off the lines that it output.
2532
2533 10. It was pointed out to me that arrays of string pointers cause lots of
2534 relocations when a shared library is dynamically loaded. A technique of
2535 using a single long string with a table of offsets can drastically reduce
2536 these. I have refactored PCRE in four places to do this. The result is
2537 dramatic:
2538
2539 Originally: 290
2540 After changing UCP table: 187
2541 After changing error message table: 43
2542 After changing table of "verbs" 36
2543 After changing table of Posix names 22
2544
2545 Thanks to the folks working on Gregex for glib for this insight.
2546
2547 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2548 unicode-properties was also set.
2549
2550 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2551
2552 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2553 checked only for CRLF.
2554
2555 14. Added casts to pcretest.c to avoid compiler warnings.
2556
2557 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2558
2559 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2560 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2561 entirely. This removes changes made in 7 above.
2562
2563 17. The CMake files have been updated, and there is now more information about
2564 building with CMake in the NON-UNIX-USE document.
2565
2566
2567 Version 7.3 28-Aug-07
2568 ---------------------
2569
2570 1. In the rejigging of the build system that eventually resulted in 7.1, the
2571 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2572 brackets there is not right, since it causes compilers to look for an
2573 installed pcre.h, not the version that is in the source that is being
2574 compiled (which of course may be different). I have changed it back to:
2575
2576 #include "pcre.h"
2577
2578 I have a vague recollection that the change was concerned with compiling in
2579 different directories, but in the new build system, that is taken care of
2580 by the VPATH setting the Makefile.
2581
2582 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2583 when the subject happened to end in the byte 0x85 (e.g. if the last
2584 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2585 characters but of course it shouldn't be taken as a newline when it is part
2586 of another character. The bug was that, for an unlimited repeat of . in
2587 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2588 characters when looking for a newline.
2589
2590 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2591
2592 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2593 in debug output.
2594
2595 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2596 long printing in the pcrecpp unittest when running under MinGW.
2597
2598 6. ESC_K was left out of the EBCDIC table.
2599
2600 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2601 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2602 limit also applies to "virtual nesting" when a pattern is recursive, and in
2603 this case 1000 isn't so big. I have been able to remove this limit at the
2604 expense of backing off one optimization in certain circumstances. Normally,
2605 when pcre_exec() would call its internal match() function recursively and
2606 immediately return the result unconditionally, it uses a "tail recursion"
2607 feature to save stack. However, when a subpattern that can match an empty
2608 string has an unlimited repetition quantifier, it no longer makes this
2609 optimization. That gives it a stack frame in which to save the data for
2610 checking that an empty string has been matched. Previously this was taken
2611 from the 1000-entry workspace that had been reserved. So now there is no
2612 explicit limit, but more stack is used.
2613
2614 8. Applied Daniel's patches to solve problems with the import/export magic
2615 syntax that is required for Windows, and which was going wrong for the
2616 pcreposix and pcrecpp parts of the library. These were overlooked when this
2617 problem was solved for the main library.
2618
2619 9. There were some crude static tests to avoid integer overflow when computing
2620 the size of patterns that contain repeated groups with explicit upper
2621 limits. As the maximum quantifier is 65535, the maximum group length was
2622 set at 30,000 so that the product of these two numbers did not overflow a
2623 32-bit integer. However, it turns out that people want to use groups that
2624 are longer than 30,000 bytes (though not repeat them that many times).
2625 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2626 made it possible to implement the integer overflow checks in a much more
2627 dynamic way, which I have now done. The artificial limitation on group
2628 length has been removed - we now have only the limit on the total length of
2629 the compiled pattern, which depends on the LINK_SIZE setting.
2630
2631 10. Fixed a bug in the documentation for get/copy named substring when
2632 duplicate names are permitted. If none of the named substrings are set, the
2633 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2634 empty string.
2635
2636 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2637 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2638 because the ] is interpreted as the first data character and the
2639 terminating ] is not found. PCRE has been made compatible with Perl in this
2640 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2641 cause memory overwriting.
2642
2643 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2644 string has been matched (to stop an infinite loop). It was not recognizing
2645 a conditional subpattern that could match an empty string if that
2646 subpattern was within another subpattern. For example, it looped when
2647 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2648 condition was not nested. This bug has been fixed.
2649
2650 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2651 past the start of the subject in the presence of bytes with the top bit
2652 set, for example "\x8aBCD".
2653
2654 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2655 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2656
2657 14. Optimized (?!) to (*FAIL).
2658
2659 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2660 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2661 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2662 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2663 does: it's just the validity check that is more restrictive.
2664
2665 16. Inserted checks for integer overflows during escape sequence (backslash)
2666 processing, and also fixed erroneous offset values for syntax errors during
2667 backslash processing.
2668
2669 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2670 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2671
2672 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2673 caused an overrun.
2674
2675 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2676 something other than just ASCII characters) inside a group that had an
2677 unlimited repeat caused a loop at compile time (while checking to see
2678 whether the group could match an empty string).
2679
2680 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2681 [\P{Any}] did so. (Error in the code for printing property names.)
2682
2683 21. An orphan \E inside a character class could cause a crash.
2684
2685 22. A repeated capturing bracket such as (A)? could cause a wild memory
2686 reference during compilation.
2687
2688 23. There are several functions in pcre_compile() that scan along a compiled
2689 expression for various reasons (e.g. to see if it's fixed length for look
2690 behind). There were bugs in these functions when a repeated \p or \P was
2691 present in the pattern. These operators have additional parameters compared
2692 with \d, etc, and these were not being taken into account when moving along
2693 the compiled data. Specifically:
2694
2695 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2696 length.
2697
2698 (b) An item such as \pL+ within a repeated group could cause crashes or
2699 loops.
2700
2701 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2702 "reference to non-existent subpattern" error.
2703
2704 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2705
2706 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2707 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2708
2709 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2710
2711 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2712 character were causing crashes (broken optimization).
2713
2714 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2715 \p or \P) caused a compile-time loop.
2716
2717 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2718 break. For example, the unstudied pattern [\r\n]A does not match the string
2719 "\r\nA" because change 7.0/46 below moves the current point on by two
2720 characters after failing to match at the start. However, the pattern \nA
2721 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2722 the same is true. There doesn't seem any very clean way out of this, but
2723 what I have chosen to do makes the common cases work: PCRE now takes note
2724 of whether there can be an explicit match for \r or \n anywhere in the
2725 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2726 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2727 pattern has explicit CR or LF references.
2728
2729 29. Added (*CR) etc for changing newline setting at start of pattern.
2730
2731
2732 Version 7.2 19-Jun-07
2733 ---------------------
2734
2735 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2736 which is apparently normally available under Windows.
2737
2738 2. Re-jig the pcregrep tests with different newline settings in an attempt
2739 to make them independent of the local environment's newline setting.
2740
2741 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2742
2743 4. Some of the "internals" tests were previously cut out when the link size
2744 was not 2, because the output contained actual offsets. The recent new
2745 "Z" feature of pcretest means that these can be cut out, making the tests
2746 usable with all link sizes.
2747
2748 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2749 stack recursion. This gives a massive performance boost under BSD, but just
2750 a small improvement under Linux. However, it saves one field in the frame
2751 in all cases.
2752
2753 6. Added more features from the forthcoming Perl 5.10:
2754
2755 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2756 recursion call. It refers to the nth most recently opened parentheses.
2757
2758 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2759 to be opened parentheses.
2760
2761 (c) Conditions that refer to capturing parentheses can be specified
2762 relatively, for example, (?(-2)... or (?(+3)...
2763
2764 (d) \K resets the start of the current match so that everything before
2765 is not part of it.
2766
2767 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2768
2769 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2770 reference syntax.
2771
2772 (g) (?| introduces a group in which the numbering of parentheses in each
2773 alternative starts with the same number.
2774
2775 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2776
2777 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2778 PCRE_INFO_JCHANGED.
2779
2780 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2781 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2782 for detecting groups that can match an empty string.
2783
2784 9. A pattern with a very large number of alternatives (more than several
2785 hundred) was running out of internal workspace during the pre-compile
2786 phase, where pcre_compile() figures out how much memory will be needed. A
2787 bit of new cunning has reduced the workspace needed for groups with
2788 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2789 workspace instead of running out of the 4096 that are available.
2790
2791 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2792
2793 11. Applied patch from Google to remove an optimization that didn't quite work.
2794 The report of the bug said:
2795
2796 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2797 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2798 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2799
2800 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2801 it matched the wrong number of bytes.
2802
2803
2804 Version 7.1 24-Apr-07
2805 ---------------------
2806
2807 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2808 that is more "standard", making use of automake and other Autotools. There
2809 is some re-arrangement of the files and adjustment of comments consequent
2810 on this.
2811
2812 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2813 for recursive directory scanning broke on some systems because the files
2814 are not scanned in any specific order and on different systems the order
2815 was different. A call to "sort" has been inserted into RunGrepTest for the
2816 approprate test as a short-term fix. In the longer term there may be an
2817 alternative.
2818
2819 3. I had an email from Eric Raymond about problems translating some of PCRE's
2820 man pages to HTML (despite the fact that I distribute HTML pages, some
2821 people do their own conversions for various reasons). The problems
2822 concerned the use of low-level troff macros .br and .in. I have therefore
2823 removed all such uses from the man pages (some were redundant, some could
2824 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2825 HTML has been updated to handle .nf/.fi and to complain if it encounters
2826 .br or .in.
2827
2828 4. Updated comments in configure.ac that get placed in config.h.in and also
2829 arranged for config.h to be included in the distribution, with the name
2830 config.h.generic, for the benefit of those who have to compile without
2831 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2832
2833 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2834 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2835 makevp.bat for the current PCRE, using the additional files
2836 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2837
2838 6. A Windows user reported a minor discrepancy with test 2, which turned out
2839 to be caused by a trailing space on an input line that had got lost in his
2840 copy. The trailing space was an accident, so I've just removed it.
2841
2842 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2843 that is needed.
2844
2845 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2846 as "const" (a) because they are and (b) because it helps the PHP
2847 maintainers who have recently made a script to detect big data structures
2848 in the php code that should be moved to the .rodata section. I remembered
2849 to update Builducptable as well, so it won't revert if ucptable.h is ever
2850 re-created.
2851
2852 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2853 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2854 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2855 support is not required. This saves 1.5-2K of code, which is important in
2856 some applications.
2857
2858 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2859 so as not to refer to the tables, even though these functions will never be
2860 called when UTF-8 support is disabled. Otherwise there are problems with a
2861 shared library.
2862
2863 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2864
2865 (a) It was defining its arguments as char * instead of void *.
2866
2867 (b) It was assuming that all moves were upwards in memory; this was true
2868 a long time ago when I wrote it, but is no longer the case.
2869
2870 The emulated memove() is provided for those environments that have neither
2871 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2872 is clearly not the case, as these two bugs were recently reported.
2873
2874 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2875 and Detrail to create the HTML documentation, the .txt form of the man
2876 pages, and it removes trailing spaces from listed files. It also creates
2877 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2878 case, it wraps all the #defines with #ifndefs. This script should be run
2879 before "make dist".
2880
2881 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2882 with Unicode property support.
2883
2884 (a) For a maximizing quantifier, if the two different cases of the
2885 character were of different lengths in their UTF-8 codings (there are
2886 some cases like this - I found 11), and the matching function had to
2887 back up over a mixture of the two cases, it incorrectly assumed they
2888 were both the same length.
2889
2890 (b) When PCRE was configured to use the heap rather than the stack for
2891 recursion during matching, it was not correctly preserving the data for
2892 the other case of a UTF-8 character when checking ahead for a match
2893 while processing a minimizing repeat. If the check also involved
2894 matching a wide character, but failed, corruption could cause an
2895 erroneous result when trying to check for a repeat of the original
2896 character.
2897
2898 13. Some tidying changes to the testing mechanism:
2899
2900 (a) The RunTest script now detects the internal link size and whether there
2901 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2902 values substituted by "configure". (The RunGrepTest script already did
2903 this for UTF-8.) The configure.ac script no longer substitutes the
2904 relevant variables.
2905
2906 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2907 with length and offset values. This means that the output is different
2908 for different internal link sizes. Test 2 is skipped for link sizes
2909 other than 2 because of this, bypassing the problem. Unfortunately,
2910 there was also a test in test 3 (the locale tests) that used /B and
2911 failed for link sizes other than 2. Rather than cut the whole test out,
2912 I have added a new /Z option to pcretest that replaces the length and
2913 offset values with spaces. This is now used to make test 3 independent
2914 of link size. (Test 2 will be tidied up later.)
2915
2916 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2917 segmentation fault instead of returning the appropriate error message.
