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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1450 - (show annotations)
Fri Jan 17 17:50:51 2014 UTC (5 years, 9 months ago) by ph10
File size: 263000 byte(s)
Fix bug in tests when 16/32 bits and --enable-bsr-anycrlf are both set. Extend 
pcretest to show the \R default.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.35-RC1 xx-xxxx-201x
5 -----------------------------
6
7 1. A new flag is set, when property checks are present in an XCLASS.
8 When this flag is not set, PCRE can perform certain optimizations
9 such as studying these XCLASS-es.
10
11 2. The auto-possessification of character sets were improved: a normal
12 and an extended character set can be compared now. Furthermore
13 the JIT compiler optimizes more character set checks.
14
15 3. Got rid of some compiler warnings for potentially uninitialized variables
16 that show up only when compiled with -O2.
17
18 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
19 of a match later then the end of the match. The pcretest program was not
20 handling the case sensibly - it was outputting from the start to the next
21 binary zero. It now reports this situation in a message, and outputs the
22 text from the end to the start.
23
24 5. Fast forward search is improved in JIT. Instead of the first three
25 characters, any three characters with fixed position can be searched.
26 Search order: first, last, middle.
27
28 6. Improve character range checks in JIT. Characters are read by an inprecise
29 function now, which returns with an unknown value if the character code is
30 above a certain treshold (e.g: 256). The only limitation is that the value
31 must be bigger than the treshold as well. This function is useful, when
32 the characters above the treshold are handled in the same way.
33
34 7. The macros whose names start with RAWUCHAR are placeholders for a future
35 mode in which only the bottom 21 bits of 32-bit data items are used. To
36 make this more memorable for those maintaining the code, the names have
37 been changed to start with UCHAR21, and an extensive comment has been added
38 to their definition.
39
40 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
41 to the export list in Makefile.am (they were accidentally omitted from the
42 8.34 tarball).
43
44 9. The informational output from pcretest used the phrase "starting byte set"
45 which is inappropriate for the 16-bit and 32-bit libraries. As the output
46 for "first char" and "need char" really means "non-UTF-char", I've changed
47 "byte" to "char", and slightly reworded the output. The documentation about
48 these values has also been (I hope) clarified.
49
50 10. Another JIT related optimization: use table jumps for selecting the correct
51 backtracking path, when more than four alternatives are present inside a
52 bracket.
53
54 11. Empty match is not possible, when the minimum length is greater than zero,
55 and there is no \K in the pattern. JIT should avoid empty match checks in
56 such cases.
57
58 12. In a caseless character class with UCP support, when a character with more
59 than one alternative case was not the first character of a range, not all
60 the alternative cases were added to the class. For example, s and \x{17f}
61 are both alternative cases for S: the class [RST] was handled correctly,
62 but [R-T] was not.
63
64 13. The configure.ac file always checked for pthread support when JIT was
65 enabled. This is not used in Windows, so I have put this test inside a
66 check for the presence of windows.h (which was already tested for).
67
68 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
69 The algorithm provides a way to skip certain starting offsets, and usually
70 faster than linear prefix searches.
71
72 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
73 as for 'fr_FR' and 'french'. For some reason, however, it then used the
74 Windows-specific input and output files, which have 'french' screwed in.
75 So this could never have worked. One of the problems with locales is that
76 they aren't always the same. I have now updated RunTest so that it checks
77 the output of the locale test (test 3) against three different output
78 files, and it allows the test to pass if any one of them matches. With luck
79 this should make the test pass on some versions of Solaris where it was
80 failing. Because of the uncertainty, the script did not used to stop if
81 test 3 failed; it now does. If further versions of a French locale ever
82 come to light, they can now easily be added.
83
84 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
85 there was a message during ./configure, but it did not stop. This now
86 provokes an error. The invalid example in README has been corrected.
87 If a value less than the minimum is given, the minimum value has always
88 been used, but now a warning is given.
89
90 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
91 was a bug in the test system, which is now fixed. Also, the list of various
92 configurations that are tested for each release did not have one with both
93 16/32 bits and --enable-bar-anycrlf. It now does.
94
95 18. pcretest was missing "-C bsr" for displaying the \R default setting.
96
97
98 Version 8.34 15-December-2013
99 -----------------------------
100
101 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
102 executable memory. Patch inspired by Carsten Klein.
103
104 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
105 this macro is never tested and has no effect, because the work to support
106 coverage involves only compiling and linking options and special targets in
107 the Makefile. The comment in config.h implied that defining the macro would
108 enable coverage support, which is totally false. There was also support for
109 setting this macro in the CMake files (my fault, I just copied it from
110 configure). SUPPORT_GCOV has now been removed.
111
112 3. Make a small performance improvement in strlen16() and strlen32() in
113 pcretest.
114
115 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
116 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
117 these cases). There was also one in pcretest.c.
118
119 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
120
121 6. In UTF mode, the code for checking whether a group could match an empty
122 string (which is used for indefinitely repeated groups to allow for
123 breaking an infinite loop) was broken when the group contained a repeated
124 negated single-character class with a character that occupied more than one
125 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
126 UTF-8 mode). The effect was undefined: the group might or might not be
127 deemed as matching an empty string, or the program might have crashed.
128
129 7. The code for checking whether a group could match an empty string was not
130 recognizing that \h, \H, \v, \V, and \R must match a character.
131
132 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
133 an empty string. If it can, pcretest shows this in its information output.
134
135 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
136 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
137 matched by pcre_exec() without using JIT:
138
139 (a) If the rest of the pattern did not match after a maximal run of
140 grapheme clusters, the code for backing up to try with fewer of them
141 did not always back up over a full grapheme when characters that do not
142 have the modifier quality were involved, e.g. Hangul syllables.
143
144 (b) If the match point in a subject started with modifier character, and
145 there was no match, the code could incorrectly back up beyond the match
146 point, and potentially beyond the first character in the subject,
147 leading to a segfault or an incorrect match result.
148
149 10. A conditional group with an assertion condition could lead to PCRE
150 recording an incorrect first data item for a match if no other first data
151 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
152 first data item, and therefore matched "ca" after "c" instead of at the
153 start.
154
155 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
156 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
157
158 12. The source of pcregrep now includes z/OS-specific code so that it can be
159 compiled for z/OS as part of the special z/OS distribution.
160
161 13. Added the -T and -TM options to pcretest.
162
163 14. The code in pcre_compile.c for creating the table of named capturing groups
164 has been refactored. Instead of creating the table dynamically during the
165 actual compiling pass, the information is remembered during the pre-compile
166 pass (on the stack unless there are more than 20 named groups, in which
167 case malloc() is used) and the whole table is created before the actual
168 compile happens. This has simplified the code (it is now nearly 150 lines
169 shorter) and prepared the way for better handling of references to groups
170 with duplicate names.
171
172 15. A back reference to a named subpattern when there is more than one of the
173 same name now checks them in the order in which they appear in the pattern.
174 The first one that is set is used for the reference. Previously only the
175 first one was inspected. This change makes PCRE more compatible with Perl.
176
177 16. Unicode character properties were updated from Unicode 6.3.0.
178
179 17. The compile-time code for auto-possessification has been refactored, based
180 on a patch by Zoltan Herczeg. It now happens after instead of during
181 compilation. The code is cleaner, and more cases are handled. The option
182 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
183 options in pcretest are provided to set it. It can also be set by
184 (*NO_AUTO_POSSESS) at the start of a pattern.
185
186 18. The character VT has been added to the default ("C" locale) set of
187 characters that match \s and are generally treated as white space,
188 following this same change in Perl 5.18. There is now no difference between
189 "Perl space" and "POSIX space". Whether VT is treated as white space in
190 other locales depends on the locale.
191
192 19. The code for checking named groups as conditions, either for being set or
193 for being recursed, has been refactored (this is related to 14 and 15
194 above). Processing unduplicated named groups should now be as fast at
195 numerical groups, and processing duplicated groups should be faster than
196 before.
197
198 20. Two patches to the CMake build system, by Alexander Barkov:
199
200 (1) Replace the "source" command by "." in CMakeLists.txt because
201 "source" is a bash-ism.
202
203 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
204 without these the CMake build does not work on Solaris.
205
206 21. Perl has changed its handling of \8 and \9. If there is no previously
207 encountered capturing group of those numbers, they are treated as the
208 literal characters 8 and 9 instead of a binary zero followed by the
209 literals. PCRE now does the same.
210
211 22. Following Perl, added \o{} to specify codepoints in octal, making it
212 possible to specify values greater than 0777 and also making them
213 unambiguous.
214
215 23. Perl now gives an error for missing closing braces after \x{... instead of
216 treating the string as literal. PCRE now does the same.
217
218 24. RunTest used to grumble if an inappropriate test was selected explicitly,
219 but just skip it when running all tests. This make it awkward to run ranges
220 of tests when one of them was inappropriate. Now it just skips any
221 inappropriate tests, as it always did when running all tests.
222
223 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
224 character types such as \d or \w, too many callouts were inserted, and the
225 data that they returned was rubbish.
226
227 26. In UCP mode, \s was not matching two of the characters that Perl matches,
228 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
229 were matched by \h. The code has now been refactored so that the lists of
230 the horizontal and vertical whitespace characters used for \h and \v (which
231 are defined only in one place) are now also used for \s.
232
233 27. Add JIT support for the 64 bit TileGX architecture.
234 Patch by Jiong Wang (Tilera Corporation).
235
236 28. Possessive quantifiers for classes (both explicit and automatically
237 generated) now use special opcodes instead of wrapping in ONCE brackets.
238
239 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
240 (because it's meaningless), this was not happening when PCRE_CASELESS was
241 set. Not wrong, but inefficient.
242
243 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
244 properties for \w, \d, etc) is present in a test regex. Otherwise if the
245 test contains no characters greater than 255, Perl doesn't realise it
246 should be using Unicode semantics.
247
248 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
249 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
250 does in Unicode mode.
251
252 32. Added the "forbid" facility to pcretest so that putting tests into the
253 wrong test files can sometimes be quickly detected.
254
255 33. There is now a limit (default 250) on the depth of nesting of parentheses.
256 This limit is imposed to control the amount of system stack used at compile
257 time. It can be changed at build time by --with-parens-nest-limit=xxx or
258 the equivalent in CMake.
259
260 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
261 errors. Perl warns for these when in warning mode, but PCRE has no facility
262 for giving warnings.
263
264 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
265 However, this was not working for (?!) because it is optimized to (*FAIL),
266 for which PCRE does not allow quantifiers. The optimization is now disabled
267 when a quantifier follows (?!). I can't see any use for this, but it makes
268 things uniform.
269
270 36. Perl no longer allows group names to start with digits, so I have made this
271 change also in PCRE. It simplifies the code a bit.
272
273 37. In extended mode, Perl ignores spaces before a + that indicates a
274 possessive quantifier. PCRE allowed a space before the quantifier, but not
275 before the possessive +. It now does.
276
277 38. The use of \K (reset reported match start) within a repeated possessive
278 group such as (a\Kb)*+ was not working.
279
280 40. Document that the same character tables must be used at compile time and
281 run time, and that the facility to pass tables to pcre_exec() and
282 pcre_dfa_exec() is for use only with saved/restored patterns.
283
284 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
285 features for Builds with MSVC:
286
287 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
288 builds). The .rc files can be used to set FileDescription and many other
289 attributes.
290
291 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
292 This allows higher-level build scripts which want .pdb files to avoid
293 hard-coding the exact files needed."
294
295 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
296 mean "start of word" and "end of word", respectively, as a transition aid.
297
298 43. A minimizing repeat of a class containing codepoints greater than 255 in
299 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
300 compiled to use the heap for recursion.
301
302 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
303 is configured.
304
305
306 Version 8.33 28-May-2013
307 ------------------------
308
309 1. Added 'U' to some constants that are compared to unsigned integers, to
310 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
311 variables that are added to signed variables, to ensure the result is
312 signed and can be negated.
313
314 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
315 CMake config files.
316
317 3. Revise the creation of config.h.generic so that all boolean macros are
318 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
319 overriding via -D on the command line possible.
320
321 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
322 to unsigned int is reported to make a quite noticeable speed difference in
323 a specific Windows environment. Testing on Linux did also appear to show
324 some benefit (and it is clearly not harmful). Also fixed the definition of
325 Xop which should be unsigned.
326
327 5. Related to (4), changing the definition of the intermediate variable cc
328 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
329 improvements.
330
331 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
332 unnecessary spaces.
333
334 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
335 and later.
336
337 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
338
339 9. Optimizing fast_forward_start_bits in JIT.
340
341 10. Adding support for callouts in JIT, and fixing some issues revealed
342 during this work. Namely:
343
344 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
345
346 (b) Minimum length was not checked before the matching is started.
347
348 11. The value of capture_last that is passed to callouts was incorrect in some
349 cases when there was a capture on one path that was subsequently abandoned
350 after a backtrack. Also, the capture_last value is now reset after a
351 recursion, since all captures are also reset in this case.
352
353 12. The interpreter no longer returns the "too many substrings" error in the
354 case when an overflowing capture is in a branch that is subsequently
355 abandoned after a backtrack.
356
357 13. In the pathological case when an offset vector of size 2 is used, pcretest
358 now prints out the matched string after a yield of 0 or 1.
359
360 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
361 Only supported by the JIT compiler at the moment.
362
363 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
364
365 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
366 offset of the starting point of the matching process, provided the offsets
367 vector is large enough.
368
369 17. The \A escape now records a lookbehind value of 1, though its execution
370 does not actually inspect the previous character. This is to ensure that,
371 in partial multi-segment matching, at least one character from the old
372 segment is retained when a new segment is processed. Otherwise, if there
373 are no lookbehinds in the pattern, \A might match incorrectly at the start
374 of a new segment.
375
376 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
377
378 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
379 gives some modest performance improvement in 8-bit mode.
380
381 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
382 be expressed in certain programming languages using Universal Character
383 Names.
384
385 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
386 which points out that "non characters" are not "characters that may not
387 appear in Unicode strings" but rather "characters that are reserved for
388 internal use and have only local meaning".
389
390 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
391 there was a conditional group that depended on an assertion, if the
392 assertion was false, the callout that immediately followed the alternation
393 in the condition was skipped when pcre_exec() was used for matching.
394
395 23. Allow an explicit callout to be inserted before an assertion that is the
396 condition for a conditional group, for compatibility with automatic
397 callouts, which always insert a callout at this point.
398
399 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
400 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
401
402 25. (*PRUNE) is now supported by the JIT compiler.
403
404 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
405
406 27. Fix the case where there are two or more SKIPs with arguments that may be
407 ignored.
408
409 28. (*SKIP) is now supported by the JIT compiler.
410
411 29. (*THEN) is now supported by the JIT compiler.
412
413 30. Update RunTest with additional test selector options.
414
415 31. The way PCRE handles backtracking verbs has been changed in two ways.