2918
2919 15. In multiline mode when the newline sequence was set to "any", the pattern
2920 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2921 This doesn't seem right; it now treats the CRLF combination as the line
2922 ending, and so does not match in that case. It's only a pattern such as ^$
2923 that would hit this one: something like ^ABC$ would have failed after \r
2924 and then tried again after \r\n.
2925
2926 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2927 in an attempt to make files that differ only in their line terminators
2928 compare equal. This works on Linux.
2929
2930 17. Under certain error circumstances pcregrep might try to free random memory
2931 as it exited. This is now fixed, thanks to valgrind.
2932
2933 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2934 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2935 was because its rules for how to advance for /g after matching an empty
2936 string at the end of a line did not allow for this case. They now check for
2937 it specially.
2938
2939 20. pcretest is supposed to handle patterns and data of any length, by
2940 extending its buffers when necessary. It was getting this wrong when the
2941 buffer for a data line had to be extended.
2942
2943 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2944 CRLF as a newline sequence.
2945
2946 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2947 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2948 I have nevertheless tidied it up.
2949
2950 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2951
2952 24. Added a man page for pcre-config.
2953
2954
2955 Version 7.0 19-Dec-06
2956 ---------------------
2957
2958 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2959 moving to gcc 4.1.1.
2960
2961 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2962 sys/time.h, which is documented as needed for this function. It doesn't
2963 seem to matter on Linux, but it showed up on some releases of OS X.
2964
2965 3. It seems that there are systems where bytes whose values are greater than
2966 127 match isprint() in the "C" locale. The "C" locale should be the
2967 default when a C program starts up. In most systems, only ASCII printing
2968 characters match isprint(). This difference caused the output from pcretest
2969 to vary, making some of the tests fail. I have changed pcretest so that:
2970
2971 (a) When it is outputting text in the compiled version of a pattern, bytes
2972 other than 32-126 are always shown as hex escapes.
2973
2974 (b) When it is outputting text that is a matched part of a subject string,
2975 it does the same, unless a different locale has been set for the match
2976 (using the /L modifier). In this case, it uses isprint() to decide.
2977
2978 4. Fixed a major bug that caused incorrect computation of the amount of memory
2979 required for a compiled pattern when options that changed within the
2980 pattern affected the logic of the preliminary scan that determines the
2981 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2982 that the computed length was too small. The symptoms of this bug were
2983 either the PCRE error "internal error: code overflow" from pcre_compile(),
2984 or a glibc crash with a message such as "pcretest: free(): invalid next
2985 size (fast)". Examples of patterns that provoked this bug (shown in
2986 pcretest format) are:
2987
2988 /(?-x: )/x
2989 /(?x)(?-x: \s*#\s*)/
2990 /((?i)[\x{c0}])/8
2991 /(?i:[\x{c0}])/8
2992
2993 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2994 is now done differently.
2995
2996 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2997 wrapper classes; (b) implement a new function in the C++ scanner that is
2998 more efficient than the old way of doing things because it avoids levels of
2999 recursion in the regex matching; (c) add a paragraph to the documentation
3000 for the FullMatch() function.
3001
3002 6. The escape sequence \n was being treated as whatever was defined as
3003 "newline". Not only was this contrary to the documentation, which states
3004 that \n is character 10 (hex 0A), but it also went horribly wrong when
3005 "newline" was defined as CRLF. This has been fixed.
3006
3007 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3008 was being set to -1 for the "end of line" case (supposedly a value that no
3009 character can have). Though this value is never used (the check for end of
3010 line is "zero bytes in current character"), it caused compiler complaints.
3011 I've changed it to 0xffffffff.
3012
3013 8. In pcre_version.c, the version string was being built by a sequence of
3014 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3015 string (as it is for production releases) called a macro with an empty
3016 argument. The C standard says the result of this is undefined. The gcc
3017 compiler treats it as an empty string (which was what was wanted) but it is
3018 reported that Visual C gives an error. The source has been hacked around to
3019 avoid this problem.
3020
3021 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3022 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3023 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3024 of them did).
3025
3026 10. Originally, pcretest opened its input and output without "b"; then I was
3027 told that "b" was needed in some environments, so it was added for release
3028 5.0 to both the input and output. (It makes no difference on Unix-like
3029 systems.) Later I was told that it is wrong for the input on Windows. I've
3030 now abstracted the modes into two macros, to make it easier to fiddle with
3031 them, and removed "b" from the input mode under Windows.
3032
3033 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3034
3035 12. Added -help and --help to pcretest as an official way of being reminded
3036 of the options.
3037
3038 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3039 and pcrecpp.cc because they annoy compilers at high warning levels.
3040
3041 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3042
3043 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3044 scripts are not C programs :-) and which was not noticed because it works
3045 on Linux.
3046
3047 16. pcretest is supposed to handle any length of pattern and data line (as one
3048 line or as a continued sequence of lines) by extending its input buffer if
3049 necessary. This feature was broken for very long pattern lines, leading to
3050 a string of junk being passed to pcre_compile() if the pattern was longer
3051 than about 50K.
3052
3053 17. I have done a major re-factoring of the way pcre_compile() computes the
3054 amount of memory needed for a compiled pattern. Previously, there was code
3055 that made a preliminary scan of the pattern in order to do this. That was
3056 OK when PCRE was new, but as the facilities have expanded, it has become
3057 harder and harder to keep it in step with the real compile phase, and there
3058 have been a number of bugs (see for example, 4 above). I have now found a
3059 cunning way of running the real compile function in a "fake" mode that
3060 enables it to compute how much memory it would need, while actually only
3061 ever using a few hundred bytes of working memory and without too many
3062 tests of the mode. This should make future maintenance and development
3063 easier. A side effect of this work is that the limit of 200 on the nesting
3064 depth of parentheses has been removed (though this was never a serious
3065 limitation, I suspect). However, there is a downside: pcre_compile() now
3066 runs more slowly than before (30% or more, depending on the pattern). I
3067 hope this isn't a big issue. There is no effect on runtime performance.
3068
3069 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3070 newline (only possible for the last line of a file) and it was a
3071 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3072
3073 19. Added additional timing features to pcretest. (1) The -tm option now times
3074 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3075 separate command line item, by a number that specifies the number of
3076 repeats to use when timing. The default is 50000; this gives better
3077 precision, but takes uncomfortably long for very large patterns.
3078
3079 20. Extended pcre_study() to be more clever in cases where a branch of a
3080 subpattern has no definite first character. For example, (a*|b*)[cd] would
3081 previously give no result from pcre_study(). Now it recognizes that the
3082 first character must be a, b, c, or d.
3083
3084 21. There was an incorrect error "recursive call could loop indefinitely" if
3085 a subpattern (or the entire pattern) that was being tested for matching an
3086 empty string contained only one non-empty item after a nested subpattern.
3087 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3088 incorrectly, because the \d was being skipped in the check.
3089
3090 22. The pcretest program now has a new pattern option /B and a command line
3091 option -b, which is equivalent to adding /B to every pattern. This causes
3092 it to show the compiled bytecode, without the additional information that
3093 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3094 is the same as /B/I).
3095
3096 23. A new optimization is now able automatically to treat some sequences such
3097 as a*b as a*+b. More specifically, if something simple (such as a character
3098 or a simple class like \d) has an unlimited quantifier, and is followed by
3099 something that cannot possibly match the quantified thing, the quantifier
3100 is automatically "possessified".
3101
3102 24. A recursive reference to a subpattern whose number was greater than 39
3103 went wrong under certain circumstances in UTF-8 mode. This bug could also
3104 have affected the operation of pcre_study().
3105
3106 25. Realized that a little bit of performance could be had by replacing
3107 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3108
3109 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3110
3111 27. Possessive quantifiers such as a++ were previously implemented by turning
3112 them into atomic groups such as ($>a+). Now they have their own opcodes,
3113 which improves performance. This includes the automatically created ones
3114 from 23 above.
3115
3116 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3117 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3118 the first matched character to be a colon. This applied both to named and
3119 numbered groups.
3120
3121 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3122
3123 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3124 building PCRE on Windows easier, so I have included it in the distribution.
3125
3126 31. There is now a check in pcretest against a ridiculously large number being
3127 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3128 loop, the loop is abandoned.
3129
3130 32. Forward references to subpatterns in conditions such as (?(2)...) where
3131 subpattern 2 is defined later cause pcre_compile() to search forwards in
3132 the pattern for the relevant set of parentheses. This search went wrong
3133 when there were unescaped parentheses in a character class, parentheses
3134 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3135
3136 33. "Subroutine" calls and backreferences were previously restricted to
3137 referencing subpatterns earlier in the regex. This restriction has now
3138 been removed.
3139
3140 34. Added a number of extra features that are going to be in Perl 5.10. On the
3141 whole, these are just syntactic alternatives for features that PCRE had
3142 previously implemented using the Python syntax or my own invention. The
3143 other formats are all retained for compatibility.
3144
3145 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3146 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3147 also .NET compatible.
3148
3149 (b) A recursion or subroutine call to a named group can now be defined as
3150 (?&name) as well as (?P>name).
3151
3152 (c) A backreference to a named group can now be defined as \k<name> or
3153 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3154 5.10, are also .NET compatible.
3155
3156 (d) A conditional reference to a named group can now use the syntax
3157 (?(<name>) or (?('name') as well as (?(name).
3158
3159 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3160 groups (named and numbered) that are never evaluated inline, but can be
3161 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3162 is always false. There may be only one alternative in such a group.
3163
3164 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3165 as the simple (?(R). The condition is true only if the most recent
3166 recursion is that of the given number or name. It does not search out
3167 through the entire recursion stack.
3168
3169 (g) The escape \gN or \g{N} has been added, where N is a positive or
3170 negative number, specifying an absolute or relative reference.
3171
3172 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3173 some "unreachable code" warnings.
3174
3175 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3176 things, this adds five new scripts.
3177
3178 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3179 There were also incompatibilities regarding the handling of \Q..\E inside
3180 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3181 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3182
3183 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3184 matches an empty string, and forcibly breaks the loop. There were bugs in
3185 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3186 against aaaa the result was just "a" rather than "aaaa", for example. Two
3187 separate and independent bugs (that affected different cases) have been
3188 fixed.
3189
3190 39. Refactored the code to abolish the use of different opcodes for small
3191 capturing bracket numbers. This is a tidy that I avoided doing when I
3192 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3193 The new approach is not only tidier, it makes it possible to reduce the
3194 memory needed to fix the previous bug (38).
3195
3196 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3197 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3198 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3199 mode.
3200
3201 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3202 report.
3203
3204 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3205 copy construction and assignment in the C++ wrapper.
3206
3207 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3208 couple of bugs that could have given wrong results in the "--newline=crlf"
3209 case.
3210
3211 44. Added a number of casts and did some reorganization of signed/unsigned int
3212 variables following suggestions from Dair Grant. Also renamed the variable
3213 "this" as "item" because it is a C++ keyword.
3214
3215 45. Arranged for dftables to add
3216
3217 #include "pcre_internal.h"
3218
3219 to pcre_chartables.c because without it, gcc 4.x may remove the array
3220 definition from the final binary if PCRE is built into a static library and
3221 dead code stripping is activated.
3222
3223 46. For an unanchored pattern, if a match attempt fails at the start of a
3224 newline sequence, and the newline setting is CRLF or ANY, and the next two
3225 characters are CRLF, advance by two characters instead of one.
3226
3227
3228 Version 6.7 04-Jul-06
3229 ---------------------
3230
3231 1. In order to handle tests when input lines are enormously long, pcretest has
3232 been re-factored so that it automatically extends its buffers when
3233 necessary. The code is crude, but this _is_ just a test program. The
3234 default size has been increased from 32K to 50K.
3235
3236 2. The code in pcre_study() was using the value of the re argument before
3237 testing it for NULL. (Of course, in any sensible call of the function, it
3238 won't be NULL.)
3239
3240 3. The memmove() emulation function in pcre_internal.h, which is used on
3241 systems that lack both memmove() and bcopy() - that is, hardly ever -
3242 was missing a "static" storage class specifier.
3243
3244 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3245 containing an extended class (one that cannot be represented by a bitmap
3246 because it contains high-valued characters or Unicode property items, e.g.
3247 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3248 pattern, but PCRE should not loop if you do not (it no longer does).
3249 [Detail: two cases were found: (a) a repeated subpattern containing an
3250 extended class; (b) a recursive reference to a subpattern that followed a
3251 previous extended class. It wasn't skipping over the extended class
3252 correctly when UTF-8 mode was not set.]
3253
3254 5. A negated single-character class was not being recognized as fixed-length
3255 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3256 compile error "lookbehind assertion is not fixed length".
3257
3258 6. The RunPerlTest auxiliary script was showing an unexpected difference
3259 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3260 write a Perl script that can interpret lines of an input file either as
3261 byte characters or as UTF-8, which is what "perltest" was being required to
3262 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3263 can't do is switch easily at run time between having the "use utf8;" pragma
3264 or not. In the end, I fudged it by using the RunPerlTest script to insert
3265 "use utf8;" explicitly for the UTF-8 tests.
3266
3267 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3268 the end of the subject string, contrary to the documentation and to what
3269 Perl does. This was true of both matching functions. Now it matches only at
3270 the start of the subject and immediately after *internal* newlines.
3271
3272 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3273 a pointer to an int instead of a pointer to an unsigned long int. This
3274 caused problems on 64-bit systems.
3275
3276 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3277 instance of the 'standard' template library not being so standard".
3278
3279 10. There was no check on the number of named subpatterns nor the maximum
3280 length of a subpattern name. The product of these values is used to compute
3281 the size of the memory block for a compiled pattern. By supplying a very
3282 long subpattern name and a large number of named subpatterns, the size
3283 computation could be caused to overflow. This is now prevented by limiting
3284 the length of names to 32 characters, and the number of named subpatterns
3285 to 10,000.
3286
3287 11. Subpatterns that are repeated with specific counts have to be replicated in
3288 the compiled pattern. The size of memory for this was computed from the
3289 length of the subpattern and the repeat count. The latter is limited to
3290 65535, but there was no limit on the former, meaning that integer overflow
3291 could in principle occur. The compiled length of a repeated subpattern is
3292 now limited to 30,000 bytes in order to prevent this.
3293
3294 12. Added the optional facility to have named substrings with the same name.
3295
3296 13. Added the ability to use a named substring as a condition, using the
3297 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3298 are numbers (not recommended). Forward references are permitted.
3299
3300 14. Added forward references in named backreferences (if you see what I mean).
3301
3302 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3303 pattern could run off the end of the subject. For example, the pattern
3304 "(?s)(.{1,5})"8 did this with the subject "ab".
3305
3306 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3307 PCRE_CASELESS was set when matching characters that were quantified with ?
3308 or *.
3309
3310 17. A character class other than a single negated character that had a minimum
3311 but no maximum quantifier - for example [ab]{6,} - was not handled
3312 correctly by pce_dfa_exec(). It would match only one character.
3313
3314 18. A valid (though odd) pattern that looked like a POSIX character
3315 class but used an invalid character after [ (for example [[,abc,]]) caused
3316 pcre_compile() to give the error "Failed: internal error: code overflow" or
3317 in some cases to crash with a glibc free() error. This could even happen if
3318 the pattern terminated after [[ but there just happened to be a sequence of
3319 letters, a binary zero, and a closing ] in the memory that followed.
3320
3321 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3322 over the years. Originally (before any Unicode support), just the bottom 8
3323 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3324 output from "man perlunicode" includes this:
3325
3326 The regular expression compiler produces polymorphic opcodes. That
3327 is, the pattern adapts to the data and automatically switches to
3328 the Unicode character scheme when presented with Unicode data--or
3329 instead uses a traditional byte scheme when presented with byte
3330 data.
3331
3332 Sadly, a wide octal escape does not cause a switch, and in a string with
3333 no other multibyte characters, these octal escapes are treated as before.
3334 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3335 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3336 Unicode string.
3337
3338 I have not perpetrated such confusion in PCRE. Up till now, it took just
3339 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3340 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3341 translate to the appropriate multibyte character.
3342
3343 29. Applied some refactoring to reduce the number of warnings from Microsoft
3344 and Borland compilers. This has included removing the fudge introduced
3345 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3346 a warning about an unused variable.
3347
3348 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3349 characters since release 4.0, because Perl (from release 5.004) does not.
3350 [Or at least, is documented not to: some releases seem to be in conflict
3351 with the documentation.] However, when a pattern was studied with
3352 pcre_study() and all its branches started with \s, PCRE still included VT
3353 as a possible starting character. Of course, this did no harm; it just
3354 caused an unnecessary match attempt.
3355
3356 22. Removed a now-redundant internal flag bit that recorded the fact that case
3357 dependency changed within the pattern. This was once needed for "required
3358 byte" processing, but is no longer used. This recovers a now-scarce options
3359 bit. Also moved the least significant internal flag bit to the most-
3360 significant bit of the word, which was not previously used (hangover from
3361 the days when it was an int rather than a uint) to free up another bit for
3362 the future.
3363
3364 23. Added support for CRLF line endings as well as CR and LF. As well as the
3365 default being selectable at build time, it can now be changed at runtime
3366 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3367 specify that it is scanning data with non-default line endings.
3368
3369 24. Changed the definition of CXXLINK to make it agree with the definition of
3370 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3371
3372 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3373 recursions. This makes a big different to stack usage for some patterns.
3374
3375 26. If a subpattern containing a named recursion or subroutine reference such
3376 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3377 the space required for the compiled pattern went wrong and gave too small a
3378 value. Depending on the environment, this could lead to "Failed: internal
3379 error: code overflow at offset 49" or "glibc detected double free or
3380 corruption" errors.
3381
3382 27. Applied patches from Google (a) to support the new newline modes and (b) to
3383 advance over multibyte UTF-8 characters in GlobalReplace.
3384
3385 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3386 difference for some implementation of PCRE in some Windows version.
3387
3388 29. Added some extra testing facilities to pcretest:
3389
3390 \q<number> in a data line sets the "match limit" value
3391 \Q<number> in a data line sets the "match recursion limt" value
3392 -S <number> sets the stack size, where <number> is in megabytes
3393
3394 The -S option isn't available for Windows.
3395
3396
3397 Version 6.6 06-Feb-06
3398 ---------------------
3399
3400 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3401 in pcreposix.h. I have copied the definition from pcre.h.
3402
3403 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3404 because pcre.h is no longer a built file.
3405
3406 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3407 not normally included in the compiled code.
3408
3409
3410 Version 6.5 01-Feb-06
3411 ---------------------
3412
3413 1. When using the partial match feature with pcre_dfa_exec(), it was not
3414 anchoring the second and subsequent partial matches at the new starting
3415 point. This could lead to incorrect results. For example, with the pattern
3416 /1234/, partially matching against "123" and then "a4" gave a match.
3417
3418 2. Changes to pcregrep:
3419
3420 (a) All non-match returns from pcre_exec() were being treated as failures
3421 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3422 error message is output. Some extra information is given for the
3423 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3424 probably the only errors that are likely to be caused by users (by
3425 specifying a regex that has nested indefinite repeats, for instance).
3426 If there are more than 20 of these errors, pcregrep is abandoned.
3427
3428 (b) A binary zero was treated as data while matching, but terminated the
3429 output line if it was written out. This has been fixed: binary zeroes
3430 are now no different to any other data bytes.
3431
3432 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3433 used to set a locale for matching. The --locale=xxxx long option has
3434 been added (no short equivalent) to specify a locale explicitly on the
3435 pcregrep command, overriding the environment variables.
3436
3437 (d) When -B was used with -n, some line numbers in the output were one less
3438 than they should have been.
3439
3440 (e) Added the -o (--only-matching) option.
3441
3442 (f) If -A or -C was used with -c (count only), some lines of context were
3443 accidentally printed for the final match.
3444
3445 (g) Added the -H (--with-filename) option.
3446
3447 (h) The combination of options -rh failed to suppress file names for files
3448 that were found from directory arguments.
3449
3450 (i) Added the -D (--devices) and -d (--directories) options.
3451
3452 (j) Added the -F (--fixed-strings) option.
3453
3454 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3455
3456 (l) Added the --colo(u)r option.
3457
3458 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3459 is not present by default.
3460
3461 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3462 items such as (?R) or (?1), when the recursion could match a number of
3463 alternatives. If it matched one of the alternatives, but subsequently,
3464 outside the recursion, there was a failure, the code tried to back up into
3465 the recursion. However, because of the way PCRE is implemented, this is not
3466 possible, and the result was an incorrect result from the match.
3467
3468 In order to prevent this happening, the specification of recursion has
3469 been changed so that all such subpatterns are automatically treated as
3470 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3471
3472 4. I had overlooked the fact that, in some locales, there are characters for
3473 which isalpha() is true but neither isupper() nor islower() are true. In
3474 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3475 and ordfeminine) are like this. This affected the treatment of \w and \W
3476 when they appeared in character classes, but not when they appeared outside
3477 a character class. The bit map for "word" characters is now created
3478 separately from the results of isalnum() instead of just taking it from the
3479 upper, lower, and digit maps. (Plus the underscore character, of course.)
3480
3481 5. The above bug also affected the handling of POSIX character classes such as
3482 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3483 permanent tables. Instead, the bit maps for such a class were previously
3484 created as the appropriate unions of the upper, lower, and digit bitmaps.
3485 Now they are created by subtraction from the [[:word:]] class, which has
3486 its own bitmap.
3487
3488 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3489 It is created by subtracting the vertical space characters (\x09, \x0a,
3490 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3491 subtraction was done in the overall bitmap for a character class, meaning
3492 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3493 be recognized. This bug has been fixed.
3494
3495 7. Patches from the folks at Google:
3496
3497 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3498 real life, but is still worth protecting against".
3499
3500 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3501 regular expressions".
3502
3503 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3504 have it.
3505
3506 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3507 "configure" and the latter not, in order to fix a problem somebody had
3508 with compiling the Arg class on HP-UX.
3509
3510 (e) Improve the error-handling of the C++ wrapper a little bit.
3511
3512 (f) New tests for checking recursion limiting.
3513
3514 8. The pcre_memmove() function, which is used only if the environment does not
3515 have a standard memmove() function (and is therefore rarely compiled),
3516 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3517 returning a result (though PCRE never actually uses the result).
3518
3519 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3520 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3521 returned instead of calling malloc() with an overflowing number that would
3522 most likely cause subsequent chaos.
3523
3524 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3525
3526 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3527 with this option is matched, the nmatch and pmatch options of regexec() are
3528 ignored.
3529
3530 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3531 provided in case anyone wants to the the POSIX interface with UTF-8
3532 strings.
3533
3534 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3535 C++ linking (needed for some HP-UX environments).
3536
3537 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3538 (unused parameter) and in the pcre_printint() function (omitted "default"
3539 switch label when the default is to do nothing).
3540
3541 15. Added some code to make it possible, when PCRE is compiled as a C++
3542 library, to replace subject pointers for pcre_exec() with a smart pointer
3543 class, thus making it possible to process discontinuous strings.
3544
3545 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3546 much the same function. They were added by different people who were trying
3547 to make PCRE easy to compile on non-Unix systems. It has been suggested
3548 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3549 for compiling on Windows systems. I have therefore replaced it with
3550 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3551 defaults to "extern" for C or "extern C" for C++, which works fine on
3552 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3553 SCOPE with something explicit in config.h. In addition:
3554
3555 (a) pcreposix.h still had just "extern" instead of either of these macros;
3556 I have replaced it with PCRE_DATA_SCOPE.
3557
3558 (b) Functions such as _pcre_xclass(), which are internal to the library,
3559 but external in the C sense, all had PCRE_EXPORT in their definitions.
3560 This is apparently wrong for the Windows case, so I have removed it.
3561 (It makes no difference on Unix-like systems.)
3562
3563 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3564 of recursive calls to match(). This is different to MATCH_LIMIT because
3565 that limits the total number of calls to match(), not all of which increase
3566 the depth of recursion. Limiting the recursion depth limits the amount of
3567 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3568 when PCRE is compiled, and changed at run time. A patch from Google adds
3569 this functionality to the C++ interface.
3570
3571 18. Changes to the handling of Unicode character properties:
3572
3573 (a) Updated the table to Unicode 4.1.0.
3574
3575 (b) Recognize characters that are not in the table as "Cn" (undefined).
3576
3577 (c) I revised the way the table is implemented to a much improved format
3578 which includes recognition of ranges. It now supports the ranges that
3579 are defined in UnicodeData.txt, and it also amalgamates other
3580 characters into ranges. This has reduced the number of entries in the
3581 table from around 16,000 to around 3,000, thus reducing its size
3582 considerably. I realized I did not need to use a tree structure after
3583 all - a binary chop search is just as efficient. Having reduced the
3584 number of entries, I extended their size from 6 bytes to 8 bytes to
3585 allow for more data.
3586
3587 (d) Added support for Unicode script names via properties such as \p{Han}.
3588
3589 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3590 matching that character.
3591
3592 20. When matching a repeated Unicode property with a minimum greater than zero,
3593 (for example \pL{2,}), PCRE could look past the end of the subject if it
3594 reached it while seeking the minimum number of characters. This could
3595 happen only if some of the characters were more than one byte long, because
3596 there is a check for at least the minimum number of bytes.