416
417 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
418 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
419 backtracking. In some cases this makes it more Perl-compatible, but Perl's
420 rather obscure rules do not always do the same thing.
421
422 (2) Previously, backtracking verbs were confined within assertions. This is
423 no longer the case for positive assertions, except for (*ACCEPT). Again,
424 this sometimes improves Perl compatibility, and sometimes does not.
425
426 32. A number of tests that were in test 2 because Perl did things differently
427 have been moved to test 1, because either Perl or PCRE has changed, and
428 these tests are now compatible.
429
430 32. Backtracking control verbs are now handled in the same way in JIT and
431 interpreter.
432
433 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
434 contained a forward subroutine reference caused a compile error.
435
436 34. Auto-detect and optimize limited repetitions in JIT.
437
438 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
439 blocking (*UTF) etc.
440
441 36. In the interpreter, maximizing pattern repetitions for characters and
442 character types now use tail recursion, which reduces stack usage.
443
444 37. The value of the max lookbehind was not correctly preserved if a compiled
445 and saved regex was reloaded on a host of different endianness.
446
447 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
448 of the compiled pattern block, expand the flags field from 16 to 32 bits
449 because it was almost full.
450
451 39. Try madvise first before posix_madvise.
452
453 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
454 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
455 original change locked out all matches of empty strings. This has been
456 changed so that one match of an empty string per line is recognized.
457 Subsequent searches on the same line (for colouring or for --only-matching,
458 for example) do not recognize empty strings.
459
460 41. Applied a user patch to fix a number of spelling mistakes in comments.
461
462 42. Data lines longer than 65536 caused pcretest to crash.
463
464 43. Clarified the data type for length and startoffset arguments for pcre_exec
465 and pcre_dfa_exec in the function-specific man pages, where they were
466 explicitly stated to be in bytes, never having been updated. I also added
467 some clarification to the pcreapi man page.
468
469 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
470 a segmentation fault.
471
472
473 Version 8.32 30-November-2012
474 -----------------------------
475
476 1. Improved JIT compiler optimizations for first character search and single
477 character iterators.
478
479 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
480 Patch by Daniel Richard G.
481
482 3. Single character iterator optimizations in the JIT compiler.
483
484 4. Improved JIT compiler optimizations for character ranges.
485
486 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
487 Reported by Giuseppe D'Angelo.
488
489 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
490 of a line, was being set incorrectly in cases where .* appeared inside
491 atomic brackets at the start of a pattern, or where there was a subsequent
492 *PRUNE or *SKIP.
493
494 7. Improved instruction cache flush for POWER/PowerPC.
495 Patch by Daniel Richard G.
496
497 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
498 grep:
499
500 (a) There is now no limit to the number of patterns to be matched.
501
502 (b) An error is given if a pattern is too long.
503
504 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
505 are now supported.
506
507 (d) --exclude-from and --include-from (multiple use) have been added.
508
509 (e) Exclusions and inclusions now apply to all files and directories, not
510 just to those obtained from scanning a directory recursively.
511
512 (f) Multiple uses of -f and --file-list are now supported.
513
514 (g) In a Windows environment, the default for -d has been changed from
515 "read" (the GNU grep default) to "skip", because otherwise the presence
516 of a directory in the file list provokes an error.
517
518 (h) The documentation has been revised and clarified in places.
519
520 9. Improve the matching speed of capturing brackets.
521
522 10. Changed the meaning of \X so that it now matches a Unicode extended
523 grapheme cluster.
524
525 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
526 out POSIX threads when JIT support is configured.
527
528 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
529
530 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
531 structure could go wrong in environments where size_t is not the same size
532 as int.
533
534 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
535 set.
536
537 15. The EBCDIC support had decayed; later updates to the code had included
538 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
539 general tidy up of EBCDIC-related issues, and the documentation was also
540 not quite right. There is now a test that can be run on ASCII systems to
541 check some of the EBCDIC-related things (but is it not a full test).
542
543 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
544 in a small tidy to the code.
545
546 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
547
548 18. If the --only-matching (-o) option in pcregrep is specified multiple
549 times, each one causes appropriate output. For example, -o1 -o2 outputs the
550 substrings matched by the 1st and 2nd capturing parentheses. A separating
551 string can be specified by --om-separator (default empty).
552
553 19. Improving the first n character searches.
554
555 20. Turn case lists for horizontal and vertical white space into macros so that
556 they are defined only once.
557
558 21. This set of changes together give more compatible Unicode case-folding
559 behaviour for characters that have more than one other case when UCP
560 support is available.
561
562 (a) The Unicode property table now has offsets into a new table of sets of
563 three or more characters that are case-equivalent. The MultiStage2.py
564 script that generates these tables (the pcre_ucd.c file) now scans
565 CaseFolding.txt instead of UnicodeData.txt for character case
566 information.
567
568 (b) The code for adding characters or ranges of characters to a character
569 class has been abstracted into a generalized function that also handles
570 case-independence. In UTF-mode with UCP support, this uses the new data
571 to handle characters with more than one other case.
572
573 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
574 whose other case is greater than 256 are now correctly matched
575 caselessly. Previously, the high codepoint matched the low one, but not
576 vice versa.
577
578 (d) The processing of \h, \H, \v, and \ in character classes now makes use
579 of the new class addition function, using character lists defined as
580 macros alongside the case definitions of 20 above.
581
582 (e) Caseless back references now work with characters that have more than
583 one other case.
584
585 (f) General caseless matching of characters with more than one other case
586 is supported.
587
588 22. Unicode character properties were updated from Unicode 6.2.0
589
590 23. Improved CMake support under Windows. Patch by Daniel Richard G.
591
592 24. Add support for 32-bit character strings, and UTF-32
593
594 25. Major JIT compiler update (code refactoring and bugfixing).
595 Experimental Sparc 32 support is added.
596
597 26. Applied a modified version of Daniel Richard G's patch to create
598 pcre.h.generic and config.h.generic by "make" instead of in the
599 PrepareRelease script.
600
601 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
602 pcre_compile.c when checking for a zero character.
603
604 28. Introducing a native interface for JIT. Through this interface, the compiled
605 machine code can be directly executed. The purpose of this interface is to
606 provide fast pattern matching, so several sanity checks are not performed.
607 However, feature tests are still performed. The new interface provides
608 1.4x speedup compared to the old one.
609
610 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
611 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
612 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
613
614 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
615 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
616 the "old" RFC 2279). Instead, it ended up passing a negative length to
617 pcre_exec().
618
619 31. Add support for GCC's visibility feature to hide internal functions.
620
621 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
622 "unknown -C option" after outputting 0 or 1.
623
624 33. There is now support for generating a code coverage report for the test
625 suite in environments where gcc is the compiler and lcov is installed. This
626 is mainly for the benefit of the developers.
627
628 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
629 unaddressable using valgrind annotations, allowing valgrind to detect
630 invalid memory accesses. This is mainly for the benefit of the developers.
631
632 25. (*UTF) can now be used to start a pattern in any of the three libraries.
633
634 26. Give configure error if --enable-cpp but no C++ compiler found.
635
636
637 Version 8.31 06-July-2012
638 -------------------------
639
640 1. Fixing a wrong JIT test case and some compiler warnings.
641
642 2. Removed a bashism from the RunTest script.
643
644 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
645 to unsigned type, result still unsigned" that was given by an MS compiler
646 on encountering the code "-sizeof(xxx)".
647
648 4. Partial matching support is added to the JIT compiler.
649
650 5. Fixed several bugs concerned with partial matching of items that consist
651 of more than one character:
652
653 (a) /^(..)\1/ did not partially match "aba" because checking references was
654 done on an "all or nothing" basis. This also applied to repeated
655 references.
656
657 (b) \R did not give a hard partial match if \r was found at the end of the
658 subject.
659
660 (c) \X did not give a hard partial match after matching one or more
661 characters at the end of the subject.
662
663 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
664 a partial match for the string "\r".
665
666 (e) When newline was set to CRLF, the metacharacter "." did not recognize
667 a partial match for a CR character at the end of the subject string.
668
669 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
670 running pcretest, the text "(JIT)" added to the output whenever JIT is
671 actually used to run the match.
672
673 7. Individual JIT compile options can be set in pcretest by following -s+[+]
674 or /S+[+] with a digit between 1 and 7.
675
676 8. OP_NOT now supports any UTF character not just single-byte ones.
677
678 9. (*MARK) control verb is now supported by the JIT compiler.
679
680 10. The command "./RunTest list" lists the available tests without actually
681 running any of them. (Because I keep forgetting what they all are.)
682
683 11. Add PCRE_INFO_MAXLOOKBEHIND.
684
685 12. Applied a (slightly modified) user-supplied patch that improves performance
686 when the heap is used for recursion (compiled with --disable-stack-for-
687 recursion). Instead of malloc and free for each heap frame each time a
688 logical recursion happens, frames are retained on a chain and re-used where
689 possible. This sometimes gives as much as 30% improvement.
690
691 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
692 call.
693
694 14. As documented, (*COMMIT) is now confined to within a positive assertion.
695
696 15. It is now possible to link pcretest with libedit as an alternative to
697 libreadline.
698
699 16. (*COMMIT) control verb is now supported by the JIT compiler.
700
701 17. The Unicode data tables have been updated to Unicode 6.1.0.
702
703 18. Added --file-list option to pcregrep.
704
705 19. Added binary file support to pcregrep, including the -a, --binary-files,
706 -I, and --text options.
707
708 20. The madvise function is renamed for posix_madvise for QNX compatibility
709 reasons. Fixed by Giuseppe D'Angelo.
710
711 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
712
713 22. Changed the option for creating symbolic links for 16-bit man pages from
714 -s to -sf so that re-installing does not cause issues.
715
716 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
717
718 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
719 matches in certain environments (the workspace was not being correctly
720 retained). Also added to pcre_dfa_exec() a simple plausibility check on
721 some of the workspace data at the beginning of a restart.
722
723 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
724 was not doing so when it should - probably a typo introduced by SVN 528
725 (change 8.10/14).
726
727 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
728 \w+ when the character tables indicated that \x{c4} was a word character.
729 There were several related cases, all because the tests for doing a table
730 lookup were testing for characters less than 127 instead of 255.
731
732 27. If a pattern contains capturing parentheses that are not used in a match,
733 their slots in the ovector are set to -1. For those that are higher than
734 any matched groups, this happens at the end of processing. In the case when
735 there were back references that the ovector was too small to contain
736 (causing temporary malloc'd memory to be used during matching), and the
737 highest capturing number was not used, memory off the end of the ovector
738 was incorrectly being set to -1. (It was using the size of the temporary
739 memory instead of the true size.)
740
741 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
742 ovector size, it uses memory at the end of the block that it has got.
743
744 29. Check for an overlong MARK name and give an error at compile time. The
745 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
746
747 30. JIT compiler update.
748
749 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
750 Rill for the patch.
751
752 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
753
754 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
755
756 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
757 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
758 was enabled.
759
760 35. Improve JIT code generation for greedy plus quantifier.
761
762 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
763 1 to "aa" instead of to an empty string. The bug affected repeated groups
764 that could potentially match an empty string.
765
766 37. Optimizing single character iterators in JIT.
767
768 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
769 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
770 codepoints that are too big for the mode are faulted, and in a UTF mode,
771 disallowed codepoints are also faulted.
772
773 39. If PCRE was compiled with UTF support, in three places in the DFA
774 matcher there was code that should only have been obeyed in UTF mode, but
775 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
776 processing when bytes with values greater than 127 were present. In 16-bit
777 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
778 both cases the values are those that cannot be the first data item in a UTF
779 character. The three items that might have provoked this were recursions,
780 possessively repeated groups, and atomic groups.
781
782 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
783 and pcregrep, because some OS require shared objects to be explicitly
784 passed to ld, causing the link step to fail if they are not.
785
786 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
787 mode, patterns that started with \h* or \R* might be incorrectly matched.
788
789
790 Version 8.30 04-February-2012
791 -----------------------------
792
793 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
794 name is defined in ctype.h.
795
796 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
797 only in quite long subpatterns.
798
799 3. Removed the function pcre_info(), which has been obsolete and deprecated
800 since it was replaced by pcre_fullinfo() in February 2000.
801
802 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
803 match a (*MARK), and the match failed at the start of the subject, a
804 reference to memory before the start of the subject could occur. This bug
805 was introduced by fix 17 of release 8.21.
806
807 5. A reference to an unset group with zero minimum repetition was giving
808 totally wrong answers (in non-JavaScript-compatibility mode). For example,
809 /(another)?(\1?)test/ matched against "hello world test". This bug was
810 introduced in release 8.13.
811
812 6. Add support for 16-bit character strings (a large amount of work involving
813 many changes and refactorings).
814
815 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
816 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
817 from a file.
818
819 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
820 rounding is not applied in this particular case).
821
822 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
823 if they appear, or are escaped, in patterns.
824
825 10. Get rid of a number of -Wunused-but-set-variable warnings.
826
827 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
828 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
829 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
830 also returns the mark "x". This bug applied to capturing parentheses,
831 non-capturing parentheses, and atomic parentheses. It also applied to some
832 assertions.
833
834 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
835 information out of configure.ac instead of relying on pcre.h.generic, which
836 is not stored in the repository.
837
838 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
839 -lreadline.
840
841 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
842
843 15. Applied Graycode's patch to put the top-level frame on the stack rather
844 than the heap when not using the stack for recursion. This gives a
845 performance improvement in many cases when recursion is not deep.
846
847 16. Experimental code added to "pcretest -C" to output the stack frame size.
848
849
850 Version 8.21 12-Dec-2011
851 ------------------------
852
853 1. Updating the JIT compiler.
854
855 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
856 are added as well.
857
858 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
859 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
860 calling _pcre_jit_exec. Some extra comments are added.
861
862 4. (*MARK) settings inside atomic groups that do not contain any capturing
863 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
864 was introduced by change 18 for 8.20.
865
866 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
867 ECMA-262 standard.
868
869 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
870 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
871 This bug was probably introduced by change 9 of 8.13.
872
873 7. While fixing 6 above, I noticed that a number of other items were being
874 incorrectly rejected as "not fixed length". This arose partly because newer
875 opcodes had not been added to the fixed-length checking code. I have (a)
876 corrected the bug and added tests for these items, and (b) arranged for an
877 error to occur if an unknown opcode is encountered while checking for fixed
878 length instead of just assuming "not fixed length". The items that were
879 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
880 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
881 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
882
883 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
884 being incorrectly compiled and would have given unpredicatble results.
885
886 9. A possessively repeated subpattern with minimum repeat count greater than
887 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
888 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
889 the first (A) could occur when it should not.
890
891 10. Add a cast and remove a redundant test from the code.
892
893 11. JIT should use pcre_malloc/pcre_free for allocation.