3597
3598 21. Refactored the implementation of \p and \P so as to be more general, to
3599 allow for more different types of property in future. This has changed the
3600 compiled form incompatibly. Anybody with saved compiled patterns that use
3601 \p or \P will have to recompile them.
3602
3603 22. Added "Any" and "L&" to the supported property types.
3604
3605 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3606 but give a compile time error if the value is greater than 0xff.
3607
3608 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3609 accidentally not being installed or uninstalled.
3610
3611 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3612 made were to insert the current release number. This seemed silly, because
3613 it made things harder for people building PCRE on systems that don't run
3614 "configure". I have turned pcre.h into a distributed file, no longer built
3615 by "configure", with the version identification directly included. There is
3616 no longer a pcre.h.in file.
3617
3618 However, this change necessitated a change to the pcre-config script as
3619 well. It is built from pcre-config.in, and one of the substitutions was the
3620 release number. I have updated configure.ac so that ./configure now finds
3621 the release number by grepping pcre.h.
3622
3623 26. Added the ability to run the tests under valgrind.
3624
3625
3626 Version 6.4 05-Sep-05
3627 ---------------------
3628
3629 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3630 "--" to be printed when multiple files were scanned, even when none of the
3631 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3632 consider it to be a bug, and have restored the previous behaviour.
3633
3634 2. A couple of code tidies to get rid of compiler warnings.
3635
3636 3. The pcretest program used to cheat by referring to symbols in the library
3637 whose names begin with _pcre_. These are internal symbols that are not
3638 really supposed to be visible externally, and in some environments it is
3639 possible to suppress them. The cheating is now confined to including
3640 certain files from the library's source, which is a bit cleaner.
3641
3642 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3643 file's purpose clearer.
3644
3645 5. Reorganized pcre_ucp_findchar().
3646
3647
3648 Version 6.3 15-Aug-05
3649 ---------------------
3650
3651 1. The file libpcre.pc.in did not have general read permission in the tarball.
3652
3653 2. There were some problems when building without C++ support:
3654
3655 (a) If C++ support was not built, "make install" and "make test" still
3656 tried to test it.
3657
3658 (b) There were problems when the value of CXX was explicitly set. Some
3659 changes have been made to try to fix these, and ...
3660
3661 (c) --disable-cpp can now be used to explicitly disable C++ support.
3662
3663 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3664 backslash in a target when C++ was disabled. This confuses some
3665 versions of "make", apparently. Using an intermediate variable solves
3666 this. (Same for CPP_LOBJ.)
3667
3668 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3669 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3670 necessary on certain architectures.
3671
3672 4. Added a setting of -export-symbols-regex to the link command to remove
3673 those symbols that are exported in the C sense, but actually are local
3674 within the library, and not documented. Their names all begin with
3675 "_pcre_". This is not a perfect job, because (a) we have to except some
3676 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3677 available (and never for static libraries). I have made a note to try to
3678 find a way round (a) in the future.
3679
3680
3681 Version 6.2 01-Aug-05
3682 ---------------------
3683
3684 1. There was no test for integer overflow of quantifier values. A construction
3685 such as {1111111111111111} would give undefined results. What is worse, if
3686 a minimum quantifier for a parenthesized subpattern overflowed and became
3687 negative, the calculation of the memory size went wrong. This could have
3688 led to memory overwriting.
3689
3690 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3691
3692 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3693 operating environments where this matters.
3694
3695 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3696 PCRE options from within the C++ wrapper.
3697
3698 5. Named capturing subpatterns were not being correctly counted when a pattern
3699 was compiled. This caused two problems: (a) If there were more than 100
3700 such subpatterns, the calculation of the memory needed for the whole
3701 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3702 back references of the form \12, where the number was greater than 9, were
3703 not recognized as back references, even though there were sufficient
3704 previous subpatterns.
3705
3706 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3707 versions of gcc, e.g. 2.95.4.
3708
3709
3710 Version 6.1 21-Jun-05
3711 ---------------------
3712
3713 1. There was one reference to the variable "posix" in pcretest.c that was not
3714 surrounded by "#if !defined NOPOSIX".
3715
3716 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3717 the cross-check on the old pcre_info() function, for the benefit of the
3718 cut-down version of PCRE that is currently imported into Exim.
3719
3720 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3721 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3722 patterns that start (?i) but I don't think that matters. The use of (?i) is
3723 just an example; this all applies to the other options as well.
3724
3725 4. Since libtool seems to echo the compile commands it is issuing, the output
3726 from "make" can be reduced a bit by putting "@" in front of each libtool
3727 compile command.
3728
3729 5. Patch from the folks at Google for configure.in to be a bit more thorough
3730 in checking for a suitable C++ installation before trying to compile the
3731 C++ stuff. This should fix a reported problem when a compiler was present,
3732 but no suitable headers.
3733
3734 6. The man pages all had just "PCRE" as their title. I have changed them to
3735 be the relevant file name. I have also arranged that these names are
3736 retained in the file doc/pcre.txt, which is a concatenation in text format
3737 of all the man pages except the little individual ones for each function.
3738
3739 7. The NON-UNIX-USE file had not been updated for the different set of source
3740 files that come with release 6. I also added a few comments about the C++
3741 wrapper.
3742
3743
3744 Version 6.0 07-Jun-05
3745 ---------------------
3746
3747 1. Some minor internal re-organization to help with my DFA experiments.
3748
3749 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3750 didn't matter for the library itself when fully configured, but did matter
3751 when compiling without UCP support, or within Exim, where the ucp files are
3752 not imported.
3753
3754 3. Refactoring of the library code to split up the various functions into
3755 different source modules. The addition of the new DFA matching code (see
3756 below) to a single monolithic source would have made it really too
3757 unwieldy, quite apart from causing all the code to be include in a
3758 statically linked application, when only some functions are used. This is
3759 relevant even without the DFA addition now that patterns can be compiled in
3760 one application and matched in another.
3761
3762 The downside of splitting up is that there have to be some external
3763 functions and data tables that are used internally in different modules of
3764 the library but which are not part of the API. These have all had their
3765 names changed to start with "_pcre_" so that they are unlikely to clash
3766 with other external names.
3767
3768 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3769 a different (DFA) algorithm. Although it is slower than the original
3770 function, it does have some advantages for certain types of matching
3771 problem.
3772
3773 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3774 including restarting after a partial match.
3775
3776 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3777 defined when compiling for Windows was sent to me. I have put it into the
3778 code, though I have no means of testing or verifying it.
3779
3780 7. Added the pcre_refcount() auxiliary function.
3781
3782 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3783 match before or at the first newline in the subject string. In pcretest,
3784 the /f option on a pattern can be used to set this.
3785
3786 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3787 would behave wrongly. This has been present in PCRE since release 4.0.
3788
3789 10. A number of changes to the pcregrep command:
3790
3791 (a) Refactored how -x works; insert ^(...)$ instead of setting
3792 PCRE_ANCHORED and checking the length, in preparation for adding
3793 something similar for -w.
3794
3795 (b) Added the -w (match as a word) option.
3796
3797 (c) Refactored the way lines are read and buffered so as to have more
3798 than one at a time available.
3799
3800 (d) Implemented a pcregrep test script.
3801
3802 (e) Added the -M (multiline match) option. This allows patterns to match
3803 over several lines of the subject. The buffering ensures that at least
3804 8K, or the rest of the document (whichever is the shorter) is available
3805 for matching (and similarly the previous 8K for lookbehind assertions).
3806
3807 (f) Changed the --help output so that it now says
3808
3809 -w, --word-regex(p)
3810
3811 instead of two lines, one with "regex" and the other with "regexp"
3812 because that confused at least one person since the short forms are the
3813 same. (This required a bit of code, as the output is generated
3814 automatically from a table. It wasn't just a text change.)
3815
3816 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3817 option but starts with a hyphen. Could be a pattern or a path name
3818 starting with a hyphen, for instance.
3819
3820 (h) "-" can be given as a file name to represent stdin.
3821
3822 (i) When file names are being printed, "(standard input)" is used for
3823 the standard input, for compatibility with GNU grep. Previously
3824 "<stdin>" was used.
3825
3826 (j) The option --label=xxx can be used to supply a name to be used for
3827 stdin when file names are being printed. There is no short form.
3828
3829 (k) Re-factored the options decoding logic because we are going to add
3830 two more options that take data. Such options can now be given in four
3831 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3832
3833 (l) Added the -A, -B, and -C options for requesting that lines of context
3834 around matches be printed.
3835
3836 (m) Added the -L option to print the names of files that do not contain
3837 any matching lines, that is, the complement of -l.
3838
3839 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3840 continue to scan other files.
3841
3842 (o) The -s option was incorrectly implemented. For compatibility with other
3843 greps, it now suppresses the error message for a non-existent or non-
3844 accessible file (but not the return code). There is a new option called
3845 -q that suppresses the output of matching lines, which was what -s was
3846 previously doing.
3847
3848 (p) Added --include and --exclude options to specify files for inclusion
3849 and exclusion when recursing.
3850
3851 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3852 Hopefully, it now does.
3853
3854 12. Missing cast in pcre_study().
3855
3856 13. Added an "uninstall" target to the makefile.
3857
3858 14. Replaced "extern" in the function prototypes in Makefile.in with
3859 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3860 world, but is set differently for Windows.
3861
3862 15. Added a second compiling function called pcre_compile2(). The only
3863 difference is that it has an extra argument, which is a pointer to an
3864 integer error code. When there is a compile-time failure, this is set
3865 non-zero, in addition to the error test pointer being set to point to an
3866 error message. The new argument may be NULL if no error number is required
3867 (but then you may as well call pcre_compile(), which is now just a
3868 wrapper). This facility is provided because some applications need a
3869 numeric error indication, but it has also enabled me to tidy up the way
3870 compile-time errors are handled in the POSIX wrapper.
3871
3872 16. Added VPATH=.libs to the makefile; this should help when building with one
3873 prefix path and installing with another. (Or so I'm told by someone who
3874 knows more about this stuff than I do.)
3875
3876 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3877 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3878 match everything, including newlines. This is not POSIX-compatible, but
3879 somebody wanted the feature. From pcretest it can be activated by using
3880 both the P and the s flags.
3881
3882 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3883
3884 19. libpcre.pc was being incorrectly installed as executable.
3885
3886 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3887 it now also looks for '\r' so that it will work unmodified on Windows.
3888
3889 21. Added Google's contributed C++ wrapper to the distribution.
3890
3891 22. Added some untidy missing memory free() calls in pcretest, to keep
3892 Electric Fence happy when testing.
3893
3894
3895
3896 Version 5.0 13-Sep-04
3897 ---------------------
3898
3899 1. Internal change: literal characters are no longer packed up into items
3900 containing multiple characters in a single byte-string. Each character
3901 is now matched using a separate opcode. However, there may be more than one
3902 byte in the character in UTF-8 mode.
3903
3904 2. The pcre_callout_block structure has two new fields: pattern_position and
3905 next_item_length. These contain the offset in the pattern to the next match
3906 item, and its length, respectively.
3907
3908 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3909 insertion of callouts before each pattern item. Added the /C option to
3910 pcretest to make use of this.
3911
3912 4. On the advice of a Windows user, the lines
3913
3914 #if defined(_WIN32) || defined(WIN32)
3915 _setmode( _fileno( stdout ), 0x8000 );
3916 #endif /* defined(_WIN32) || defined(WIN32) */
3917
3918 have been added to the source of pcretest. This apparently does useful
3919 magic in relation to line terminators.
3920
3921 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3922 for the benefit of those environments where the "b" makes a difference.
3923
3924 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3925 to know about it. I have put a hack into configure.in that adds in code
3926 to set GCC=yes if CC=icc. This seems to end up at a point in the
3927 generated configure script that is early enough to affect the setting of
3928 compiler options, which is what is needed, but I have no means of testing
3929 whether it really works. (The user who reported this had patched the
3930 generated configure script, which of course I cannot do.)
3931
3932 LATER: After change 22 below (new libtool files), the configure script
3933 seems to know about icc (and also ecc). Therefore, I have commented out
3934 this hack in configure.in.
3935
3936 7. Added support for pkg-config (2 patches were sent in).
3937
3938 8. Negated POSIX character classes that used a combination of internal tables
3939 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3940 [[:^ascii]]. Typically, they would match almost any characters. The other
3941 POSIX classes were not broken in this way.
3942
3943 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3944 to find the match, as PCRE was deluded into thinking that the match had to
3945 start at the start point or following a newline. The same bug applied to
3946 patterns with negative forward assertions or any backward assertions
3947 preceding ".*" at the start, unless the pattern required a fixed first
3948 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3949
3950 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3951 starting at the last subject character, bytes beyond the end of the subject
3952 string were read.