894
895 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
896 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
897 is anything other than /usr, -L is still shown).
898
899 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
900
901 14. Perl does not support \N without a following name in a [] class; PCRE now
902 also gives an error.
903
904 15. If a forward reference was repeated with an upper limit of around 2000,
905 it caused the error "internal error: overran compiling workspace". The
906 maximum number of forward references (including repeats) was limited by the
907 internal workspace, and dependent on the LINK_SIZE. The code has been
908 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
909 the default depends on LINK_SIZE. There is a new upper limit (for safety)
910 of around 200,000 forward references. While doing this, I also speeded up
911 the filling in of repeated forward references.
912
913 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
914 incorrectly expecting the subject to contain another "a" after the start.
915
916 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
917 in the match, the SKIP should be ignored. This was not happening; instead
918 the SKIP was being treated as NOMATCH. For patterns such as
919 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
920 tested.
921
922 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
923 now much more compatible with Perl, in particular in cases where the result
924 is a non-match for a non-anchored pattern. For example, if
925 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
926 "m", where previously it did not return a name. A side effect of this
927 change is that for partial matches, the last encountered mark name is
928 returned, as for non matches. A number of tests that were previously not
929 Perl-compatible have been moved into the Perl-compatible test files. The
930 refactoring has had the pleasing side effect of removing one argument from
931 the match() function, thus reducing its stack requirements.
932
933 19. If the /S+ option was used in pcretest to study a pattern using JIT,
934 subsequent uses of /S (without +) incorrectly behaved like /S+.
935
936 21. Retrieve executable code size support for the JIT compiler and fixing
937 some warnings.
938
939 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
940 not work when the shorter character appeared right at the end of the
941 subject string.
942
943 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
944 systems.
945
946 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
947 output it when the /M option is used in pcretest.
948
949 25. The CheckMan script was not being included in the distribution. Also, added
950 an explicit "perl" to run Perl scripts from the PrepareRelease script
951 because this is reportedly needed in Windows.
952
953 26. If study data was being save in a file and studying had not found a set of
954 "starts with" bytes for the pattern, the data written to the file (though
955 never used) was taken from uninitialized memory and so caused valgrind to
956 complain.
957
958 27. Updated RunTest.bat as provided by Sheri Pierce.
959
960 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
961
962 29. Computation of memory usage for the table of capturing group names was
963 giving an unnecessarily large value.
964
965
966 Version 8.20 21-Oct-2011
967 ------------------------
968
969 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
970 a POSIX class. After further experiments with Perl, which convinced me that
971 Perl has bugs and confusions, a closing square bracket is no longer allowed
972 in a POSIX name. This bug also affected patterns with classes that started
973 with full stops.
974
975 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
976 captured substring, but while checking the failing first alternative,
977 substring 1 is temporarily captured. If the output vector supplied to
978 pcre_exec() was not big enough for this capture, the yield of the function
979 was still zero ("insufficient space for captured substrings"). This cannot
980 be totally fixed without adding another stack variable, which seems a lot
981 of expense for a edge case. However, I have improved the situation in cases
982 such as /(a)(b)x|abc/ matched against "abc", where the return code
983 indicates that fewer than the maximum number of slots in the ovector have
984 been set.
985
986 3. Related to (2) above: when there are more back references in a pattern than
987 slots in the output vector, pcre_exec() uses temporary memory during
988 matching, and copies in the captures as far as possible afterwards. It was
989 using the entire output vector, but this conflicts with the specification
990 that only 2/3 is used for passing back captured substrings. Now it uses
991 only the first 2/3, for compatibility. This is, of course, another edge
992 case.
993
994 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
995 main code base, and can be used by building with --enable-jit. When this is
996 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
997 runtime --no-jit option is given.
998
999 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1000 ovector, the return from the function was zero, implying that there were
1001 other matches that did not fit. The correct "exactly full" value is now
1002 returned.
1003
1004 6. If a subpattern that was called recursively or as a subroutine contained
1005 (*PRUNE) or any other control that caused it to give a non-standard return,
1006 invalid errors such as "Error -26 (nested recursion at the same subject
1007 position)" or even infinite loops could occur.
1008
1009 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1010 computing the minimum length on reaching *ACCEPT, and so ended up with the
1011 wrong value of 1 rather than 0. Further investigation indicates that
1012 computing a minimum subject length in the presence of *ACCEPT is difficult
1013 (think back references, subroutine calls), and so I have changed the code
1014 so that no minimum is registered for a pattern that contains *ACCEPT.
1015
1016 8. If (*THEN) was present in the first (true) branch of a conditional group,
1017 it was not handled as intended. [But see 16 below.]
1018
1019 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1020 Sheri Pierce.
1021
1022 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1023 the first byte in a match must be "a".
1024
1025 11. Change 17 for 8.13 increased the recursion depth for patterns like
1026 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1027 pattern contains any instances of (*THEN). If it does not, the old
1028 optimizations are restored. It would be nice to do this on a per-group
1029 basis, but at the moment that is not feasible.
1030
1031 12. In some environments, the output of pcretest -C is CRLF terminated. This
1032 broke RunTest's code that checks for the link size. A single white space
1033 character after the value is now allowed for.
1034
1035 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1036 For "fr", it uses the Windows-specific input and output files.
1037
1038 14. If (*THEN) appeared in a group that was called recursively or as a
1039 subroutine, it did not work as intended. [But see next item.]
1040
1041 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1042 pattern fragments (but not containing any | characters). If A and B are
1043 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1044 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1045 D. In other words, Perl considers parentheses that do not contain any |
1046 characters to be part of a surrounding alternative, whereas PCRE was
1047 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1048 differently. PCRE now behaves in the same way as Perl, except in the case
1049 of subroutine/recursion calls such as (?1) which have in any case always
1050 been different (but PCRE had them first :-).
1051
1052 16. Related to 15 above: Perl does not treat the | in a conditional group as
1053 creating alternatives. Such a group is treated in the same way as an
1054 ordinary group without any | characters when processing (*THEN). PCRE has
1055 been changed to match Perl's behaviour.
1056
1057 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1058 RunGrepTest script failed.
1059
1060 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1061 inevitable for groups that contain captures, but it can lead to a lot of
1062 stack use in large patterns. The old behaviour has been restored for atomic
1063 groups that do not contain any capturing parentheses.
1064
1065 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1066 suppress the check for a minimum subject length at run time. (If it was
1067 given to pcre_exec() or pcre_dfa_exec() it did work.)
1068
1069 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1070 fail to work when decoding hex characters in data strings in EBCDIC
1071 environments.
1072
1073 21. It appears that in at least one Mac OS environment, the isxdigit() function
1074 is implemented as a macro that evaluates to its argument more than once,
1075 contravening the C 90 Standard (I haven't checked a later standard). There
1076 was an instance in pcretest which caused it to go wrong when processing
1077 \x{...} escapes in subject strings. The has been rewritten to avoid using
1078 things like p++ in the argument of isxdigit().
1079
1080
1081 Version 8.13 16-Aug-2011
1082 ------------------------
1083
1084 1. The Unicode data tables have been updated to Unicode 6.0.0.
1085
1086 2. Two minor typos in pcre_internal.h have been fixed.
1087
1088 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1089 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1090 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1091
1092 4. There were a number of related bugs in the code for matching backrefences
1093 caselessly in UTF-8 mode when codes for the characters concerned were
1094 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1095 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1096 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1097 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1098 2-byte code at the end of the subject (it thought there wasn't enough data
1099 left).
1100
1101 5. Comprehensive information about what went wrong is now returned by
1102 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1103 as the output vector has at least 2 elements. The offset of the start of
1104 the failing character and a reason code are placed in the vector.
1105
1106 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1107 now returned is for the first byte of the failing character, instead of the
1108 last byte inspected. This is an incompatible change, but I hope it is small
1109 enough not to be a problem. It makes the returned offset consistent with
1110 pcre_exec() and pcre_dfa_exec().
1111
1112 7. pcretest now gives a text phrase as well as the error number when
1113 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1114 failure, the offset and reason code are output.
1115
1116 8. When \R was used with a maximizing quantifier it failed to skip backwards
1117 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1118 back over a single character (\n). This seems wrong (because it treated the
1119 two characters as a single entity when going forwards), conflicts with the
1120 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1121 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1122 has been changed.
1123
1124 9. Some internal refactoring has changed the processing so that the handling
1125 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1126 time (the PCRE_DOTALL option was changed this way some time ago: version
1127 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1128 which was always a bit of a fudge. It also means that there is one less
1129 argument for the match() function, which reduces its stack requirements
1130 slightly. This change also fixes an incompatibility with Perl: the pattern
1131 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1132
1133 10. More internal refactoring has drastically reduced the number of recursive
1134 calls to match() for possessively repeated groups such as (abc)++ when
1135 using pcre_exec().
1136
1137 11. While implementing 10, a number of bugs in the handling of groups were
1138 discovered and fixed:
1139
1140 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1141 (a|)*(?1) gave a compile-time internal error.
1142 ((a|)+)+ did not notice that the outer group could match an empty string.
1143 (^a|^)+ was not marked as anchored.
1144 (.*a|.*)+ was not marked as matching at start or after a newline.
1145
1146 12. Yet more internal refactoring has removed another argument from the match()
1147 function. Special calls to this function are now indicated by setting a
1148 value in a variable in the "match data" data block.
1149
1150 13. Be more explicit in pcre_study() instead of relying on "default" for
1151 opcodes that mean there is no starting character; this means that when new
1152 ones are added and accidentally left out of pcre_study(), testing should
1153 pick them up.
1154
1155 14. The -s option of pcretest has been documented for ages as being an old
1156 synonym of -m (show memory usage). I have changed it to mean "force study
1157 for every regex", that is, assume /S for every regex. This is similar to -i
1158 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1159 using it. It makes it easier to run collections of tests with and without
1160 study enabled, and thereby test pcre_study() more easily. All the standard
1161 tests are now run with and without -s (but some patterns can be marked as
1162 "never study" - see 20 below).
1163
1164 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1165 restoration of the capturing data to the outer values was not happening
1166 correctly.
1167
1168 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1169 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1170 pattern had matched an empty string, and so incorrectly returned a no
1171 match.
1172
1173 17. There was optimizing code for the last branch of non-capturing parentheses,
1174 and also for the obeyed branch of a conditional subexpression, which used
1175 tail recursion to cut down on stack usage. Unfortunately, now that there is
1176 the possibility of (*THEN) occurring in these branches, tail recursion is
1177 no longer possible because the return has to be checked for (*THEN). These
1178 two optimizations have therefore been removed. [But see 8.20/11 above.]
1179
1180 18. If a pattern containing \R was studied, it was assumed that \R always
1181 matched two bytes, thus causing the minimum subject length to be
1182 incorrectly computed because \R can also match just one byte.
1183
1184 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1185 was incorrectly computed.
1186
1187 20. If /S is present twice on a test pattern in pcretest input, it now
1188 *disables* studying, thereby overriding the use of -s on the command line
1189 (see 14 above). This is necessary for one or two tests to keep the output
1190 identical in both cases.
1191
1192 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1193 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1194
1195 22. When an atomic group that contained a capturing parenthesis was
1196 successfully matched, but the branch in which it appeared failed, the
1197 capturing was not being forgotten if a higher numbered group was later
1198 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1199 group 1 to "a", when in fact it should be unset. This applied to multi-
1200 branched capturing and non-capturing groups, repeated or not, and also to
1201 positive assertions (capturing in negative assertions does not happen
1202 in PCRE) and also to nested atomic groups.
1203
1204 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1205 subject after a captured substring, to make it easier to tell which of a
1206 number of identical substrings has been captured.
1207
1208 24. The way atomic groups are processed by pcre_exec() has been changed so that
1209 if they are repeated, backtracking one repetition now resets captured
1210 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1211 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1212 "aaa". Previously, it would have been "a". As part of this code
1213 refactoring, the way recursive calls are handled has also been changed.
1214
1215 25. If an assertion condition captured any substrings, they were not passed
1216 back unless some other capturing happened later. For example, if
1217 (?(?=(a))a) was matched against "a", no capturing was returned.
1218
1219 26. When studying a pattern that contained subroutine calls or assertions,
1220 the code for finding the minimum length of a possible match was handling
1221 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1222 group 1 called group 2 while simultaneously a separate group 2 called group
1223 1). A stack overflow occurred in this case. I have fixed this by limiting
1224 the recursion depth to 10.
1225
1226 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1227 Fortmann. This supports explicit test numbers on the command line, and has
1228 argument validation and error reporting.
1229
1230 28. An instance of \X with an unlimited repeat could fail if at any point the
1231 first character it looked at was a mark character.
1232
1233 29. Some minor code refactoring concerning Unicode properties and scripts
1234 should reduce the stack requirement of match() slightly.
1235
1236 30. Added the '=' option to pcretest to check the setting of unused capturing
1237 slots at the end of the pattern, which are documented as being -1, but are
1238 not included in the return count.
1239
1240 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1241 compiled something random. Now it gives a compile-time error (as does
1242 Perl).
1243
1244 32. A *MARK encountered during the processing of a positive assertion is now
1245 recorded and passed back (compatible with Perl).
1246
1247 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1248 had alternative anchored branches, the search for a second match in a line
1249 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1250 matched the line "0102" twice. The same bug affected patterns that started
1251 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1252 twice.
1253
1254 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1255 does, and because of capturing effects, quantifying parenthesized
1256 assertions may at times be useful. Quantifiers are now allowed for
1257 parenthesized assertions.
1258
1259 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1260
1261 36. \g was being checked for fancy things in a character class, when it should
1262 just be a literal "g".
1263
1264 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1265 appearance of a nested POSIX class supersedes an apparent external class.
1266 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1267 unescaped square brackets may also appear as part of class names. For
1268 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1269 more like Perl. (But see 8.20/1 above.)
1270
1271 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1272 was because it thought it was \N{name}, which is not supported).
1273
1274 39. Add minix to OS list not supporting the -S option in pcretest.
1275
1276 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1277 cannot analyze patterns in sufficient detail to catch mutual recursions
1278 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1279 subgroup is called recursively as a subpattern for a second time at the
1280 same position in the subject string. In previous releases this might have
1281 been caught by the recursion limit, or it might have run out of stack.
1282
1283 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1284 happen only once. PCRE was, however incorrectly giving a compile time error
1285 "recursive call could loop indefinitely" because it cannot analyze the
1286 pattern in sufficient detail. The compile time test no longer happens when
1287 PCRE is compiling a conditional subpattern, but actual runaway loops are
1288 now caught at runtime (see 40 above).
1289
1290 42. It seems that Perl allows any characters other than a closing parenthesis
1291 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1292 has been changed to be the same.
1293
1294 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1295 as not to get warnings when autogen.sh is called. Also changed
1296 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1297
1298 44. To help people who use pcregrep to scan files containing exceedingly long
1299 lines, the following changes have been made:
1300
1301 (a) The default value of the buffer size parameter has been increased from
1302 8K to 20K. (The actual buffer used is three times this size.)