3953
3954 11. Renamed the variable "class" as "classbits" to make life easier for C++
3955 users. (Previously there was a macro definition, but it apparently wasn't
3956 enough.)
3957
3958 12. Added the new field "tables" to the extra data so that tables can be passed
3959 in at exec time, or the internal tables can be re-selected. This allows
3960 a compiled regex to be saved and re-used at a later time by a different
3961 program that might have everything at different addresses.
3962
3963 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3964 -R library as well as a -L library.
3965
3966 14. The debugging options of pcretest (-d on the command line or D on a
3967 pattern) showed incorrect output for anything following an extended class
3968 that contained multibyte characters and which was followed by a quantifier.
3969
3970 15. Added optional support for general category Unicode character properties
3971 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3972 support. It adds about 90K to the size of the library. The meanings of the
3973 inbuilt class escapes such as \d and \s have NOT been changed.
3974
3975 16. Updated pcredemo.c to include calls to free() to release the memory for the
3976 compiled pattern.
3977
3978 17. The generated file chartables.c was being created in the source directory
3979 instead of in the building directory. This caused the build to fail if the
3980 source directory was different from the building directory, and was
3981 read-only.
3982
3983 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3984 file. No doubt somebody will tell me if they don't make sense... Also added
3985 Dan Mooney's comments about building on OpenVMS.
3986
3987 19. Added support for partial matching via the PCRE_PARTIAL option for
3988 pcre_exec() and the \P data escape in pcretest.
3989
3990 20. Extended pcretest with 3 new pattern features:
3991
3992 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3993 write the compiled pattern to the file whose name is "rest-of-line".
3994 This is a straight binary dump of the data, with the saved pointer to
3995 the character tables forced to be NULL. The study data, if any, is
3996 written too. After writing, pcretest reads a new pattern.
3997
3998 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3999 compiled pattern from the given file. There must not be any
4000 occurrences of "<" in the file name (pretty unlikely); if there are,
4001 pcretest will instead treat the initial "<" as a pattern delimiter.
4002 After reading in the pattern, pcretest goes on to read data lines as
4003 usual.
4004
4005 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4006 and 16-bit fields in a compiled pattern, to simulate a pattern that
4007 was compiled on a host of opposite endianness.
4008
4009 21. The pcre-exec() function can now cope with patterns that were compiled on
4010 hosts of opposite endianness, with this restriction:
4011
4012 As for any compiled expression that is saved and used later, the tables
4013 pointer field cannot be preserved; the extra_data field in the arguments
4014 to pcre_exec() should be used to pass in a tables address if a value
4015 other than the default internal tables were used at compile time.
4016
4017 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4018 now diagnosed as an error. Previously, most of the time, a negative number
4019 would have been treated as zero, but if in addition "ovector" was passed as
4020 NULL, a crash could occur.
4021
4022 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4023 new versions from the libtool 1.5 distribution (the last one is a copy of
4024 a file called libtool.m4). This seems to have fixed the need to patch
4025 "configure" to support Darwin 1.3 (which I used to do). However, I still
4026 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4027 workstation).
4028
4029 24. Changed the PCRE licence to be the more standard "BSD" licence.
4030
4031
4032 Version 4.5 01-Dec-03
4033 ---------------------
4034
4035 1. There has been some re-arrangement of the code for the match() function so
4036 that it can be compiled in a version that does not call itself recursively.
4037 Instead, it keeps those local variables that need separate instances for
4038 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4039 needs to "recurse". Keeping track of where control must go is done by means
4040 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4041 hide most of the details from the main code, and operates only if
4042 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4043 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4044 operating.
4045
4046 To make it easier for callers to provide specially tailored get/free
4047 functions for this usage, two new functions, pcre_stack_malloc, and
4048 pcre_stack_free, are used. They are always called in strict stacking order,
4049 and the size of block requested is always the same.
4050
4051 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4052 PCRE has been compiled to use the stack or the heap for recursion. The
4053 -C option of pcretest uses this to show which version is compiled.
4054
4055 A new data escape \S, is added to pcretest; it causes the amounts of store
4056 obtained and freed by both kinds of malloc/free at match time to be added
4057 to the output.
4058
4059 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4060 what's available on my current Linux desktop machine.
4061
4062 3. When matching a UTF-8 string, the test for a valid string at the start has
4063 been extended. If start_offset is not zero, PCRE now checks that it points
4064 to a byte that is the start of a UTF-8 character. If not, it returns
4065 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4066 this is necessary because there may be backward assertions in the pattern.
4067 When matching the same subject several times, it may save resources to use
4068 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4069
4070 4. The code for checking the validity of UTF-8 strings has been tightened so
4071 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4072 containing "overlong sequences".
4073
4074 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4075 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4076 should have been "&", but it just so happened that all the cases this let
4077 through by mistake were picked up later in the function.
4078
4079 6. I had used a variable called "isblank" - this is a C99 function, causing
4080 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4081
4082 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4083 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4084 so the version is shown; (c) stop "make test" echoing "./RunTest".
4085
4086 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4087
4088 9. The prototype for memmove() for systems that don't have it was using
4089 size_t, but the inclusion of the header that defines size_t was later. I've
4090 moved the #includes for the C headers earlier to avoid this.
4091
4092 10. Added some adjustments to the code to make it easier to compiler on certain
4093 special systems:
4094
4095 (a) Some "const" qualifiers were missing.
4096 (b) Added the macro EXPORT before all exported functions; by default this
4097 is defined to be empty.
4098 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4099 that it reads its output file name as an argument instead of writing
4100 to the standard output and assuming this can be redirected.
4101
4102 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4103 class containing characters with values greater than 255, PCRE compilation
4104 went into a loop.
4105
4106 12. A recursive reference to a subpattern that was within another subpattern
4107 that had a minimum quantifier of zero caused PCRE to crash. For example,
4108 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4109 recursion. If the recursively-called subpattern itself had a zero repeat,
4110 that was OK.
4111
4112 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4113 buffer into which it was copied (for escape processing) was still set at
4114 1024, so long lines caused crashes.
4115
4116 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4117 "internal error: code overflow...". This applied to any character class
4118 that was followed by a possessive quantifier.
4119
4120 15. Modified the Makefile to add libpcre.la as a prerequisite for
4121 libpcreposix.la because I was told this is needed for a parallel build to
4122 work.
4123
4124 16. If a pattern that contained .* following optional items at the start was
4125 studied, the wrong optimizing data was generated, leading to matching
4126 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4127 matching string must start with a or b or c. The correct conclusion for
4128 this pattern is that a match can start with any character.
4129
4130
4131 Version 4.4 13-Aug-03
4132 ---------------------
4133
4134 1. In UTF-8 mode, a character class containing characters with values between
4135 127 and 255 was not handled correctly if the compiled pattern was studied.
4136 In fixing this, I have also improved the studying algorithm for such
4137 classes (slightly).
4138
4139 2. Three internal functions had redundant arguments passed to them. Removal
4140 might give a very teeny performance improvement.
4141
4142 3. Documentation bug: the value of the capture_top field in a callout is *one
4143 more than* the number of the hightest numbered captured substring.
4144
4145 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4146 in incorrectly linking with a previously installed version. They now link
4147 explicitly with libpcre.la.
4148
4149 5. configure.in no longer needs to recognize Cygwin specially.
4150
4151 6. A problem in pcre.in for Windows platforms is fixed.
4152
4153 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4154 pcretest, it used to include the size of the study block as part of its
4155 output. Unfortunately, the structure contains a field that has a different
4156 size on different hardware architectures. This meant that the tests that
4157 showed this size failed. As the block is currently always of a fixed size,
4158 this information isn't actually particularly useful in pcretest output, so
4159 I have just removed it.
4160
4161 8. Three pre-processor statements accidentally did not start in column 1.
4162 Sadly, there are *still* compilers around that complain, even though
4163 standard C has not required this for well over a decade. Sigh.
4164
4165 9. In pcretest, the code for checking callouts passed small integers in the
4166 callout_data field, which is a void * field. However, some picky compilers
4167 complained about the casts involved for this on 64-bit systems. Now
4168 pcretest passes the address of the small integer instead, which should get
4169 rid of the warnings.
4170
4171 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4172 both compile and run time, and gives an error if an invalid UTF-8 sequence
4173 is found. There is a option for disabling this check in cases where the
4174 string is known to be correct and/or the maximum performance is wanted.
4175
4176 11. In response to a bug report, I changed one line in Makefile.in from
4177
4178 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4179 to
4180 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4181
4182 to look similar to other lines, but I have no way of telling whether this
4183 is the right thing to do, as I do not use Windows. No doubt I'll get told
4184 if it's wrong...
4185
4186
4187 Version 4.3 21-May-03
4188 ---------------------
4189
4190 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4191 Makefile.
4192
4193 2. Some refactoring to improve the quality of the code:
4194
4195 (i) The utf8_table... variables are now declared "const".
4196
4197 (ii) The code for \cx, which used the "case flipping" table to upper case
4198 lower case letters, now just substracts 32. This is ASCII-specific,
4199 but the whole concept of \cx is ASCII-specific, so it seems
4200 reasonable.
4201
4202 (iii) PCRE was using its character types table to recognize decimal and
4203 hexadecimal digits in the pattern. This is silly, because it handles
4204 only 0-9, a-f, and A-F, but the character types table is locale-
4205 specific, which means strange things might happen. A private
4206 table is now used for this - though it costs 256 bytes, a table is
4207 much faster than multiple explicit tests. Of course, the standard
4208 character types table is still used for matching digits in subject
4209 strings against \d.
4210
4211 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4212 ending in _t are). So I've renamed it as ESC_tee.
4213
4214 3. The first argument for regexec() in the POSIX wrapper should have been
4215 defined as "const".
4216
4217 4. Changed pcretest to use malloc() for its buffers so that they can be
4218 Electric Fenced for debugging.
4219
4220 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4221 to read one or more bytes before the start of the subject string. Often this
4222 had no effect on PCRE's behaviour, but in some circumstances it could
4223 provoke a segmentation fault.
4224
4225 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4226 to try to read one or more bytes before the start of the subject string.
4227
4228 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4229 UTF-8 support could misbehave in various ways if the subject string
4230 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4231 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4232 back over UTF-8 characters.)
4233
4234
4235 Version 4.2 14-Apr-03
4236 ---------------------
4237
4238 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4239
4240 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4241 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4242 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4243 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4244 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4245 and BUILD_EXEEXT
4246 Note: automatic setting of the BUILD variables is not yet working
4247 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4248 compile-time but not at link-time
4249 [LINK]: use for linking executables only
4250 make different versions for Windows and non-Windows
4251 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4252 libraries
4253 [LINK_FOR_BUILD]: new variable
4254 [OBJEXT]: use throughout
4255 [EXEEXT]: use throughout
4256 <winshared>: new target
4257 <wininstall>: new target
4258 <dftables.o>: use native compiler
4259 <dftables>: use native linker
4260 <install>: handle Windows platform correctly
4261 <clean>: ditto
4262 <check>: ditto
4263 copy DLL to top builddir before testing
4264
4265 As part of these changes, -no-undefined was removed again. This was reported
4266 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4267 in any case.
4268
4269 3. Some tidies to get rid of compiler warnings:
4270
4271 . In the match_data structure, match_limit was an unsigned long int, whereas
4272 match_call_count was an int. I've made them both unsigned long ints.
4273
4274 . In pcretest the fact that a const uschar * doesn't automatically cast to
4275 a void * provoked a warning.
4276
4277 . Turning on some more compiler warnings threw up some "shadow" variables
4278 and a few more missing casts.
4279
4280 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4281 option, a class that contained a single character with a value between 128
4282 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4283
4284 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4285 option, a class that contained several characters, but with at least one
4286 whose value was between 128 and 255 caused PCRE to crash.
4287
4288
4289 Version 4.1 12-Mar-03
4290 ---------------------
4291
4292 1. Compiling with gcc -pedantic found a couple of places where casts were
4293 needed, and a string in dftables.c that was longer than standard compilers are
4294 required to support.
4295
4296 2. Compiling with Sun's compiler found a few more places where the code could
4297 be tidied up in order to avoid warnings.
4298
4299 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4300 first of these names is deprecated in the latest Autoconf in favour of the name
4301 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4302 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4303 analogy I have changed it to CFLAGS_FOR_BUILD.
4304
4305 4. Added -no-undefined to the linking command in the Makefile, because this is
4306 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4307 linking step for the pcreposix library.
4308
4309 5. PCRE was failing to diagnose the case of two named groups with the same
4310 name.