1303
1304 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1305 PCRE is built.
1306
1307 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1308 to be set at run time.
1309
1310 (d) Numerical values in pcregrep options can be followed by K or M, for
1311 example --buffer-size=50K.
1312
1313 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1314 given and the return code is set to 2.
1315
1316 45. Add a pointer to the latest mark to the callout data block.
1317
1318 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1319 partial match of an empty string instead of no match. This was specific to
1320 the use of ".".
1321
1322 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1323 complete match instead of a partial match. This bug was dependent on both
1324 the PCRE_UTF8 and PCRE_DOTALL options being set.
1325
1326 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1327 starting byte set, because \b was not being ignored.
1328
1329
1330 Version 8.12 15-Jan-2011
1331 ------------------------
1332
1333 1. Fixed some typos in the markup of the man pages, and wrote a script that
1334 checks for such things as part of the documentation building process.
1335
1336 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1337 --match-limit and --recursion-limit options (added for 8.11). In
1338 particular, this made one of the standard tests fail. (The integer value
1339 went into the wrong half of a long int.)
1340
1341 3. If the --colour option was given to pcregrep with -v (invert match), it
1342 did strange things, either producing crazy output, or crashing. It should,
1343 of course, ignore a request for colour when reporting lines that do not
1344 match.
1345
1346 4. Another pcregrep bug caused similar problems if --colour was specified with
1347 -M (multiline) and the pattern match finished with a line ending.
1348
1349 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1350 matched in multiline mode, the following line was shown as part of the
1351 match. This seems wrong, so I have changed it.
1352
1353 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1354 the check for further matches in the same line (so they could be coloured)
1355 to overrun the end of the current line. If another match was found, it was
1356 incorrectly shown (and then shown again when found in the next line).
1357
1358 7. If pcregrep was compiled under Windows, there was a reference to the
1359 function pcregrep_exit() before it was defined. I am assuming this was
1360 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1361 reported by a user. I've moved the definition above the reference.
1362
1363
1364 Version 8.11 10-Dec-2010
1365 ------------------------
1366
1367 1. (*THEN) was not working properly if there were untried alternatives prior
1368 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1369 backtracked to try for "b" instead of moving to the next alternative branch
1370 at the same level (in this case, to look for "c"). The Perl documentation
1371 is clear that when (*THEN) is backtracked onto, it goes to the "next
1372 alternative in the innermost enclosing group".
1373
1374 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1375 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1376 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1377 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1378 (*THEN).
1379
1380 3. If \s appeared in a character class, it removed the VT character from
1381 the class, even if it had been included by some previous item, for example
1382 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1383 of \s, but is part of the POSIX "space" class.)
1384
1385 4. A partial match never returns an empty string (because you can always
1386 match an empty string at the end of the subject); however the checking for
1387 an empty string was starting at the "start of match" point. This has been
1388 changed to the "earliest inspected character" point, because the returned
1389 data for a partial match starts at this character. This means that, for
1390 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1391 (previously it gave "no match").
1392
1393 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1394 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1395 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1396 has an implication that the given string is incomplete (because a partial
1397 match is preferred over a full match). For this reason, these items now
1398 give a partial match in this situation. [Aside: previously, the one case
1399 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1400 match rather than a full match, which was wrong by the old rules, but is
1401 now correct.]
1402
1403 6. There was a bug in the handling of #-introduced comments, recognized when
1404 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1405 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1406 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1407 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1408 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1409 places in pcre_compile().
1410
1411 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1412 comments when looking ahead for named forward references to subpatterns,
1413 the only newline sequence it recognized was NL. It now handles newlines
1414 according to the set newline convention.
1415
1416 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1417 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1418 cater for a lack of strerror(). These oversights have been fixed.
1419
1420 9. Added --match-limit and --recursion-limit to pcregrep.
1421
1422 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1423
1424 11. When the -o option was used, pcregrep was setting a return code of 1, even
1425 when matches were found, and --line-buffered was not being honoured.
1426
1427 12. Added an optional parentheses number to the -o and --only-matching options
1428 of pcregrep.
1429
1430 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1431 can match an empty string. The code to do it in pcretest and pcredemo
1432 needed fixing:
1433
1434 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1435 only one byte after an empty string match just before CRLF (this case
1436 just got forgotten; "any" and "anycrlf" were OK).
1437
1438 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1439 mode when an empty string match preceded an ASCII character followed by
1440 a non-ASCII character. (The code for advancing by one character rather
1441 than one byte was nonsense.)
1442
1443 (c) The pcredemo.c sample program did not have any code at all to handle
1444 the cases when CRLF is a valid newline sequence.
1445
1446 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1447 as a starting offset was within the subject string. There is now a new
1448 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1449 negative or greater than the length of the string. In order to test this,
1450 pcretest is extended to allow the setting of negative starting offsets.
1451
1452 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1453 starting offset points to the beginning of a UTF-8 character was
1454 unnecessarily clumsy. I tidied it up.
1455
1456 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1457 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1458
1459 17. Nobody had reported that the --include_dir option, which was added in
1460 release 7.7 should have been called --include-dir (hyphen, not underscore)
1461 for compatibility with GNU grep. I have changed it to --include-dir, but
1462 left --include_dir as an undocumented synonym, and the same for
1463 --exclude-dir, though that is not available in GNU grep, at least as of
1464 release 2.5.4.
1465
1466 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1467 characters from a string of bytes) have been redefined so as not to use
1468 loops, in order to improve performance in some environments. At the same
1469 time, I abstracted some of the common code into auxiliary macros to save
1470 repetition (this should not affect the compiled code).
1471
1472 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1473 compile-time error is now given if \c is not followed by an ASCII
1474 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1475 different, and any byte value is allowed.)
1476
1477 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1478 START_OPTIMIZE option, which is now allowed at compile time - but just
1479 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1480 to pcregrep and other applications that have no direct access to PCRE
1481 options. The new /Y option in pcretest sets this option when calling
1482 pcre_compile().
1483
1484 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1485 back references. Groups containing recursive back references were forced to
1486 be atomic by that change, but in the case of named groups, the amount of
1487 memory required was incorrectly computed, leading to "Failed: internal
1488 error: code overflow". This has been fixed.
1489
1490 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1491 pcretest.c, to avoid build problems in some Borland environments.
1492
1493
1494 Version 8.10 25-Jun-2010
1495 ------------------------
1496
1497 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1498 THEN.
1499
1500 2. (*ACCEPT) was not working when inside an atomic group.
1501
1502 3. Inside a character class, \B is treated as a literal by default, but
1503 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1504 causes the error). The code is unchanged, but I tidied the documentation.
1505
1506 4. Inside a character class, PCRE always treated \R and \X as literals,
1507 whereas Perl faults them if its -w option is set. I have changed PCRE so
1508 that it faults them when PCRE_EXTRA is set.
1509
1510 5. Added support for \N, which always matches any character other than
1511 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1512
1513 6. When compiling pcregrep with newer versions of gcc which may have
1514 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1515 declared with attribute warn_unused_result" were given. Just casting the
1516 result to (void) does not stop the warnings; a more elaborate fudge is
1517 needed. I've used a macro to implement this.
1518
1519 7. Minor change to pcretest.c to avoid a compiler warning.
1520
1521 8. Added four artifical Unicode properties to help with an option to make
1522 \s etc use properties (see next item). The new properties are: Xan
1523 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1524
1525 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1526 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1527 this option. Modified pcretest to add /W to test this facility. Added
1528 REG_UCP to make it available via the POSIX interface.
1529
1530 10. Added --line-buffered to pcregrep.
1531
1532 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1533 studied, and the match started with a letter with a code point greater than
1534 127 whose first byte was different to the first byte of the other case of
1535 the letter, the other case of this starting letter was not recognized
1536 (#976).
1537
1538 12. If a pattern that was studied started with a repeated Unicode property
1539 test, for example, \p{Nd}+, there was the theoretical possibility of
1540 setting up an incorrect bitmap of starting bytes, but fortunately it could
1541 not have actually happened in practice until change 8 above was made (it
1542 added property types that matched character-matching opcodes).
1543
1544 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1545 possible starting bytes for non-anchored patterns.
1546
1547 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1548 \R, and also a number of cases that involve Unicode properties, both
1549 explicit and implicit when PCRE_UCP is set.
1550
1551 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1552 input, it could crash or give wrong results if characters with values
1553 greater than 0xc0 were present in the subject string. (Detail: it assumed
1554 UTF-8 input when processing these items.)
1555
1556 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1557 size_t is 64-bit (#991).
1558
1559 17. Added a check for running out of memory when PCRE is compiled with
1560 --disable-stack-for-recursion (#990).
1561
1562 18. If the last data line in a file for pcretest does not have a newline on
1563 the end, a newline was missing in the output.
1564
1565 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1566 less than 128) in its various bitmaps. However, there is a facility for
1567 generating tables according to the current locale when PCRE is compiled. It
1568 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1569 characters, are recognized by isspace() and therefore were getting set in
1570 these tables, and indeed these tables seem to approximate to ISO 8859. This
1571 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1572 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1573 which of course cannot start UTF-8 characters. I have changed the code so
1574 that only real ASCII characters (less than 128) and the correct starting
1575 bytes for UTF-8 encodings are set for characters greater than 127 when in
1576 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1577 altogether.)
1578
1579 20. Added the /T option to pcretest so as to be able to run tests with non-
1580 standard character tables, thus making it possible to include the tests
1581 used for 19 above in the standard set of tests.
1582
1583 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1584 reference to a subpattern the other side of a comment that contains an
1585 opening parenthesis caused either an internal compiling error, or a
1586 reference to the wrong subpattern.
1587
1588
1589 Version 8.02 19-Mar-2010
1590 ------------------------
1591
1592 1. The Unicode data tables have been updated to Unicode 5.2.0.
1593
1594 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1595 configured.
1596
1597 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1598 original author of that file, following a query about its status.
1599
1600 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1601 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1602
1603 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1604 quantifier applied to a forward-referencing subroutine call, could compile
1605 incorrect code or give the error "internal error: previously-checked
1606 referenced subpattern not found".
1607
1608 6. Both MS Visual Studio and Symbian OS have problems with initializing
1609 variables to point to external functions. For these systems, therefore,
1610 pcre_malloc etc. are now initialized to local functions that call the
1611 relevant global functions.
1612
1613 7. There were two entries missing in the vectors called coptable and poptable
1614 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1615 I've fixed the data, and added a kludgy way of testing at compile time that
1616 the lengths are correct (equal to the number of opcodes).
1617
1618 8. Following on from 7, I added a similar kludge to check the length of the
1619 eint vector in pcreposix.c.
1620
1621 9. Error texts for pcre_compile() are held as one long string to avoid too
1622 much relocation at load time. To find a text, the string is searched,
1623 counting zeros. There was no check for running off the end of the string,
1624 which could happen if a new error number was added without updating the
1625 string.
1626
1627 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1628
1629 11. \K was not working if it appeared in an atomic group or in a group that
1630 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1631 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1632 the assertion is positive, but not if it is negative.
1633
1634 12. Change 11 fortuitously reduced the size of the stack frame used in the
1635 "match()" function of pcre_exec.c by one pointer. Forthcoming
1636 implementation of support for (*MARK) will need an extra pointer on the
1637 stack; I have reserved it now, so that the stack frame size does not
1638 decrease.
1639
1640 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1641 item in branch that calls a recursion is a subroutine call - as in the
1642 second branch in the above example - was incorrectly given the compile-
1643 time error "recursive call could loop indefinitely" because pcre_compile()
1644 was not correctly checking the subroutine for matching a non-empty string.
1645
1646 14. The checks for overrunning compiling workspace could trigger after an
1647 overrun had occurred. This is a "should never occur" error, but it can be
1648 triggered by pathological patterns such as hundreds of nested parentheses.
1649 The checks now trigger 100 bytes before the end of the workspace.
1650
1651 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1652
1653
1654 Version 8.01 19-Jan-2010
1655 ------------------------
1656
1657 1. If a pattern contained a conditional subpattern with only one branch (in
1658 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1659 computed the wrong minimum data length (which is of course zero for such
1660 subpatterns). This could cause incorrect "no match" results.
1661
1662 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1663 the pattern is reset in the first branch, pcre_compile() failed with
1664 "internal error: code overflow at offset...". This happened only when
1665 the reset was to the original external option setting. (An optimization
1666 abstracts leading options settings into an external setting, which was the
1667 cause of this.)
1668
1669 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1670 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1671 assertion pattern did not match (meaning that the assertion was true), it
1672 was incorrectly treated as false if the SKIP had been reached during the
1673 matching. This also applied to assertions used as conditions.
1674
1675 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1676 assertion subpattern, including such a pattern used as a condition,
1677 unpredictable results occurred, instead of the error return
1678 PCRE_ERROR_DFA_UITEM.
1679
1680 5. The C++ GlobalReplace function was not working like Perl for the special
1681 situation when an empty string is matched. It now does the fancy magic
1682 stuff that is necessary.
1683
1684 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1685 removed. (These were left over from very, very early versions of PCRE.)
1686
1687 7. Some cosmetic changes to the code to make life easier when compiling it
1688 as part of something else:
1689
1690 (a) Change DEBUG to PCRE_DEBUG.
1691
1692 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1693 called "current" as "current_branch", to prevent a collision with the
1694 Linux macro when compiled as a kernel module.
1695
1696 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1697 prevent a collision with the Linux macro when compiled as a kernel
1698 module.
1699
1700 8. In pcre_compile() there are some checks for integer overflows that used to
1701 cast potentially large values to (double). This has been changed to that
1702 when building, a check for int64_t is made, and if it is found, it is used
1703 instead, thus avoiding the use of floating point arithmetic. (There is no
1704 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1705 double.
1706
1707 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1708 2005 (difference between two addresses compared to an unsigned value).
1709
1710 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1711 custom one, because of the following reported problem in Windows:
1712
1713 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1714 under Win32.
1715 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1716 therefore missing the function definition.
1717 - The compiler thus generates a "C" signature for the test function.
1718 - The linker fails to find the "C" function.
1719 - PCRE fails to configure if asked to do so against libbz2.
1720
1721 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1722 messages were output:
1723
1724 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1725 rerunning libtoolize, to keep the correct libtool macros in-tree.
1726 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1727
1728 I have done both of these things.
1729
1730 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1731 most of the time, it *can* run out if it is given a pattern that contains a
1732 runaway infinite recursion. I updated the discussion in the pcrestack man
1733 page.
1734
1735 13. Now that we have gone to the x.xx style of version numbers, the minor
1736 version may start with zero. Using 08 or 09 is a bad idea because users
1737 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1738 interpreted as invalid octal numbers. I've updated the previous comment in
1739 configure.ac, and also added a check that gives an error if 08 or 09 are
1740 used.