4311
4312 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4313 literal character that is needed in the subject for a match, and scans along to
4314 ensure that it is present before embarking on the full matching process. This
4315 saves time in cases of nested unlimited repeats that are never going to match.
4316 Problem: the scan can take a lot of time if the subject is very long (e.g.
4317 megabytes), thus penalizing straightforward matches. It is now done only if the
4318 amount of subject to be scanned is less than 1000 bytes.
4319
4320 7. A lesser problem with the same optimization is that it was recording the
4321 first character of an anchored pattern as "needed", thus provoking a search
4322 right along the subject, even when the first match of the pattern was going to
4323 fail. The "needed" character is now not set for anchored patterns, unless it
4324 follows something in the pattern that is of non-fixed length. Thus, it still
4325 fulfils its original purpose of finding quick non-matches in cases of nested
4326 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4327
4328
4329 Version 4.0 17-Feb-03
4330 ---------------------
4331
4332 1. If a comment in an extended regex that started immediately after a meta-item
4333 extended to the end of string, PCRE compiled incorrect data. This could lead to
4334 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4335
4336 2. Moved to autoconf 2.53 and libtool 1.4.2.
4337
4338 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4339 the special perltest8 script is no longer needed - all the tests can be run
4340 from a single perltest script.
4341
4342 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4343 by \s. It has now been removed in PCRE. This means it isn't recognized as
4344 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4345 class [:space:] *does* include VT, thereby creating a mess.
4346
4347 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4348 space and tab.
4349
4350 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4351 its new features into the main test script, reducing the number of scripts.
4352
4353 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4354 were backward compatible, and made the (?i) apply to the whole pattern, as if
4355 /i were given. Now it behaves more logically, and applies the option setting
4356 only to what follows. PCRE has been changed to follow suit. However, if it
4357 finds options settings right at the start of the pattern, it extracts them into
4358 the global options, as before. Thus, they show up in the info data.
4359
4360 8. Added support for the \Q...\E escape sequence. Characters in between are
4361 treated as literals. This is slightly different from Perl in that $ and @ are
4362 also handled as literals inside the quotes. In Perl, they will cause variable
4363 interpolation. Note the following examples:
4364
4365 Pattern PCRE matches Perl matches
4366
4367 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4368 \Qabc\$xyz\E abc\$xyz abc\$xyz
4369 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4370
4371 For compatibility with Perl, \Q...\E sequences are recognized inside character
4372 classes as well as outside them.
4373
4374 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4375 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4376 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4377 signed/unsigned warnings.
4378
4379 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4380 option for pcretest, so I've replaced it by a simple function that does just
4381 that job.
4382
4383 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4384 "pcregrep -".
4385
4386 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4387 Java package. This provides some syntactic sugar for simple cases of what my
4388 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4389 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4390 item, you can use this simplified notation. Note that only makes sense with
4391 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4392 greediness, whatever the setting of the PCRE_UNGREEDY option.
4393
4394 13. A change of greediness default within a pattern was not taking effect at
4395 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4396 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4397 was abstracted outside.
4398
4399 14. PCRE now supports the \G assertion. It is true when the current matching
4400 position is at the start point of the match. This differs from \A when the
4401 starting offset is non-zero. Used with the /g option of pcretest (or similar
4402 code), it works in the same way as it does for Perl's /g option. If all
4403 alternatives of a regex begin with \G, the expression is anchored to the start
4404 match position, and the "anchored" flag is set in the compiled expression.
4405
4406 15. Some bugs concerning the handling of certain option changes within patterns
4407 have been fixed. These applied to options other than (?ims). For example,
4408 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4409 been the other way round. Some of this was related to change 7 above.
4410
4411 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4412 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4413 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4414 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4415
4416 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4417 mode. Unlike ".", it always matches newline, whatever the setting of
4418 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4419 assertions. Perl allows it, but it doesn't (in general) work because it can't
4420 calculate the length of the lookbehind. At least, that's the case for Perl
4421 5.8.0 - I've been told they are going to document that it doesn't work in
4422 future.
4423
4424 18. Added an error diagnosis for escapes that PCRE does not support: these are
4425 \L, \l, \N, \P, \p, \U, \u, and \X.
4426
4427 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4428 reading past the end of the pattern in cases such as /[abcd/.
4429
4430 20. PCRE was getting more memory than necessary for patterns with classes that
4431 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4432
4433 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4434 compiling PCRE for use with Virtual Pascal.
4435
4436 22. Small fix to the Makefile to make it work properly if the build is done
4437 outside the source tree.
4438
4439 23. Added a new extension: a condition to go with recursion. If a conditional
4440 subpattern starts with (?(R) the "true" branch is used if recursion has
4441 happened, whereas the "false" branch is used only at the top level.
4442
4443 24. When there was a very long string of literal characters (over 255 bytes
4444 without UTF support, over 250 bytes with UTF support), the computation of how
4445 much memory was required could be incorrect, leading to segfaults or other
4446 strange effects.
4447
4448 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4449 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4450 there was a subsequent back reference to those brackets. This meant that, for
4451 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4452 possible to check for precisely this case. All we can do is abandon the
4453 optimization if .* occurs inside capturing brackets when there are any back
4454 references whatsoever. (See below for a better fix that came later.)
4455
4456 26. The handling of the optimization for finding the first character of a
4457 non-anchored pattern, and for finding a character that is required later in the
4458 match were failing in some cases. This didn't break the matching; it just
4459 failed to optimize when it could. The way this is done has been re-implemented.
4460
4461 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4462
4463 28. Added a new feature that provides some of the functionality that Perl
4464 provides with (?{...}). The facility is termed a "callout". The way it is done
4465 in PCRE is for the caller to provide an optional function, by setting
4466 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4467 global variable. By default it is unset, which disables all calling out. To get
4468 the function called, the regex must include (?C) at appropriate points. This
4469 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4470 This provides a means of identifying different callout points. When PCRE
4471 reaches such a point in the regex, if pcre_callout has been set, the external
4472 function is called. It is provided with data in a structure called
4473 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4474 matching continues; if it returns a non-zero value, the match at the current
4475 point fails. However, backtracking will occur if possible. [This was changed
4476 later and other features added - see item 49 below.]
4477
4478 29. pcretest is upgraded to test the callout functionality. It provides a
4479 callout function that displays information. By default, it shows the start of
4480 the match and the current position in the text. There are some new data escapes
4481 to vary what happens:
4482
4483 \C+ in addition, show current contents of captured substrings
4484 \C- do not supply a callout function
4485 \C!n return 1 when callout number n is reached
4486 \C!n!m return 1 when callout number n is reached for the mth time
4487
4488 30. If pcregrep was called with the -l option and just a single file name, it
4489 output "<stdin>" if a match was found, instead of the file name.
4490
4491 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4492 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4493 pcre_exec(). This saves a malloc/free per call. The default value of
4494 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4495 when configuring.
4496
4497 32. The default maximum size of a compiled pattern is 64K. There have been a
4498 few cases of people hitting this limit. The code now uses macros to handle the
4499 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4500 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4501 configuring. Tests 2 and 5 work only with 2-byte links because they output
4502 debugging information about compiled patterns.
4503
4504 33. Internal code re-arrangements:
4505
4506 (a) Moved the debugging function for printing out a compiled regex into
4507 its own source file (printint.c) and used #include to pull it into
4508 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4509 separate copies.
4510
4511 (b) Defined the list of op-code names for debugging as a macro in
4512 internal.h so that it is next to the definition of the opcodes.
4513
4514 (c) Defined a table of op-code lengths for simpler skipping along compiled
4515 code. This is again a macro in internal.h so that it is next to the
4516 definition of the opcodes.
4517
4518 34. Added support for recursive calls to individual subpatterns, along the
4519 lines of Robin Houston's patch (but implemented somewhat differently).
4520
4521 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4522 allow it to read and process whole directories in Win32. This code was
4523 contributed by Lionel Fourquaux; it has not been tested by me.
4524
4525 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4526 used to name a group. Names consist of alphanumerics and underscores, and must
4527 be unique. Back references use the syntax (?P=name) and recursive calls use
4528 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4529 numbers. The function pcre_fullinfo() can be used after compilation to extract
4530 a name/number map. There are three relevant calls:
4531
4532 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4533 PCRE_INFO_NAMECOUNT yields the number of entries
4534 PCRE_INFO_NAMETABLE yields a pointer to the map.
4535
4536 The map is a vector of fixed-size entries. The size of each entry depends on
4537 the length of the longest name used. The first two bytes of each entry are the
4538 group number, most significant byte first. There follows the corresponding
4539 name, zero terminated. The names are in alphabetical order.
4540
4541 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4542 case instead of 255. Making it the same both with and without UTF-8 support
4543 means that the same test output works with both.
4544
4545 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4546 calling malloc() with a zero argument.
4547
4548 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4549 optimization. I've improved things by keeping a bitmap of backreferences with
4550 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4551 fact referenced, the optimization can be applied. It is unlikely that a
4552 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4553 the match to follow \n) will appear inside brackets with a number greater than
4554 31, but if it does, any back reference > 31 suppresses the optimization.
4555
4556 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4557 of disabling numbered capturing parentheses. Any opening parenthesis that is
4558 not followed by ? behaves as if it were followed by ?: but named parentheses
4559 can still be used for capturing (and they will acquire numbers in the usual
4560 way).
4561
4562 41. Redesigned the return codes from the match() function into yes/no/error so
4563 that errors can be passed back from deep inside the nested calls. A malloc
4564 failure while inside a recursive subpattern call now causes the
4565 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4566
4567 42. It is now possible to set a limit on the number of times the match()
4568 function is called in a call to pcre_exec(). This facility makes it possible to
4569 limit the amount of recursion and backtracking, though not in a directly
4570 obvious way, because the match() function is used in a number of different
4571 circumstances. The count starts from zero for each position in the subject
4572 string (for non-anchored patterns). The default limit is, for compatibility, a
4573 large number, namely 10 000 000. You can change this in two ways:
4574
4575 (a) When configuring PCRE before making, you can use --with-match-limit=n
4576 to set a default value for the compiled library.
4577
4578 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4579 a different value is set. See 45 below.
4580
4581 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4582
4583 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4584 of things that can be changed at compile time. The first argument specifies
4585 what is wanted and the second points to where the information is to be placed.
4586 The current list of available information is:
4587
4588 PCRE_CONFIG_UTF8
4589
4590 The output is an integer that is set to one if UTF-8 support is available;
4591 otherwise it is set to zero.
4592
4593 PCRE_CONFIG_NEWLINE
4594
4595 The output is an integer that it set to the value of the code that is used for
4596 newline. It is either LF (10) or CR (13).
4597
4598 PCRE_CONFIG_LINK_SIZE
4599
4600 The output is an integer that contains the number of bytes used for internal
4601 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4602
4603 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4604
4605 The output is an integer that contains the threshold above which the POSIX
4606 interface uses malloc() for output vectors. See item 31 above.
4607
4608 PCRE_CONFIG_MATCH_LIMIT
4609
4610 The output is an unsigned integer that contains the default limit of the number
4611 of match() calls in a pcre_exec() execution. See 42 above.
4612
4613 44. pcretest has been upgraded by the addition of the -C option. This causes it
4614 to extract all the available output from the new pcre_config() function, and to
4615 output it. The program then exits immediately.
4616
4617 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4618 order to support additional features. One way would have been to define
4619 pcre_exec2() (for example) with extra arguments, but this would not have been
4620 extensible, and would also have required all calls to the original function to
4621 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4622 is used for passing in "extra" data from pcre_study().
4623
4624 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4625 contains the following fields:
4626
4627 flags a bitmap indicating which of the following fields are set
4628 study_data opaque data from pcre_study()
4629 match_limit a way of specifying a limit on match() calls for a specific
4630 call to pcre_exec()
4631 callout_data data for callouts (see 49 below)
4632
4633 The flag bits are also defined in pcre.h, and are
4634
4635 PCRE_EXTRA_STUDY_DATA
4636 PCRE_EXTRA_MATCH_LIMIT
4637 PCRE_EXTRA_CALLOUT_DATA
4638
4639 The pcre_study() function now returns one of these new pcre_extra blocks, with
4640 the actual study data pointed to by the study_data field, and the
4641 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4642 before. That is, this change is entirely upwards-compatible and requires no
4643 change to existing code.
4644
4645 If you want to pass in additional data to pcre_exec(), you can either place it
4646 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4647 block.
4648
4649 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4650 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4651 times with different match limits, until it finds the minimum value needed for
4652 pcre_exec() to complete. The value is then output. This can be instructive; for
4653 most simple matches the number is quite small, but for pathological cases it
4654 gets very large very quickly.