1741
1742 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1743 causing partial matching to fail when the end of the subject matched \W
1744 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1745
1746 15. There were some discrepancies between the declarations in pcre_internal.h
1747 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1748 their definitions. The declarations used "const uschar *" and the
1749 definitions used USPTR. Even though USPTR is normally defined as "const
1750 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1751 reported that: "This difference in casting confuses some C++ compilers, for
1752 example, SunCC recognizes above declarations as different functions and
1753 generates broken code for hbpcre." I have changed the declarations to use
1754 USPTR.
1755
1756 16. GNU libtool is named differently on some systems. The autogen.sh script now
1757 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1758 (FreeBSD).
1759
1760 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1761 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1762 comment: "Figure out how to create a longlong from a string: strtoll and
1763 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1764 instance, but it only takes 2 args instead of 3!"
1765
1766 18. A subtle bug concerned with back references has been fixed by a change of
1767 specification, with a corresponding code fix. A pattern such as
1768 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1769 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1770 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1771 same bug. Such groups have to be quantified to be useful, or contained
1772 inside another quantified group. (If there's no repetition, the reference
1773 can never match.) The problem arises because, having left the group and
1774 moved on to the rest of the pattern, a later failure that backtracks into
1775 the group uses the captured value from the final iteration of the group
1776 rather than the correct earlier one. I have fixed this in PCRE by forcing
1777 any group that contains a reference to itself to be an atomic group; that
1778 is, there cannot be any backtracking into it once it has completed. This is
1779 similar to recursive and subroutine calls.
1780
1781
1782 Version 8.00 19-Oct-09
1783 ----------------------
1784
1785 1. The table for translating pcre_compile() error codes into POSIX error codes
1786 was out-of-date, and there was no check on the pcre_compile() error code
1787 being within the table. This could lead to an OK return being given in
1788 error.
1789
1790 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1791 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1792 in a Windows environment.
1793
1794 3. The pcregrep --count option prints the count for each file even when it is
1795 zero, as does GNU grep. However, pcregrep was also printing all files when
1796 --files-with-matches was added. Now, when both options are given, it prints
1797 counts only for those files that have at least one match. (GNU grep just
1798 prints the file name in this circumstance, but including the count seems
1799 more useful - otherwise, why use --count?) Also ensured that the
1800 combination -clh just lists non-zero counts, with no names.
1801
1802 4. The long form of the pcregrep -F option was incorrectly implemented as
1803 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1804 but it seems right to fix it, and I didn't think it was worth preserving
1805 the old behaviour.
1806
1807 5. The command line items --regex=pattern and --regexp=pattern were not
1808 recognized by pcregrep, which required --regex pattern or --regexp pattern
1809 (with a space rather than an '='). The man page documented the '=' forms,
1810 which are compatible with GNU grep; these now work.
1811
1812 6. No libpcreposix.pc file was created for pkg-config; there was just
1813 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1814
1815 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1816 when UCP support is not needed, by modifying the Python script that
1817 generates it from Unicode data files. This should not matter if the module
1818 is correctly used as a library, but I received one complaint about 50K of
1819 unwanted data. My guess is that the person linked everything into his
1820 program rather than using a library. Anyway, it does no harm.
1821
1822 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1823 was a minimum greater than 1 for a wide character in a possessive
1824 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1825 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1826 character. Chaos in the form of incorrect output or a compiling loop could
1827 result.
1828
1829 9. The restrictions on what a pattern can contain when partial matching is
1830 requested for pcre_exec() have been removed. All patterns can now be
1831 partially matched by this function. In addition, if there are at least two
1832 slots in the offset vector, the offset of the earliest inspected character
1833 for the match and the offset of the end of the subject are set in them when
1834 PCRE_ERROR_PARTIAL is returned.
1835
1836 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1837 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1838 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1839 and may be more useful for multi-segment matching.
1840
1841 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1842 used to be given if ever the end of the subject was reached; now it is
1843 given only if matching could not proceed because another character was
1844 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1845 string "Z", which now yields "no match" instead of "partial match". In the
1846 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1847 final character ended with (*FAIL).
1848
1849 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1850 if the pattern had a "must contain" character that was already found in the
1851 earlier partial match, unless partial matching was again requested. For
1852 example, with the pattern /dog.(body)?/, the "must contain" character is
1853 "g". If the first part-match was for the string "dog", restarting with
1854 "sbody" failed. This bug has been fixed.
1855
1856 13. The string returned by pcre_dfa_exec() after a partial match has been
1857 changed so that it starts at the first inspected character rather than the
1858 first character of the match. This makes a difference only if the pattern
1859 starts with a lookbehind assertion or \b or \B (\K is not supported by
1860 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1861 matching functions compatible, and I think it's the right thing to do.
1862
1863 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1864 so that the demonstration program is easily available in environments where
1865 PCRE has not been installed from source.
1866
1867 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1868 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1869 library.
1870
1871 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1872 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1873 is not the first non-POSIX option to be added. Clearly some people find
1874 these options useful.
1875
1876 17. If a caller to the POSIX matching function regexec() passes a non-zero
1877 value for nmatch with a NULL value for pmatch, the value of
1878 nmatch is forced to zero.
1879
1880 18. RunGrepTest did not have a test for the availability of the -u option of
1881 the diff command, as RunTest does. It now checks in the same way as
1882 RunTest, and also checks for the -b option.
1883
1884 19. If an odd number of negated classes containing just a single character
1885 interposed, within parentheses, between a forward reference to a named
1886 subpattern and the definition of the subpattern, compilation crashed with
1887 an internal error, complaining that it could not find the referenced
1888 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1889 [The bug was that it was starting one character too far in when skipping
1890 over the character class, thus treating the ] as data rather than
1891 terminating the class. This meant it could skip too much.]
1892
1893 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1894 /g option in pcretest when the pattern contains \K, which makes it possible
1895 to have an empty string match not at the start, even when the pattern is
1896 anchored. Updated pcretest and pcredemo to use this option.
1897
1898 21. If the maximum number of capturing subpatterns in a recursion was greater
1899 than the maximum at the outer level, the higher number was returned, but
1900 with unset values at the outer level. The correct (outer level) value is
1901 now given.
1902
1903 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1904 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1905 make it do so. The string so far is captured, making this feature
1906 compatible with Perl.
1907
1908 23. The tests have been re-organized, adding tests 11 and 12, to make it
1909 possible to check the Perl 5.10 features against Perl 5.10.
1910
1911 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1912 pattern matches a fixed length string. PCRE did not allow this; now it
1913 does. Neither allows recursion.
1914
1915 25. I finally figured out how to implement a request to provide the minimum
1916 length of subject string that was needed in order to match a given pattern.
1917 (It was back references and recursion that I had previously got hung up
1918 on.) This code has now been added to pcre_study(); it finds a lower bound
1919 to the length of subject needed. It is not necessarily the greatest lower
1920 bound, but using it to avoid searching strings that are too short does give
1921 some useful speed-ups. The value is available to calling programs via
1922 pcre_fullinfo().
1923
1924 26. While implementing 25, I discovered to my embarrassment that pcretest had
1925 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1926 study optimizations had never been tested with that matching function.
1927 Oops. What is worse, even when it was passed study data, there was a bug in
1928 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1929 were also very few tests of studied patterns with pcre_dfa_exec().
1930
1931 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1932 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1933 on the other side of the coin, they are no longer allowed to have different
1934 names, because these cannot be distinguished in PCRE, and this has caused
1935 confusion. (This is a difference from Perl.)
1936
1937 28. When duplicate subpattern names are present (necessarily with different
1938 numbers, as required by 27 above), and a test is made by name in a
1939 conditional pattern, either for a subpattern having been matched, or for
1940 recursion in such a pattern, all the associated numbered subpatterns are
1941 tested, and the overall condition is true if the condition is true for any
1942 one of them. This is the way Perl works, and is also more like the way
1943 testing by number works.
1944
1945
1946 Version 7.9 11-Apr-09
1947 ---------------------
1948
1949 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1950 (pcretest), all targets were linked against these libraries. This included
1951 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1952 libraries. This caused unwanted dependencies to be created. This problem
1953 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1954 pcretest is linked with readline.
1955
1956 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1957 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1958 moved outside it again, because FALSE and TRUE are already defined in AIX,
1959 but BOOL is not.
1960
1961 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1962 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1963
1964 4. The pcregrep documentation said spaces were inserted as well as colons (or
1965 hyphens) following file names and line numbers when outputting matching
1966 lines. This is not true; no spaces are inserted. I have also clarified the
1967 wording for the --colour (or --color) option.
1968
1969 5. In pcregrep, when --colour was used with -o, the list of matching strings
1970 was not coloured; this is different to GNU grep, so I have changed it to be
1971 the same.
1972
1973 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1974 each matching line was coloured. Now it goes on to look for further matches
1975 of any of the test patterns, which is the same behaviour as GNU grep.
1976
1977 7. A pattern that could match an empty string could cause pcregrep to loop; it
1978 doesn't make sense to accept an empty string match in pcregrep, so I have
1979 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1980 seems to be how GNU grep behaves. [But see later change 40 for release
1981 8.33.]
1982
1983 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1984 start or after a newline", because the conditional assertion was not being
1985 correctly handled. The rule now is that both the assertion and what follows
1986 in the first alternative must satisfy the test.
1987
1988 9. If auto-callout was enabled in a pattern with a conditional group whose
1989 condition was an assertion, PCRE could crash during matching, both with
1990 pcre_exec() and pcre_dfa_exec().
1991
1992 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1993 used for matching.
1994
1995 11. Unicode property support in character classes was not working for
1996 characters (bytes) greater than 127 when not in UTF-8 mode.
1997
1998 12. Added the -M command line option to pcretest.
1999
2000 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2001
2002 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2003
2004 16. Added comments and documentation about mis-use of no_arg in the C++
2005 wrapper.
2006
2007 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2008 from Martin Jerabek that uses macro names for all relevant character and
2009 string constants.
2010
2011 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2012 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2013 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2014 these, but not everybody uses configure.
2015
2016 19. A conditional group that had only one branch was not being correctly
2017 recognized as an item that could match an empty string. This meant that an
2018 enclosing group might also not be so recognized, causing infinite looping
2019 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2020 with the subject "ab", where knowledge that the repeated group can match
2021 nothing is needed in order to break the loop.
2022
2023 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2024 exec(), but without supplying a callout function, matching went wrong.
2025
2026 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2027 leak if the size of the offset vector was greater than 30. When the vector
2028 is smaller, the saved offsets during recursion go onto a local stack
2029 vector, but for larger vectors malloc() is used. It was failing to free
2030 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2031 error, in fact).
2032
2033 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2034 heapframe that is used only when UTF-8 support is enabled. This caused no
2035 problem, but was untidy.
2036
2037 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2038 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2039 included within another project.
2040
2041 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2042 slightly modified by me:
2043
2044 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2045 not building pcregrep.
2046
2047 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2048 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2049
2050 25. Forward references, both numeric and by name, in patterns that made use of
2051 duplicate group numbers, could behave incorrectly or give incorrect errors,
2052 because when scanning forward to find the reference group, PCRE was not
2053 taking into account the duplicate group numbers. A pattern such as
2054 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2055
2056 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2057 the feature of a custom pointer more persuasive (as requested by a user).
2058
2059 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2060 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2061 pre-defined.
2062
2063 28. Added support for (*UTF8) at the start of a pattern.
2064
2065 29. Arrange for flags added by the "release type" setting in CMake to be shown
2066 in the configuration summary.
2067
2068
2069 Version 7.8 05-Sep-08
2070 ---------------------
2071
2072 1. Replaced UCP searching code with optimized version as implemented for Ad
2073 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2074 stage table and inline lookup instead of a function, giving speed ups of 2
2075 to 5 times on some simple patterns that I tested. Permission was given to
2076 distribute the MultiStage2.py script that generates the tables (it's not in
2077 the tarball, but is in the Subversion repository).
2078
2079 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2080 scripts.
2081
2082 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2083 a group with a zero qualifier. The result of the study could be incorrect,
2084 or the function might crash, depending on the pattern.
2085
2086 4. Caseless matching was not working for non-ASCII characters in back
2087 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2088 It now works when Unicode Property Support is available.
2089
2090 5. In pcretest, an escape such as \x{de} in the data was always generating
2091 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2092 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2093 truncation.
2094
2095 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2096
2097 7. Added two (int) casts to pcregrep when printing the difference of two
2098 pointers, in case they are 64-bit values.
2099
2100 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2101 test 2 if it fails.
2102
2103 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2104 and a #define of that name to empty if it is not externally set. This is to
2105 allow users of MSVC to set it if necessary.
2106
2107 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2108 the convenience functions in the pcre_get.c source file.
2109
2110 11. An option change at the start of a pattern that had top-level alternatives
2111 could cause overwriting and/or a crash. This command provoked a crash in
2112 some environments:
2113
2114 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2115
2116 This potential security problem was recorded as CVE-2008-2371.
2117
2118 12. For a pattern where the match had to start at the beginning or immediately
2119 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2120 pcre_dfa_exec() could read past the end of the passed subject if there was
2121 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2122 pcretest so that it places the subject at the end of its malloc-ed buffer.
2123
2124 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2125 exec() might read past the end of the data buffer in UTF-8 mode.
2126
2127 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2128 the data contained the byte 0x85 as part of a UTF-8 character within its
2129 first line. This applied both to normal and DFA matching.
2130
2131 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2132 /^[^d]*?$/8 failed to match "abc".
2133
2134 16. Added a missing copyright notice to pcrecpp_internal.h.
2135
2136 17. Make it more clear in the documentation that values returned from
2137 pcre_exec() in ovector are byte offsets, not character counts.
2138
2139 18. Tidied a few places to stop certain compilers from issuing warnings.
2140
2141 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2142 supplied by Stefan Weber. I made a further small update for 7.8 because
2143 there is a change of source arrangements: the pcre_searchfuncs.c module is
2144 replaced by pcre_ucd.c.
2145
2146
2147 Version 7.7 07-May-08
2148 ---------------------
2149
2150 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2151 a string to a long long, pretend we don't even have a long long." This is
2152 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2153
2154 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2155 pre-7.6 versions, which defined a global no_arg variable instead of putting
2156 it in the RE class. (See also #8 below.)
2157
2158 3. Remove a line of dead code, identified by coverity and reported by Nuno
2159 Lopes.
2160
2161 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2162
2163 (1) The include/exclude patterns were being applied to the whole pathnames
2164 of files, instead of just to the final components.
2165
2166 (2) If there was more than one level of directory, the subdirectories were
2167 skipped unless they satisfied the include/exclude conditions. This is
2168 inconsistent with GNU grep (and could even be seen as contrary to the
2169 pcregrep specification - which I improved to make it absolutely clear).
2170 The action now is always to scan all levels of directory, and just
2171 apply the include/exclude patterns to regular files.
2172
2173 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2174 --exclude_dir in the tests to avoid scanning .svn directories.
2175
2176 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2177 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2178 doesn't support NULs in patterns.
2179
2180 7. Added some missing "const"s to declarations of static tables in
2181 pcre_compile.c and pcre_dfa_exec.c.
2182
2183 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2184 caused by fix #2 above. (Subsequently also a second patch to fix the
2185 first patch. And a third patch - this was a messy problem.)
2186
2187 9. Applied Craig's patch to remove the use of push_back().
2188
2189 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2190 matching function regexec().
2191
2192 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2193 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2194 references. PCRE supports relative numbers with this syntax (I don't think
2195 Oniguruma does).
2196
2197 12. Previously, a group with a zero repeat such as (...){0} was completely
2198 omitted from the compiled regex. However, this means that if the group
2199 was called as a subroutine from elsewhere in the pattern, things went wrong
2200 (an internal error was given). Such groups are now left in the compiled
2201 pattern, with a new opcode that causes them to be skipped at execution
2202 time.
2203
2204 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2205 to the way PCRE behaves:
2206
2207 (a) A lone ] character is dis-allowed (Perl treats it as data).
2208
2209 (b) A back reference to an unmatched subpattern matches an empty string
2210 (Perl fails the current match path).
2211
2212 (c) A data ] in a character class must be notated as \] because if the
2213 first data character in a class is ], it defines an empty class. (In
2214 Perl it is not possible to have an empty class.) The empty class []
2215 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2216 The negative empty class [^] matches any one character, independently
2217 of the DOTALL setting.
2218
2219 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2220 non-existent subpattern following a character class starting with ']' and
2221 containing () gave an internal compiling error instead of "reference to
2222 non-existent subpattern". Fortunately, when the pattern did exist, the
2223 compiled code was correct. (When scanning forwards to check for the
2224 existence of the subpattern, it was treating the data ']' as terminating
2225 the class, so got the count wrong. When actually compiling, the reference
2226 was subsequently set up correctly.)
2227
2228 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2229 it was being rejected as not supported by pcre_dfa_exec(), even though
2230 other assertions are supported. I have made pcre_dfa_exec() support
2231 (*FAIL).
2232
2233 16. The implementation of 13c above involved the invention of a new opcode,
2234 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2235 cannot be changed at match time, I realized I could make a small
2236 improvement to matching performance by compiling OP_ALLANY instead of
2237 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2238 on the OP_ANY path.
2239
2240 17. Compiling pcretest on Windows with readline support failed without the
2241 following two fixes: (1) Make the unistd.h include conditional on
2242 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2243
2244 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2245 ncurses library to be included for pcretest when ReadLine support is
2246 requested, but also to allow for it to be overridden. This patch came from
2247 Daniel Bergström.
2248
2249 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2250 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2251 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2252 spotting this.
2253
2254
2255 Version 7.6 28-Jan-08
2256 ---------------------
2257
2258 1. A character class containing a very large number of characters with
2259 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2260 overflow.
2261
2262 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2263 HAVE_LONG_LONG is not defined.
2264
2265 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2266 bring them up to date and include new features. This patch includes:
2267
2268 - Fixed PH's badly added libz and libbz2 support.
2269 - Fixed a problem with static linking.
2270 - Added pcredemo. [But later removed - see 7 below.]
2271 - Fixed dftables problem and added an option.
2272 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2273 HAVE_LONG_LONG.
2274 - Added readline support for pcretest.
2275 - Added an listing of the option settings after cmake has run.
2276
2277 4. A user submitted a patch to Makefile that makes it easy to create
2278 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2279 Makefile.am that cause it to include this special target, without
2280 affecting anything else. Note that the same mingw target plus all
2281 the other distribution libraries and programs are now supported
2282 when configuring with CMake (see 6 below) instead of with
2283 Configure/Make.
2284
2285 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2286 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2287 exported in the Windows port". It has not yet been confirmed that the patch
2288 solves the problem, but it does no harm.
2289
2290 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2291 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2292 with CMake, and also correct the comment about stack recursion.
2293
2294 7. Remove the automatic building of pcredemo from the ./configure system and
2295 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2296 of a program that users should build themselves after PCRE is installed, so
2297 building it automatically is not really right. What is more, it gave
2298 trouble in some build environments.
2299
2300 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2301
2302
2303 Version 7.5 10-Jan-08
2304 ---------------------
2305
2306 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2307 values in parens when parsing an RE using the C++ wrapper."
2308
2309 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2310 Characters greater than 255 were excluded from the class instead of being
2311 included.
2312
2313 3. The same bug as (2) above applied to negated POSIX classes such as
2314 [:^space:].
2315
2316 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2317 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2318 I have changed it.
2319
2320 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2321 first named subpattern) and a construct such as (?&a) would reference the
2322 first named subpattern whose name started with "a" (in other words, the
2323 length check was missing). Both these problems are fixed. "Subpattern name
2324 expected" is now given for (?&) (a zero-length name), and this patch also
2325 makes it give the same error for \k'' (previously it complained that that
2326 was a reference to a non-existent subpattern).
2327
2328 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2329 this is right because (?- can be followed by option settings as well as by
2330 digits. I have, however, made the messages clearer.
2331
2332 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2333 than the number used in the conditional) now cause a compile-time error.
2334 This is actually not compatible with Perl, which accepts such patterns, but
2335 treats the conditional as always being FALSE (as PCRE used to), but it
2336 seems to me that giving a diagnostic is better.
2337
2338 8. Change "alphameric" to the more common word "alphanumeric" in comments
2339 and messages.
2340
2341 9. Fix two occurrences of "backslash" in comments that should have been
2342 "backspace".
2343
2344 10. Remove two redundant lines of code that can never be obeyed (their function
2345 was moved elsewhere).
2346
2347 11. The program that makes PCRE's Unicode character property table had a bug
2348 which caused it to generate incorrect table entries for sequences of
2349 characters that have the same character type, but are in different scripts.
2350 It amalgamated them into a single range, with the script of the first of
2351 them. In other words, some characters were in the wrong script. There were
2352 thirteen such cases, affecting characters in the following ranges:
2353
2354 U+002b0 - U+002c1
2355 U+0060c - U+0060d
2356 U+0061e - U+00612
2357 U+0064b - U+0065e
2358 U+0074d - U+0076d
2359 U+01800 - U+01805
2360 U+01d00 - U+01d77
2361 U+01d9b - U+01dbf
2362 U+0200b - U+0200f
2363 U+030fc - U+030fe
2364 U+03260 - U+0327f
2365 U+0fb46 - U+0fbb1
2366 U+10450 - U+1049d
2367
2368 12. The -o option (show only the matching part of a line) for pcregrep was not
2369 compatible with GNU grep in that, if there was more than one match in a
2370 line, it showed only the first of them. It now behaves in the same way as
2371 GNU grep.
2372
2373 13. If the -o and -v options were combined for pcregrep, it printed a blank
2374 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2375 does the same. The return code can be used to tell if there were any
2376 non-matching lines.
2377
2378 14. Added --file-offsets and --line-offsets to pcregrep.
2379
2380 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2381 infinitely looping recursion. The bug was that positive lookaheads were not
2382 being skipped when checking for a possible empty match (negative lookaheads
2383 and both kinds of lookbehind were skipped).
2384
2385 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2386 inclusion of <windows.h> to before rather than after the definition of
2387 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2388
2389 17. Specifying a possessive quantifier with a specific limit for a Unicode
2390 character property caused pcre_compile() to compile bad code, which led at
2391 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2392 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2393 caused the error; without that there was no problem.
2394
2395 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2396
2397 19. Added --enable-pcretest-libreadline.
2398
2399 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2400 RE::GlobalReplace(). As a result, the number of replacements returned was
2401 double what it should be. I removed one of the increments, but Craig sent a
2402 later patch that removed the other one (the right fix) and added unit tests
2403 that check the return values (which was not done before).
2404
2405 21. Several CMake things:
2406
2407 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2408 the names libpcre and libpcreposix, not just pcre and pcreposix.
2409
2410 (2) The above change means that pcretest and pcregrep are now correctly
2411 linked with the newly-built libraries, not previously installed ones.
2412
2413 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2414
2415 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2416 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2417 UTF-8 newline character). The key issue is that the pattern starts .*;
2418 this means that the match must be either at the beginning, or after a
2419 newline. The bug was in the code for advancing after a failed match and
2420 checking that the new position followed a newline. It was not taking
2421 account of UTF-8 characters correctly.
2422
2423 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2424 character classes. PCRE was not treating the sequence [:...:] as a
2425 character class unless the ... were all letters. Perl, however, seems to
2426 allow any characters between [: and :], though of course it rejects as
2427 unknown any "names" that contain non-letters, because all the known class
2428 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2429 for example, whereas PCRE did not - it did not recognize a POSIX character
2430 class. This seemed a bit dangerous, so the code has been changed to be
2431 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2432 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2433 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2434 Perl does, and where it didn't before.
2435
2436 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2437 Windows environments %n is disabled by default.
2438
2439
2440 Version 7.4 21-Sep-07
2441 ---------------------
2442
2443 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2444 means that a class such as [\s] counted as "explicit reference to CR or
2445 LF". That isn't really right - the whole point of the change was to try to
2446 help when there was an actual mention of one of the two characters. So now
2447 the change happens only if \r or \n (or a literal CR or LF) character is
2448 encountered.
2449
2450 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2451 of both had grown to the point where there were only 3 bits left.
2452 Fortunately, there was spare space in the data structure, and so I have
2453 moved the internal flags into a new 16-bit field to free up more option
2454 bits.
2455
2456 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2457 but did not set the internal JCHANGED flag - either of these is enough to
2458 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2459 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2460 start sets both bits.
2461
2462 4. Added options (at build time, compile time, exec time) to change \R from
2463 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2464
2465 5. doc/pcresyntax.html was missing from the distribution.
2466
2467 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2468 compatibility, even though it is no longer used.
2469
2470 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2471 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2472 windows.h file is present (where different names are used). [This was
2473 reversed later after testing - see 16 below.]
2474
2475 8. Changed all #include <config.h> to #include "config.h". There were also
2476 some further <pcre.h> cases that I changed to "pcre.h".
2477
2478 9. When pcregrep was used with the --colour option, it missed the line ending
2479 sequence off the lines that it output.
2480
2481 10. It was pointed out to me that arrays of string pointers cause lots of
2482 relocations when a shared library is dynamically loaded. A technique of
2483 using a single long string with a table of offsets can drastically reduce
2484 these. I have refactored PCRE in four places to do this. The result is
2485 dramatic:
2486
2487 Originally: 290
2488 After changing UCP table: 187
2489 After changing error message table: 43
2490 After changing table of "verbs" 36
2491 After changing table of Posix names 22
2492
2493 Thanks to the folks working on Gregex for glib for this insight.
2494
2495 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2496 unicode-properties was also set.
2497
2498 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2499
2500 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2501 checked only for CRLF.
2502
2503 14. Added casts to pcretest.c to avoid compiler warnings.
2504
2505 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2506
2507 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2508 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2509 entirely. This removes changes made in 7 above.
2510
2511 17. The CMake files have been updated, and there is now more information about
2512 building with CMake in the NON-UNIX-USE document.
2513
2514
2515 Version 7.3 28-Aug-07
2516 ---------------------
2517
2518 1. In the rejigging of the build system that eventually resulted in 7.1, the
2519 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2520 brackets there is not right, since it causes compilers to look for an
2521 installed pcre.h, not the version that is in the source that is being
2522 compiled (which of course may be different). I have changed it back to:
2523
2524 #include "pcre.h"
2525
2526 I have a vague recollection that the change was concerned with compiling in
2527 different directories, but in the new build system, that is taken care of
2528 by the VPATH setting the Makefile.
2529
2530 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2531 when the subject happened to end in the byte 0x85 (e.g. if the last
2532 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2533 characters but of course it shouldn't be taken as a newline when it is part
2534 of another character. The bug was that, for an unlimited repeat of . in
2535 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2536 characters when looking for a newline.
2537
2538 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2539
2540 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2541 in debug output.
2542
2543 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2544 long printing in the pcrecpp unittest when running under MinGW.
2545
2546 6. ESC_K was left out of the EBCDIC table.
2547
2548 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2549 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2550 limit also applies to "virtual nesting" when a pattern is recursive, and in
2551 this case 1000 isn't so big. I have been able to remove this limit at the
2552 expense of backing off one optimization in certain circumstances. Normally,
2553 when pcre_exec() would call its internal match() function recursively and
2554 immediately return the result unconditionally, it uses a "tail recursion"
2555 feature to save stack. However, when a subpattern that can match an empty
2556 string has an unlimited repetition quantifier, it no longer makes this
2557 optimization. That gives it a stack frame in which to save the data for
2558 checking that an empty string has been matched. Previously this was taken
2559 from the 1000-entry workspace that had been reserved. So now there is no
2560 explicit limit, but more stack is used.
2561
2562 8. Applied Daniel's patches to solve problems with the import/export magic
2563 syntax that is required for Windows, and which was going wrong for the
2564 pcreposix and pcrecpp parts of the library. These were overlooked when this
2565 problem was solved for the main library.
2566
2567 9. There were some crude static tests to avoid integer overflow when computing
2568 the size of patterns that contain repeated groups with explicit upper
2569 limits. As the maximum quantifier is 65535, the maximum group length was
2570 set at 30,000 so that the product of these two numbers did not overflow a
2571 32-bit integer. However, it turns out that people want to use groups that
2572 are longer than 30,000 bytes (though not repeat them that many times).
2573 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2574 made it possible to implement the integer overflow checks in a much more
2575 dynamic way, which I have now done. The artificial limitation on group
2576 length has been removed - we now have only the limit on the total length of
2577 the compiled pattern, which depends on the LINK_SIZE setting.
2578
2579 10. Fixed a bug in the documentation for get/copy named substring when
2580 duplicate names are permitted. If none of the named substrings are set, the
2581 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2582 empty string.
2583
2584 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2585 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2586 because the ] is interpreted as the first data character and the
2587 terminating ] is not found. PCRE has been made compatible with Perl in this
2588 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2589 cause memory overwriting.
2590
2591 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2592 string has been matched (to stop an infinite loop). It was not recognizing
2593 a conditional subpattern that could match an empty string if that
2594 subpattern was within another subpattern. For example, it looped when
2595 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2596 condition was not nested. This bug has been fixed.
2597
2598 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2599 past the start of the subject in the presence of bytes with the top bit
2600 set, for example "\x8aBCD".
2601
2602 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2603 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2604
2605 14. Optimized (?!) to (*FAIL).
2606
2607 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2608 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2609 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2610 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2611 does: it's just the validity check that is more restrictive.