4655
4656 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4657 returns the size of the data block pointed to by the study_data field in a
4658 pcre_extra block, that is, the value that was passed as the argument to
4659 pcre_malloc() when PCRE was getting memory in which to place the information
4660 created by pcre_study(). The fourth argument should point to a size_t variable.
4661 pcretest has been extended so that this information is shown after a successful
4662 pcre_study() call when information about the compiled regex is being displayed.
4663
4664 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4665 because what follows is always an absolute path. (Later: it turns out that this
4666 is more than cosmetic for MinGW, because it doesn't like empty path
4667 components.)
4668
4669 49. Some changes have been made to the callout feature (see 28 above):
4670
4671 (i) A callout function now has three choices for what it returns:
4672
4673 0 => success, carry on matching
4674 > 0 => failure at this point, but backtrack if possible
4675 < 0 => serious error, return this value from pcre_exec()
4676
4677 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4678 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4679 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4680 use by callout functions. It will never be used by PCRE itself.
4681
4682 (ii) The pcre_extra structure (see 45 above) has a void * field called
4683 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4684 pcre_callout_block structure has a field of the same name. The contents of
4685 the field passed in the pcre_extra structure are passed to the callout
4686 function in the corresponding field in the callout block. This makes it
4687 easier to use the same callout-containing regex from multiple threads. For
4688 testing, the pcretest program has a new data escape
4689
4690 \C*n pass the number n (may be negative) as callout_data
4691
4692 If the callout function in pcretest receives a non-zero value as
4693 callout_data, it returns that value.
4694
4695 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4696 there were some redundant $(CFLAGS) in commands that are now specified as
4697 $(LINK), which already includes $(CFLAGS).
4698
4699 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4700 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4701 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4702 one-byte characters throughout. Note that case-insensitive matching applies
4703 only to characters whose values are less than 256. PCRE doesn't support the
4704 notion of cases for higher-valued characters.
4705
4706 (i) A character class whose characters are all within 0-255 is handled as
4707 a bit map, and the map is inverted for negative classes. Previously, a
4708 character > 255 always failed to match such a class; however it should
4709 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4710
4711 (ii) A negated character class with a single character < 255 is coded as
4712 "not this character" (OP_NOT). This wasn't working properly when the test
4713 character was multibyte, either singly or repeated.
4714
4715 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4716 mode, for example: \x{100}{2,3}.
4717
4718 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4719 singly or repeated) now correctly test multibyte characters. However,
4720 PCRE doesn't recognize any characters with values greater than 255 as
4721 digits, spaces, or word characters. Such characters always match \D, \S,
4722 and \W, and never match \d, \s, or \w.
4723
4724 (v) Classes may now contain characters and character ranges with values
4725 greater than 255. For example: [ab\x{100}-\x{400}].
4726
4727 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4728 PCRE in UTF-8 mode.
4729
4730 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4731 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4732 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4733 value.)
4734
4735 53. The single man page has become too large. I have therefore split it up into
4736 a number of separate man pages. These also give rise to individual HTML pages;
4737 these are now put in a separate directory, and there is an index.html page that
4738 lists them all. Some hyperlinking between the pages has been installed.
4739
4740 54. Added convenience functions for handling named capturing parentheses.
4741
4742 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4743 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4744 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4745 are faulted.
4746
4747 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4748 calling configure. These values are used when compiling the dftables.c program
4749 which is run to generate the source of the default character tables. They
4750 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4751 you will need to set these values.
4752
4753 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4754
4755
4756 Version 3.9 02-Jan-02
4757 ---------------------
4758
4759 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4760
4761 2. If --disable-static was given, the building process failed when trying to
4762 build pcretest and pcregrep. (For some reason it was using libtool to compile
4763 them, which is not right, as they aren't part of the library.)
4764
4765
4766 Version 3.8 18-Dec-01
4767 ---------------------
4768
4769 1. The experimental UTF-8 code was completely screwed up. It was packing the
4770 bytes in the wrong order. How dumb can you get?
4771
4772
4773 Version 3.7 29-Oct-01
4774 ---------------------
4775
4776 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4777 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4778 this didn't happen under Solaris 8, where I normally test things.
4779
4780 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4781 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4782 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4783 this fix an hour or so after the initial 3.7 release.)
4784
4785
4786 Version 3.6 23-Oct-01
4787 ---------------------
4788
4789 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4790 offsets passed as NULL with zero offset count.
4791
4792 2. The config.guess and config.sub files had not been updated when I moved to
4793 the latest autoconf.
4794
4795
4796 Version 3.5 15-Aug-01
4797 ---------------------
4798
4799 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4800 had been forgotten.
4801
4802 2. By using declared but undefined structures, we can avoid using "void"
4803 definitions in pcre.h while keeping the internal definitions of the structures
4804 private.
4805
4806 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4807 user point of view, this means that both static and shared libraries are built
4808 by default, but this can be individually controlled. More of the work of
4809 handling this static/shared cases is now inside libtool instead of PCRE's make
4810 file.
4811
4812 4. The pcretest utility is now installed along with pcregrep because it is
4813 useful for users (to test regexs) and by doing this, it automatically gets
4814 relinked by libtool. The documentation has been turned into a man page, so
4815 there are now .1, .txt, and .html versions in /doc.
4816
4817 5. Upgrades to pcregrep:
4818 (i) Added long-form option names like gnu grep.
4819 (ii) Added --help to list all options with an explanatory phrase.
4820 (iii) Added -r, --recursive to recurse into sub-directories.
4821 (iv) Added -f, --file to read patterns from a file.
4822
4823 6. pcre_exec() was referring to its "code" argument before testing that
4824 argument for NULL (and giving an error if it was NULL).
4825
4826 7. Upgraded Makefile.in to allow for compiling in a different directory from
4827 the source directory.
4828
4829 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4830 options bits, the pointer it was passed was to an int instead of to an unsigned
4831 long int. This mattered only on 64-bit systems.
4832
4833 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4834 generated) instead of pcre.in, which it its source. Also made the same change
4835 in several of the .c files.
4836
4837 10. A new release of gcc defines printf() as a macro, which broke pcretest
4838 because it had an ifdef in the middle of a string argument for printf(). Fixed
4839 by using separate calls to printf().
4840
4841 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4842 script, to force use of CR or LF instead of \n in the source. On non-Unix
4843 systems, the value can be set in config.h.
4844
4845 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4846 absolute limit. Changed the text of the error message to make this clear, and
4847 likewise updated the man page.
4848
4849 13. The limit of 99 on the number of capturing subpatterns has been removed.
4850 The new limit is 65535, which I hope will not be a "real" limit.
4851
4852
4853 Version 3.4 22-Aug-00
4854 ---------------------
4855
4856 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4857
4858 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4859
4860
4861 Version 3.3 01-Aug-00
4862 ---------------------
4863
4864 1. If an octal character was given, but the value was greater than \377, it
4865 was not getting masked to the least significant bits, as documented. This could
4866 lead to crashes in some systems.
4867
4868 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4869 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4870
4871 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4872 These just pass their arguments on to (pcre_free)(), but they are provided
4873 because some uses of PCRE bind it to non-C systems that can call its functions,
4874 but cannot call free() or pcre_free() directly.
4875
4876 4. Add "make test" as a synonym for "make check". Corrected some comments in
4877 the Makefile.
4878
4879 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4880 Makefile.
4881
4882 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4883 command called pgrep for grepping around the active processes.
4884
4885 7. Added the beginnings of support for UTF-8 character strings.
4886
4887 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4888 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4889 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4890 out for the ar command.)
4891
4892
4893 Version 3.2 12-May-00
4894 ---------------------
4895
4896 This is purely a bug fixing release.
4897
4898 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4899 of ZA. This was just one example of several cases that could provoke this bug,
4900 which was introduced by change 9 of version 2.00. The code for breaking
4901 infinite loops after an iteration that matches an empty string was't working
4902 correctly.
4903
4904 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4905 when matched against abbab (for example). After matching an empty string, it
4906 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4907 caused it to match further down the string than it should.
4908
4909 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4910 was there because it doesn't seem to be needed, and it causes trouble on some
4911 systems, as it is not a Standard C header. It has been removed.
4912
4913 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4914 were reported on the Macintosh. The changes were from
4915
4916 while ((c = *(++ptr)) != 0 && c != '\n');
4917 to
4918 while ((c = *(++ptr)) != 0 && c != '\n') ;
4919
4920 Totally extraordinary, but if that's what it takes...
4921
4922 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4923 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4924 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4925 assumes the way PCRE uses memmove() (always moving upwards).
4926
4927 6. PCRE is being used in one environment where strchr() is not available. There
4928 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4929 faster code anyway.
4930
4931
4932 Version 3.1 09-Feb-00
4933 ---------------------
4934
4935 The only change in this release is the fixing of some bugs in Makefile.in for
4936 the "install" target:
4937
4938 (1) It was failing to install pcreposix.h.
4939
4940 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4941
4942
4943 Version 3.0 01-Feb-00
4944 ---------------------
4945
4946 1. Add support for the /+ modifier to perltest (to output $` like it does in
4947 pcretest).
4948
4949 2. Add support for the /g modifier to perltest.
4950
4951 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4952 matches null strings.
4953
4954 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4955 pattern. Perl treats empty patterns specially - it reuses the most recent
4956 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4957 effect.
4958
4959 5. The POSIX interface was broken in that it was just handing over the POSIX
4960 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4961 required a bigger vector, with some working space on the end. This means that
4962 the POSIX wrapper now has to get and free some memory, and copy the results.
4963
4964 6. Added some simple autoconf support, placing the test data and the
4965 documentation in separate directories, re-organizing some of the
4966 information files, and making it build pcre-config (a GNU standard). Also added
4967 libtool support for building PCRE as a shared library, which is now the
4968 default.
4969
4970 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4971 09 are not valid octal constants. Single digits will be used for minor values
4972 less than 10.
4973
4974 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4975 existing programs that set these in the POSIX interface can use PCRE without
4976 modification.
4977
4978 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4979 return all that pcre_info() returns, plus additional data. The pcre_info()
4980 function is retained for compatibility, but is considered to be obsolete.
4981
4982 10. Added experimental recursion feature (?R) to handle one common case that
4983 Perl 5.6 will be able to do with (?p{...}).
4984
4985 11. Added support for POSIX character classes like [:alpha:], which Perl is
4986 adopting.
4987
4988
4989 Version 2.08 31-Aug-99
4990 ----------------------
4991
4992 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4993 trying to match at the startoffset position, but instead was moving forward to
4994 the next newline as if a previous match had failed.
4995
4996 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4997 and could get into a loop if a null string was matched other than at the start
4998 of the subject.
4999
5000 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5001 be distinguished at compile time, and for completeness also added PCRE_DATE.
5002
5003 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5004 in GnuWin32 environments.
5005
5006
5007 Version 2.07 29-Jul-99
5008 ----------------------
5009
5010 1. The documentation is now supplied in plain text form and HTML as well as in
5011 the form of man page sources.
5012
5013 2. C++ compilers don't like assigning (void *) values to other pointer types.
5014 In particular this affects malloc(). Although there is no problem in Standard
5015 C, I've put in casts to keep C++ compilers happy.
5016
5017 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5018 should be (const char *).
5019
5020 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5021 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5022 However, I haven't made this a standard facility. The documentation doesn't
5023 mention it, and the Makefile doesn't support it.
5024
5025 5. The Makefile now contains an "install" target, with editable destinations at
5026 the top of the file. The pcretest program is not installed.
5027
5028 6. pgrep -V now gives the PCRE version number and date.
5029
5030 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5031 causing the entire string to be ignored, instead of just the last character.
5032
5033 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5034 non-matching string, it can take a very, very long time, even for strings of
5035 quite modest length, because of the nested recursion. PCRE now does better in
5036 some of these cases. It does this by remembering the last required literal
5037 character in the pattern, and pre-searching the subject to ensure it is present
5038 before running the real match. In other words, it applies a heuristic to detect
5039 some types of certain failure quickly, and in the above example, if presented
5040 with a string that has no trailing " it gives "no match" very quickly.
5041
5042 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5043 other alternatives are tried instead.
5044
5045
5046 Version 2.06 09-Jun-99
5047 ----------------------
5048
5049 1. Change pcretest's output for amount of store used to show just the code
5050 space, because the remainder (the data block) varies in size between 32-bit and
5051 64-bit systems.
5052
5053 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5054 start matching at. This allows lookbehinds to work when searching for multiple
5055 occurrences in a string.
5056
5057 3. Added additional options to pcretest for testing multiple occurrences:
5058
5059 /+ outputs the rest of the string that follows a match
5060 /g loops for multiple occurrences, using the new startoffset argument
5061 /G loops for multiple occurrences by passing an incremented pointer
5062
5063 4. PCRE wasn't doing the "first character" optimization for patterns starting
5064 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5065 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5066 the letter 'x'. On long subject strings, this gives a significant speed-up.