2612
2613 16. Inserted checks for integer overflows during escape sequence (backslash)
2614 processing, and also fixed erroneous offset values for syntax errors during
2615 backslash processing.
2616
2617 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2618 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2619
2620 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2621 caused an overrun.
2622
2623 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2624 something other than just ASCII characters) inside a group that had an
2625 unlimited repeat caused a loop at compile time (while checking to see
2626 whether the group could match an empty string).
2627
2628 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2629 [\P{Any}] did so. (Error in the code for printing property names.)
2630
2631 21. An orphan \E inside a character class could cause a crash.
2632
2633 22. A repeated capturing bracket such as (A)? could cause a wild memory
2634 reference during compilation.
2635
2636 23. There are several functions in pcre_compile() that scan along a compiled
2637 expression for various reasons (e.g. to see if it's fixed length for look
2638 behind). There were bugs in these functions when a repeated \p or \P was
2639 present in the pattern. These operators have additional parameters compared
2640 with \d, etc, and these were not being taken into account when moving along
2641 the compiled data. Specifically:
2642
2643 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2644 length.
2645
2646 (b) An item such as \pL+ within a repeated group could cause crashes or
2647 loops.
2648
2649 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2650 "reference to non-existent subpattern" error.
2651
2652 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2653
2654 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2655 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2656
2657 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2658
2659 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2660 character were causing crashes (broken optimization).
2661
2662 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2663 \p or \P) caused a compile-time loop.
2664
2665 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2666 break. For example, the unstudied pattern [\r\n]A does not match the string
2667 "\r\nA" because change 7.0/46 below moves the current point on by two
2668 characters after failing to match at the start. However, the pattern \nA
2669 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2670 the same is true. There doesn't seem any very clean way out of this, but
2671 what I have chosen to do makes the common cases work: PCRE now takes note
2672 of whether there can be an explicit match for \r or \n anywhere in the
2673 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2674 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2675 pattern has explicit CR or LF references.
2676
2677 29. Added (*CR) etc for changing newline setting at start of pattern.
2678
2679
2680 Version 7.2 19-Jun-07
2681 ---------------------
2682
2683 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2684 which is apparently normally available under Windows.
2685
2686 2. Re-jig the pcregrep tests with different newline settings in an attempt
2687 to make them independent of the local environment's newline setting.
2688
2689 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2690
2691 4. Some of the "internals" tests were previously cut out when the link size
2692 was not 2, because the output contained actual offsets. The recent new
2693 "Z" feature of pcretest means that these can be cut out, making the tests
2694 usable with all link sizes.
2695
2696 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2697 stack recursion. This gives a massive performance boost under BSD, but just
2698 a small improvement under Linux. However, it saves one field in the frame
2699 in all cases.
2700
2701 6. Added more features from the forthcoming Perl 5.10:
2702
2703 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2704 recursion call. It refers to the nth most recently opened parentheses.
2705
2706 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2707 to be opened parentheses.
2708
2709 (c) Conditions that refer to capturing parentheses can be specified
2710 relatively, for example, (?(-2)... or (?(+3)...
2711
2712 (d) \K resets the start of the current match so that everything before
2713 is not part of it.
2714
2715 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2716
2717 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2718 reference syntax.
2719
2720 (g) (?| introduces a group in which the numbering of parentheses in each
2721 alternative starts with the same number.
2722
2723 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2724
2725 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2726 PCRE_INFO_JCHANGED.
2727
2728 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2729 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2730 for detecting groups that can match an empty string.
2731
2732 9. A pattern with a very large number of alternatives (more than several
2733 hundred) was running out of internal workspace during the pre-compile
2734 phase, where pcre_compile() figures out how much memory will be needed. A
2735 bit of new cunning has reduced the workspace needed for groups with
2736 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2737 workspace instead of running out of the 4096 that are available.
2738
2739 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2740
2741 11. Applied patch from Google to remove an optimization that didn't quite work.
2742 The report of the bug said:
2743
2744 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2745 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2746 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2747
2748 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2749 it matched the wrong number of bytes.
2750
2751
2752 Version 7.1 24-Apr-07
2753 ---------------------
2754
2755 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2756 that is more "standard", making use of automake and other Autotools. There
2757 is some re-arrangement of the files and adjustment of comments consequent
2758 on this.
2759
2760 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2761 for recursive directory scanning broke on some systems because the files
2762 are not scanned in any specific order and on different systems the order
2763 was different. A call to "sort" has been inserted into RunGrepTest for the
2764 approprate test as a short-term fix. In the longer term there may be an
2765 alternative.
2766
2767 3. I had an email from Eric Raymond about problems translating some of PCRE's
2768 man pages to HTML (despite the fact that I distribute HTML pages, some
2769 people do their own conversions for various reasons). The problems
2770 concerned the use of low-level troff macros .br and .in. I have therefore
2771 removed all such uses from the man pages (some were redundant, some could
2772 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2773 HTML has been updated to handle .nf/.fi and to complain if it encounters
2774 .br or .in.
2775
2776 4. Updated comments in configure.ac that get placed in config.h.in and also
2777 arranged for config.h to be included in the distribution, with the name
2778 config.h.generic, for the benefit of those who have to compile without
2779 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2780
2781 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2782 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2783 makevp.bat for the current PCRE, using the additional files
2784 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2785
2786 6. A Windows user reported a minor discrepancy with test 2, which turned out
2787 to be caused by a trailing space on an input line that had got lost in his
2788 copy. The trailing space was an accident, so I've just removed it.
2789
2790 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2791 that is needed.
2792
2793 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2794 as "const" (a) because they are and (b) because it helps the PHP
2795 maintainers who have recently made a script to detect big data structures
2796 in the php code that should be moved to the .rodata section. I remembered
2797 to update Builducptable as well, so it won't revert if ucptable.h is ever
2798 re-created.
2799
2800 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2801 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2802 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2803 support is not required. This saves 1.5-2K of code, which is important in
2804 some applications.
2805
2806 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2807 so as not to refer to the tables, even though these functions will never be
2808 called when UTF-8 support is disabled. Otherwise there are problems with a
2809 shared library.
2810
2811 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2812
2813 (a) It was defining its arguments as char * instead of void *.
2814
2815 (b) It was assuming that all moves were upwards in memory; this was true
2816 a long time ago when I wrote it, but is no longer the case.
2817
2818 The emulated memove() is provided for those environments that have neither
2819 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2820 is clearly not the case, as these two bugs were recently reported.
2821
2822 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2823 and Detrail to create the HTML documentation, the .txt form of the man
2824 pages, and it removes trailing spaces from listed files. It also creates
2825 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2826 case, it wraps all the #defines with #ifndefs. This script should be run
2827 before "make dist".
2828
2829 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2830 with Unicode property support.
2831
2832 (a) For a maximizing quantifier, if the two different cases of the
2833 character were of different lengths in their UTF-8 codings (there are
2834 some cases like this - I found 11), and the matching function had to
2835 back up over a mixture of the two cases, it incorrectly assumed they
2836 were both the same length.
2837
2838 (b) When PCRE was configured to use the heap rather than the stack for
2839 recursion during matching, it was not correctly preserving the data for
2840 the other case of a UTF-8 character when checking ahead for a match
2841 while processing a minimizing repeat. If the check also involved
2842 matching a wide character, but failed, corruption could cause an
2843 erroneous result when trying to check for a repeat of the original
2844 character.
2845
2846 13. Some tidying changes to the testing mechanism:
2847
2848 (a) The RunTest script now detects the internal link size and whether there
2849 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2850 values substituted by "configure". (The RunGrepTest script already did
2851 this for UTF-8.) The configure.ac script no longer substitutes the
2852 relevant variables.
2853
2854 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2855 with length and offset values. This means that the output is different
2856 for different internal link sizes. Test 2 is skipped for link sizes
2857 other than 2 because of this, bypassing the problem. Unfortunately,
2858 there was also a test in test 3 (the locale tests) that used /B and
2859 failed for link sizes other than 2. Rather than cut the whole test out,
2860 I have added a new /Z option to pcretest that replaces the length and
2861 offset values with spaces. This is now used to make test 3 independent
2862 of link size. (Test 2 will be tidied up later.)
2863
2864 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2865 segmentation fault instead of returning the appropriate error message.
2866
2867 15. In multiline mode when the newline sequence was set to "any", the pattern
2868 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2869 This doesn't seem right; it now treats the CRLF combination as the line
2870 ending, and so does not match in that case. It's only a pattern such as ^$
2871 that would hit this one: something like ^ABC$ would have failed after \r
2872 and then tried again after \r\n.
2873
2874 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2875 in an attempt to make files that differ only in their line terminators
2876 compare equal. This works on Linux.
2877
2878 17. Under certain error circumstances pcregrep might try to free random memory
2879 as it exited. This is now fixed, thanks to valgrind.
2880
2881 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2882 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2883 was because its rules for how to advance for /g after matching an empty
2884 string at the end of a line did not allow for this case. They now check for
2885 it specially.
2886
2887 20. pcretest is supposed to handle patterns and data of any length, by
2888 extending its buffers when necessary. It was getting this wrong when the
2889 buffer for a data line had to be extended.
2890
2891 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2892 CRLF as a newline sequence.
2893
2894 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2895 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2896 I have nevertheless tidied it up.
2897
2898 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2899
2900 24. Added a man page for pcre-config.
2901
2902
2903 Version 7.0 19-Dec-06
2904 ---------------------
2905
2906 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2907 moving to gcc 4.1.1.
2908
2909 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2910 sys/time.h, which is documented as needed for this function. It doesn't
2911 seem to matter on Linux, but it showed up on some releases of OS X.
2912
2913 3. It seems that there are systems where bytes whose values are greater than
2914 127 match isprint() in the "C" locale. The "C" locale should be the
2915 default when a C program starts up. In most systems, only ASCII printing
2916 characters match isprint(). This difference caused the output from pcretest
2917 to vary, making some of the tests fail. I have changed pcretest so that:
2918
2919 (a) When it is outputting text in the compiled version of a pattern, bytes
2920 other than 32-126 are always shown as hex escapes.
2921
2922 (b) When it is outputting text that is a matched part of a subject string,
2923 it does the same, unless a different locale has been set for the match
2924 (using the /L modifier). In this case, it uses isprint() to decide.
2925
2926 4. Fixed a major bug that caused incorrect computation of the amount of memory
2927 required for a compiled pattern when options that changed within the
2928 pattern affected the logic of the preliminary scan that determines the
2929 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2930 that the computed length was too small. The symptoms of this bug were
2931 either the PCRE error "internal error: code overflow" from pcre_compile(),
2932 or a glibc crash with a message such as "pcretest: free(): invalid next
2933 size (fast)". Examples of patterns that provoked this bug (shown in
2934 pcretest format) are:
2935
2936 /(?-x: )/x
2937 /(?x)(?-x: \s*#\s*)/
2938 /((?i)[\x{c0}])/8
2939 /(?i:[\x{c0}])/8
2940
2941 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2942 is now done differently.
2943
2944 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2945 wrapper classes; (b) implement a new function in the C++ scanner that is
2946 more efficient than the old way of doing things because it avoids levels of
2947 recursion in the regex matching; (c) add a paragraph to the documentation
2948 for the FullMatch() function.
2949
2950 6. The escape sequence \n was being treated as whatever was defined as
2951 "newline". Not only was this contrary to the documentation, which states
2952 that \n is character 10 (hex 0A), but it also went horribly wrong when
2953 "newline" was defined as CRLF. This has been fixed.
2954
2955 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2956 was being set to -1 for the "end of line" case (supposedly a value that no
2957 character can have). Though this value is never used (the check for end of
2958 line is "zero bytes in current character"), it caused compiler complaints.
2959 I've changed it to 0xffffffff.
2960
2961 8. In pcre_version.c, the version string was being built by a sequence of
2962 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2963 string (as it is for production releases) called a macro with an empty
2964 argument. The C standard says the result of this is undefined. The gcc
2965 compiler treats it as an empty string (which was what was wanted) but it is
2966 reported that Visual C gives an error. The source has been hacked around to
2967 avoid this problem.
2968
2969 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2970 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2971 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2972 of them did).
2973
2974 10. Originally, pcretest opened its input and output without "b"; then I was
2975 told that "b" was needed in some environments, so it was added for release
2976 5.0 to both the input and output. (It makes no difference on Unix-like
2977 systems.) Later I was told that it is wrong for the input on Windows. I've
2978 now abstracted the modes into two macros, to make it easier to fiddle with
2979 them, and removed "b" from the input mode under Windows.
2980
2981 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2982
2983 12. Added -help and --help to pcretest as an official way of being reminded
2984 of the options.
2985
2986 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2987 and pcrecpp.cc because they annoy compilers at high warning levels.
2988
2989 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2990
2991 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2992 scripts are not C programs :-) and which was not noticed because it works
2993 on Linux.
2994
2995 16. pcretest is supposed to handle any length of pattern and data line (as one
2996 line or as a continued sequence of lines) by extending its input buffer if
2997 necessary. This feature was broken for very long pattern lines, leading to
2998 a string of junk being passed to pcre_compile() if the pattern was longer
2999 than about 50K.
3000
3001 17. I have done a major re-factoring of the way pcre_compile() computes the
3002 amount of memory needed for a compiled pattern. Previously, there was code
3003 that made a preliminary scan of the pattern in order to do this. That was
3004 OK when PCRE was new, but as the facilities have expanded, it has become
3005 harder and harder to keep it in step with the real compile phase, and there
3006 have been a number of bugs (see for example, 4 above). I have now found a
3007 cunning way of running the real compile function in a "fake" mode that
3008 enables it to compute how much memory it would need, while actually only
3009 ever using a few hundred bytes of working memory and without too many
3010 tests of the mode. This should make future maintenance and development
3011 easier. A side effect of this work is that the limit of 200 on the nesting
3012 depth of parentheses has been removed (though this was never a serious
3013 limitation, I suspect). However, there is a downside: pcre_compile() now
3014 runs more slowly than before (30% or more, depending on the pattern). I
3015 hope this isn't a big issue. There is no effect on runtime performance.
3016
3017 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3018 newline (only possible for the last line of a file) and it was a
3019 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3020
3021 19. Added additional timing features to pcretest. (1) The -tm option now times
3022 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3023 separate command line item, by a number that specifies the number of
3024 repeats to use when timing. The default is 50000; this gives better
3025 precision, but takes uncomfortably long for very large patterns.
3026
3027 20. Extended pcre_study() to be more clever in cases where a branch of a
3028 subpattern has no definite first character. For example, (a*|b*)[cd] would
3029 previously give no result from pcre_study(). Now it recognizes that the
3030 first character must be a, b, c, or d.
3031
3032 21. There was an incorrect error "recursive call could loop indefinitely" if
3033 a subpattern (or the entire pattern) that was being tested for matching an
3034 empty string contained only one non-empty item after a nested subpattern.