5067
5068
5069 Version 2.05 21-Apr-99
5070 ----------------------
5071
5072 1. Changed the type of magic_number from int to long int so that it works
5073 properly on 16-bit systems.
5074
5075 2. Fixed a bug which caused patterns starting with .* not to work correctly
5076 when the subject string contained newline characters. PCRE was assuming
5077 anchoring for such patterns in all cases, which is not correct because .* will
5078 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5079 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5080 must be retried after every newline in the subject.
5081
5082
5083 Version 2.04 18-Feb-99
5084 ----------------------
5085
5086 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5087 computation of the store needed to hold the pattern was incorrect (too large).
5088 If such patterns were nested a few deep, this could multiply and become a real
5089 problem.
5090
5091 2. Added /M option to pcretest to show the memory requirement of a specific
5092 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5093
5094 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5095 compiled in such a way that the backtracking after subsequent failure was
5096 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5097 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5098
5099
5100 Version 2.03 02-Feb-99
5101 ----------------------
5102
5103 1. Fixed typo and small mistake in man page.
5104
5105 2. Added 4th condition (GPL supersedes if conflict) and created separate
5106 LICENCE file containing the conditions.
5107
5108 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5109 Perl, that is the internal \ allows the delimiter to be included in the
5110 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5111 the final delimiter, add \ to the end of the pattern (to test the error).
5112
5113 4. Added the convenience functions for extracting substrings after a successful
5114 match. Updated pcretest to make it able to test these functions.
5115
5116
5117 Version 2.02 14-Jan-99
5118 ----------------------
5119
5120 1. Initialized the working variables associated with each extraction so that
5121 their saving and restoring doesn't refer to uninitialized store.
5122
5123 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5124 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5125 fix the problem.
5126
5127 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5128 calls, and wasn't printing the correct value for compiling calls. Increased the
5129 default value of LOOPREPEAT, and the number of significant figures in the
5130 times.
5131
5132 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5133
5134 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5135 a building problem on Windows NT with a FAT file system.
5136
5137
5138 Version 2.01 21-Oct-98
5139 ----------------------
5140
5141 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5142 to character tables built by pcre_maketables() in the current locale. If NULL
5143 is passed, the default tables are used.
5144
5145
5146 Version 2.00 24-Sep-98
5147 ----------------------
5148
5149 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5150 it any more.
5151
5152 2. Allow quantification of (?>) groups, and make it work correctly.
5153
5154 3. The first character computation wasn't working for (?>) groups.
5155
5156 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5157 end of the subject) and add 5.005's \z, which really does match only at the
5158 very end of the subject.
5159
5160 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5161
5162 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5163 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5164 localized options. All options to pcre_study() were also removed.
5165
5166 7. Add other new features from 5.005:
5167
5168 $(?<= positive lookbehind
5169 $(?<! negative lookbehind
5170 (?imsx-imsx) added the unsetting capability
5171 such a setting is global if at outer level; local otherwise
5172 (?imsx-imsx:) non-capturing groups with option setting
5173 (?(cond)re|re) conditional pattern matching
5174
5175 A backreference to itself in a repeated group matches the previous
5176 captured string.
5177
5178 8. General tidying up of studying (both automatic and via "study")
5179 consequential on the addition of new assertions.
5180
5181 9. As in 5.005, unlimited repeated groups that could match an empty substring
5182 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5183 runtime if any iteration does actually match an empty substring.
5184
5185 10. Include the RunTest script in the distribution.
5186
5187 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5188 discrepancies, some of which were old and were also with respect to 5.004. They
5189 have now been fixed.
5190
5191
5192 Version 1.09 28-Apr-98
5193 ----------------------
5194
5195 1. A negated single character class followed by a quantifier with a minimum
5196 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5197 program crashes, or just wrong answers. This did not apply to negated classes
5198 containing more than one character, or to minima other than one.
5199
5200
5201 Version 1.08 27-Mar-98
5202 ----------------------
5203
5204 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5205
5206 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5207 latter must appear before anything that relies on it in the pattern.
5208
5209
5210 Version 1.07 16-Feb-98
5211 ----------------------
5212
5213 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5214 repeat of a potentially empty string).
5215
5216
5217 Version 1.06 23-Jan-98
5218 ----------------------
5219
5220 1. Added Markus Oberhumer's little patches for C++.
5221
5222 2. Literal strings longer than 255 characters were broken.
5223
5224
5225 Version 1.05 23-Dec-97
5226 ----------------------
5227
5228 1. Negated character classes containing more than one character were failing if
5229 PCRE_CASELESS was set at run time.
5230
5231
5232 Version 1.04 19-Dec-97
5233 ----------------------
5234
5235 1. Corrected the man page, where some "const" qualifiers had been omitted.
5236
5237 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5238 input syntax.
5239
5240 3. Fixed memory leak which occurred when a regex with back references was
5241 matched with an offsets vector that wasn't big enough. The temporary memory
5242 that is used in this case wasn't being freed if the match failed.
5243
5244 4. Tidied pcretest to ensure it frees memory that it gets.
5245
5246 5. Temporary memory was being obtained in the case where the passed offsets
5247 vector was exactly big enough.
5248
5249 6. Corrected definition of offsetof() from change 5 below.
5250
5251 7. I had screwed up change 6 below and broken the rules for the use of
5252 setjmp(). Now fixed.
5253
5254
5255 Version 1.03 18-Dec-97
5256 ----------------------
5257
5258 1. A erroneous regex with a missing opening parenthesis was correctly
5259 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5260 on some systems.
5261
5262 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5263 it was reported that one broken compiler failed on the former because "code" is
5264 also an independent variable.
5265
5266 3. The erroneous regex a[]b caused an array overrun reference.
5267
5268 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5269 fail on data ending with that character. (It was going on too far, and checking
5270 the next character, typically a binary zero.) This was specific to the
5271 optimized code for single-character negative classes.
5272
5273 5. Added a contributed patch from the TIN world which does the following:
5274
5275 + Add an undef for memmove, in case the the system defines a macro for it.
5276
5277 + Add a definition of offsetof(), in case there isn't one. (I don't know
5278 the reason behind this - offsetof() is part of the ANSI standard - but
5279 it does no harm).
5280
5281 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5282 most of the places where whitespace preceded '#'. I have given up and
5283 allowed the remaining 2 cases to be at the margin.
5284
5285 + Rename some variables in pcre to eliminate shadowing. This seems very
5286 pedantic, but does no harm, of course.
5287
5288 6. Moved the call to setjmp() into its own function, to get rid of warnings
5289 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5290
5291 7. Constructs such as \d{8,} were compiling into the equivalent of
5292 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5293 outcome, but in this particular case used more store than had been allocated,
5294 which caused the bug to be discovered because it threw up an internal error.
5295
5296 8. The debugging code in both pcre and pcretest for outputting the compiled
5297 form of a regex was going wrong in the case of back references followed by
5298 curly-bracketed repeats.
5299
5300
5301 Version 1.02 12-Dec-97
5302 ----------------------
5303
5304 1. Typos in pcre.3 and comments in the source fixed.
5305
5306 2. Applied a contributed patch to get rid of places where it used to remove
5307 'const' from variables, and fixed some signed/unsigned and uninitialized
5308 variable warnings.
5309
5310 3. Added the "runtest" target to Makefile.
5311
5312 4. Set default compiler flag to -O2 rather than just -O.
5313
5314
5315 Version 1.01 19-Nov-97
5316 ----------------------
5317
5318 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5319 like /([ab]*)*/, that is, for classes with more than one character in them.
5320
5321 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5322 as /((?>a*))*/ (a PCRE_EXTRA facility).
5323
5324
5325 Version 1.00 18-Nov-97
5326 ----------------------
5327
5328 1. Added compile-time macros to support systems such as SunOS4 which don't have
5329 memmove() or strerror() but have other things that can be used instead.
5330
5331 2. Arranged that "make clean" removes the executables.
5332
5333
5334 Version 0.99 27-Oct-97
5335 ----------------------
5336
5337 1. Fixed bug in code for optimizing classes with only one character. It was
5338 initializing a 32-byte map regardless, which could cause it to run off the end
5339 of the memory it had got.
5340
5341 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5342
5343
5344 Version 0.98 22-Oct-97
5345 ----------------------
5346
5347 1. Fixed bug in code for handling temporary memory usage when there are more
5348 back references than supplied space in the ovector. This could cause segfaults.
5349
5350
5351 Version 0.97 21-Oct-97
5352 ----------------------
5353
5354 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5355
5356 2. Optimized negated single characters not to use a bit map.
5357
5358 3. Brought error texts together as macro definitions; clarified some of them;
5359 fixed one that was wrong - it said "range out of order" when it meant "invalid
5360 escape sequence".
5361
5362 4. Changed some char * arguments to const char *.
5363
5364 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5365
5366 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5367 pcretest.
5368
5369
5370 Version 0.96 16-Oct-97
5371 ----------------------
5372
5373 1. Added a simple "pgrep" utility to the distribution.
5374
5375 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5376 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5377 where "ddd" means "one or more decimal digits".
5378
5379 3. Fixed serious bug. If a pattern had a back reference, but the call to
5380 pcre_exec() didn't supply a large enough ovector to record the related
5381 identifying subpattern, the match always failed. PCRE now remembers the number
5382 of the largest back reference, and gets some temporary memory in which to save
5383 the offsets during matching if necessary, in order to ensure that
5384 backreferences always work.
5385
5386 4. Increased the compatibility with Perl in a number of ways:
5387
5388 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5389 to request this handling. The option can be set at compile or exec time.
5390
5391 (b) $ matches before a terminating newline by default; an option
5392 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5393 mode). The option can be set at compile or exec time.
5394
5395 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5396 the same as Perl's. If the decimal number it represents is less than 10
5397 or there aren't that many previous left capturing parentheses, an octal
5398 escape is read. Inside a character class, it's always an octal escape,
5399 even if it is a single digit.
5400
5401 (d) An escaped but undefined alphabetic character is taken as a literal,
5402 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5403 escapes.
5404
5405 (e) {0} is now permitted. (The previous item is removed from the compiled
5406 pattern).
5407
5408 5. Changed all the names of code files so that the basic parts are no longer
5409 than 10 characters, and abolished the teeny "globals.c" file.
5410
5411 6. Changed the handling of character classes; they are now done with a 32-byte
5412 bit map always.
5413
5414 7. Added the -d and /D options to pcretest to make it possible to look at the
5415 internals of compilation without having to recompile pcre.
5416
5417
5418 Version 0.95 23-Sep-97
5419 ----------------------
5420
5421 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5422 \x20 at the start of a run of normal characters. These were being treated as
5423 real characters, instead of the source characters being re-checked.
5424
5425
5426 Version 0.94 18-Sep-97
5427 ----------------------
5428
5429 1. The functions are now thread-safe, with the caveat that the global variables
5430 containing pointers to malloc() and free() or alternative functions are the
5431 same for all threads.
5432
5433 2. Get pcre_study() to generate a bitmap of initial characters for non-
5434 anchored patterns when this is possible, and use it if passed to pcre_exec().
5435
5436
5437 Version 0.93 15-Sep-97
5438 ----------------------
5439
5440 1. /(b)|(:+)/ was computing an incorrect first character.
5441
5442 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5443 but not actually doing anything yet.
5444
5445 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5446 as Perl does (e.g. [-az] or [az-]).
5447
5448 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5449 all possible positions.
5450
5451 5. Split up into different modules to avoid including unneeded functions in a
5452 compiled binary. However, compile and exec are still in one module. The "study"
5453 function is split off.
5454
5455 6. The character tables are now in a separate module whose source is generated
5456 by an auxiliary program - but can then be edited by hand if required. There are
5457 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5458 toupper() in the code.
5459
5460 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5461 make them global. Abolish the function for setting them, as the caller can now
5462 set them directly.
5463
5464
5465 Version 0.92 11-Sep-97
5466 ----------------------
5467
5468 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5469 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5470
5471 2. Caseless matching was not working in character classes if the characters in
5472 the pattern were in upper case.
5473
5474 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5475
5476 4. Make PCRE_ANCHORED public and accept as a compile option.
5477
5478 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5479 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5480 pass them.
5481
5482 6. Give an error if bad option bits passed at compile or run time.
5483
5484 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5485 pcretest to cause it to pass that flag.
5486
5487 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5488 options, and the first character, if set.
5489
5490 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5491
5492
5493 Version 0.91 10-Sep-97
5494 ----------------------
5495
5496 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5497 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5498
5499 2. PCRE was looping on encountering an indefinitely repeated back reference to
5500 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5501 Perl does - treats the match as successful.
5502
5503 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5