3035 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3036 incorrectly, because the \d was being skipped in the check.
3037
3038 22. The pcretest program now has a new pattern option /B and a command line
3039 option -b, which is equivalent to adding /B to every pattern. This causes
3040 it to show the compiled bytecode, without the additional information that
3041 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3042 is the same as /B/I).
3043
3044 23. A new optimization is now able automatically to treat some sequences such
3045 as a*b as a*+b. More specifically, if something simple (such as a character
3046 or a simple class like \d) has an unlimited quantifier, and is followed by
3047 something that cannot possibly match the quantified thing, the quantifier
3048 is automatically "possessified".
3049
3050 24. A recursive reference to a subpattern whose number was greater than 39
3051 went wrong under certain circumstances in UTF-8 mode. This bug could also
3052 have affected the operation of pcre_study().
3053
3054 25. Realized that a little bit of performance could be had by replacing
3055 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3056
3057 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3058
3059 27. Possessive quantifiers such as a++ were previously implemented by turning
3060 them into atomic groups such as ($>a+). Now they have their own opcodes,
3061 which improves performance. This includes the automatically created ones
3062 from 23 above.
3063
3064 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3065 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3066 the first matched character to be a colon. This applied both to named and
3067 numbered groups.
3068
3069 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3070
3071 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3072 building PCRE on Windows easier, so I have included it in the distribution.
3073
3074 31. There is now a check in pcretest against a ridiculously large number being
3075 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3076 loop, the loop is abandoned.
3077
3078 32. Forward references to subpatterns in conditions such as (?(2)...) where
3079 subpattern 2 is defined later cause pcre_compile() to search forwards in
3080 the pattern for the relevant set of parentheses. This search went wrong
3081 when there were unescaped parentheses in a character class, parentheses
3082 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3083
3084 33. "Subroutine" calls and backreferences were previously restricted to
3085 referencing subpatterns earlier in the regex. This restriction has now
3086 been removed.
3087
3088 34. Added a number of extra features that are going to be in Perl 5.10. On the
3089 whole, these are just syntactic alternatives for features that PCRE had
3090 previously implemented using the Python syntax or my own invention. The
3091 other formats are all retained for compatibility.
3092
3093 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3094 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3095 also .NET compatible.
3096
3097 (b) A recursion or subroutine call to a named group can now be defined as
3098 (?&name) as well as (?P>name).
3099
3100 (c) A backreference to a named group can now be defined as \k<name> or
3101 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3102 5.10, are also .NET compatible.
3103
3104 (d) A conditional reference to a named group can now use the syntax
3105 (?(<name>) or (?('name') as well as (?(name).
3106
3107 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3108 groups (named and numbered) that are never evaluated inline, but can be
3109 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3110 is always false. There may be only one alternative in such a group.
3111
3112 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3113 as the simple (?(R). The condition is true only if the most recent
3114 recursion is that of the given number or name. It does not search out
3115 through the entire recursion stack.
3116
3117 (g) The escape \gN or \g{N} has been added, where N is a positive or
3118 negative number, specifying an absolute or relative reference.
3119
3120 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3121 some "unreachable code" warnings.
3122
3123 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3124 things, this adds five new scripts.
3125
3126 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3127 There were also incompatibilities regarding the handling of \Q..\E inside
3128 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3129 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3130
3131 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3132 matches an empty string, and forcibly breaks the loop. There were bugs in
3133 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3134 against aaaa the result was just "a" rather than "aaaa", for example. Two
3135 separate and independent bugs (that affected different cases) have been
3136 fixed.
3137
3138 39. Refactored the code to abolish the use of different opcodes for small
3139 capturing bracket numbers. This is a tidy that I avoided doing when I
3140 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3141 The new approach is not only tidier, it makes it possible to reduce the
3142 memory needed to fix the previous bug (38).
3143
3144 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3145 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3146 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3147 mode.
3148
3149 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3150 report.
3151
3152 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3153 copy construction and assignment in the C++ wrapper.
3154
3155 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3156 couple of bugs that could have given wrong results in the "--newline=crlf"
3157 case.
3158
3159 44. Added a number of casts and did some reorganization of signed/unsigned int
3160 variables following suggestions from Dair Grant. Also renamed the variable
3161 "this" as "item" because it is a C++ keyword.
3162
3163 45. Arranged for dftables to add
3164
3165 #include "pcre_internal.h"
3166
3167 to pcre_chartables.c because without it, gcc 4.x may remove the array
3168 definition from the final binary if PCRE is built into a static library and
3169 dead code stripping is activated.
3170
3171 46. For an unanchored pattern, if a match attempt fails at the start of a
3172 newline sequence, and the newline setting is CRLF or ANY, and the next two
3173 characters are CRLF, advance by two characters instead of one.
3174
3175
3176 Version 6.7 04-Jul-06
3177 ---------------------
3178
3179 1. In order to handle tests when input lines are enormously long, pcretest has
3180 been re-factored so that it automatically extends its buffers when
3181 necessary. The code is crude, but this _is_ just a test program. The
3182 default size has been increased from 32K to 50K.
3183
3184 2. The code in pcre_study() was using the value of the re argument before
3185 testing it for NULL. (Of course, in any sensible call of the function, it
3186 won't be NULL.)
3187
3188 3. The memmove() emulation function in pcre_internal.h, which is used on
3189 systems that lack both memmove() and bcopy() - that is, hardly ever -
3190 was missing a "static" storage class specifier.
3191
3192 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3193 containing an extended class (one that cannot be represented by a bitmap
3194 because it contains high-valued characters or Unicode property items, e.g.
3195 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3196 pattern, but PCRE should not loop if you do not (it no longer does).
3197 [Detail: two cases were found: (a) a repeated subpattern containing an
3198 extended class; (b) a recursive reference to a subpattern that followed a
3199 previous extended class. It wasn't skipping over the extended class
3200 correctly when UTF-8 mode was not set.]
3201
3202 5. A negated single-character class was not being recognized as fixed-length
3203 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3204 compile error "lookbehind assertion is not fixed length".
3205
3206 6. The RunPerlTest auxiliary script was showing an unexpected difference
3207 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3208 write a Perl script that can interpret lines of an input file either as
3209 byte characters or as UTF-8, which is what "perltest" was being required to
3210 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3211 can't do is switch easily at run time between having the "use utf8;" pragma
3212 or not. In the end, I fudged it by using the RunPerlTest script to insert
3213 "use utf8;" explicitly for the UTF-8 tests.
3214
3215 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3216 the end of the subject string, contrary to the documentation and to what
3217 Perl does. This was true of both matching functions. Now it matches only at
3218 the start of the subject and immediately after *internal* newlines.
3219
3220 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3221 a pointer to an int instead of a pointer to an unsigned long int. This
3222 caused problems on 64-bit systems.
3223
3224 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3225 instance of the 'standard' template library not being so standard".
3226
3227 10. There was no check on the number of named subpatterns nor the maximum
3228 length of a subpattern name. The product of these values is used to compute
3229 the size of the memory block for a compiled pattern. By supplying a very
3230 long subpattern name and a large number of named subpatterns, the size
3231 computation could be caused to overflow. This is now prevented by limiting
3232 the length of names to 32 characters, and the number of named subpatterns
3233 to 10,000.
3234
3235 11. Subpatterns that are repeated with specific counts have to be replicated in
3236 the compiled pattern. The size of memory for this was computed from the
3237 length of the subpattern and the repeat count. The latter is limited to
3238 65535, but there was no limit on the former, meaning that integer overflow
3239 could in principle occur. The compiled length of a repeated subpattern is
3240 now limited to 30,000 bytes in order to prevent this.
3241
3242 12. Added the optional facility to have named substrings with the same name.
3243
3244 13. Added the ability to use a named substring as a condition, using the
3245 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3246 are numbers (not recommended). Forward references are permitted.
3247
3248 14. Added forward references in named backreferences (if you see what I mean).
3249
3250 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3251 pattern could run off the end of the subject. For example, the pattern
3252 "(?s)(.{1,5})"8 did this with the subject "ab".
3253
3254 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3255 PCRE_CASELESS was set when matching characters that were quantified with ?
3256 or *.
3257
3258 17. A character class other than a single negated character that had a minimum
3259 but no maximum quantifier - for example [ab]{6,} - was not handled
3260 correctly by pce_dfa_exec(). It would match only one character.
3261
3262 18. A valid (though odd) pattern that looked like a POSIX character
3263 class but used an invalid character after [ (for example [[,abc,]]) caused
3264 pcre_compile() to give the error "Failed: internal error: code overflow" or
3265 in some cases to crash with a glibc free() error. This could even happen if
3266 the pattern terminated after [[ but there just happened to be a sequence of
3267 letters, a binary zero, and a closing ] in the memory that followed.
3268
3269 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3270 over the years. Originally (before any Unicode support), just the bottom 8
3271 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3272 output from "man perlunicode" includes this:
3273
3274 The regular expression compiler produces polymorphic opcodes. That
3275 is, the pattern adapts to the data and automatically switches to
3276 the Unicode character scheme when presented with Unicode data--or
3277 instead uses a traditional byte scheme when presented with byte
3278 data.
3279
3280 Sadly, a wide octal escape does not cause a switch, and in a string with
3281 no other multibyte characters, these octal escapes are treated as before.
3282 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3283 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3284 Unicode string.
3285
3286 I have not perpetrated such confusion in PCRE. Up till now, it took just
3287 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3288 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3289 translate to the appropriate multibyte character.
3290
3291 29. Applied some refactoring to reduce the number of warnings from Microsoft
3292 and Borland compilers. This has included removing the fudge introduced
3293 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3294 a warning about an unused variable.
3295
3296 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3297 characters since release 4.0, because Perl (from release 5.004) does not.
3298 [Or at least, is documented not to: some releases seem to be in conflict
3299 with the documentation.] However, when a pattern was studied with
3300 pcre_study() and all its branches started with \s, PCRE still included VT
3301 as a possible starting character. Of course, this did no harm; it just
3302 caused an unnecessary match attempt.
3303
3304 22. Removed a now-redundant internal flag bit that recorded the fact that case
3305 dependency changed within the pattern. This was once needed for "required
3306 byte" processing, but is no longer used. This recovers a now-scarce options
3307 bit. Also moved the least significant internal flag bit to the most-
3308 significant bit of the word, which was not previously used (hangover from
3309 the days when it was an int rather than a uint) to free up another bit for
3310 the future.
3311
3312 23. Added support for CRLF line endings as well as CR and LF. As well as the
3313 default being selectable at build time, it can now be changed at runtime
3314 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3315 specify that it is scanning data with non-default line endings.
3316
3317 24. Changed the definition of CXXLINK to make it agree with the definition of
3318 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3319
3320 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3321 recursions. This makes a big different to stack usage for some patterns.
3322
3323 26. If a subpattern containing a named recursion or subroutine reference such
3324 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3325 the space required for the compiled pattern went wrong and gave too small a
3326 value. Depending on the environment, this could lead to "Failed: internal
3327 error: code overflow at offset 49" or "glibc detected double free or
3328 corruption" errors.
3329
3330 27. Applied patches from Google (a) to support the new newline modes and (b) to
3331 advance over multibyte UTF-8 characters in GlobalReplace.
3332
3333 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3334 difference for some implementation of PCRE in some Windows version.
3335
3336 29. Added some extra testing facilities to pcretest:
3337
3338 \q<number> in a data line sets the "match limit" value
3339 \Q<number> in a data line sets the "match recursion limt" value
3340 -S <number> sets the stack size, where <number> is in megabytes
3341
3342 The -S option isn't available for Windows.
3343
3344
3345 Version 6.6 06-Feb-06
3346 ---------------------
3347
3348 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3349 in pcreposix.h. I have copied the definition from pcre.h.
3350
3351 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3352 because pcre.h is no longer a built file.
3353
3354 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3355 not normally included in the compiled code.
3356
3357
3358 Version 6.5 01-Feb-06
3359 ---------------------
3360
3361 1. When using the partial match feature with pcre_dfa_exec(), it was not
3362 anchoring the second and subsequent partial matches at the new starting
3363 point. This could lead to incorrect results. For example, with the pattern
3364 /1234/, partially matching against "123" and then "a4" gave a match.
3365
3366 2. Changes to pcregrep:
3367
3368 (a) All non-match returns from pcre_exec() were being treated as failures
3369 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3370 error message is output. Some extra information is given for the
3371 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3372 probably the only errors that are likely to be caused by users (by
3373 specifying a regex that has nested indefinite repeats, for instance).
3374 If there are more than 20 of these errors, pcregrep is abandoned.
3375
3376 (b) A binary zero was treated as data while matching, but terminated the
3377 output line if it was written out. This has been fixed: binary zeroes
3378 are now no different to any other data bytes.
3379
3380 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3381 used to set a locale for matching. The --locale=xxxx long option has
3382 been added (no short equivalent) to specify a locale explicitly on the
3383 pcregrep command, overriding the environment variables.
3384
3385 (d) When -B was used with -n, some line numbers in the output were one less
3386 than they should have been.
3387
3388 (e) Added the -o (--only-matching) option.
3389
3390 (f) If -A or -C was used with -c (count only), some lines of context were
3391 accidentally printed for the final match.
3392
3393 (g) Added the -H (--with-filename) option.
3394
3395 (h) The combination of options -rh failed to suppress file names for files
3396 that were found from directory arguments.
3397
3398 (i) Added the -D (--devices) and -d (--directories) options.
3399
3400 (j) Added the -F (--fixed-strings) option.
3401
3402 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3403
3404 (l) Added the --colo(u)r option.
3405
3406 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3407 is not present by default.
3408
3409 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3410 items such as (?R) or (?1), when the recursion could match a number of
3411 alternatives. If it matched one of the alternatives, but subsequently,
3412 outside the recursion, there was a failure, the code tried to back up into
3413 the recursion. However, because of the way PCRE is implemented, this is not
3414 possible, and the result was an incorrect result from the match.
3415
3416 In order to prevent this happening, the specification of recursion has
3417 been changed so that all such subpatterns are automatically treated as
3418 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3419
3420 4. I had overlooked the fact that, in some locales, there are characters for
3421 which isalpha() is true but neither isupper() nor islower() are true. In
3422 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3423 and ordfeminine) are like this. This affected the treatment of \w and \W
3424 when they appeared in character classes, but not when they appeared outside
3425 a character class. The bit map for "word" characters is now created
3426 separately from the results of isalnum() instead of just taking it from the
3427 upper, lower, and digit maps. (Plus the underscore character, of course.)
3428
3429 5. The above bug also affected the handling of POSIX character classes such as
3430 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3431 permanent tables. Instead, the bit maps for such a class were previously
3432 created as the appropriate unions of the upper, lower, and digit bitmaps.
3433 Now they are created by subtraction from the [[:word:]] class, which has
3434 its own bitmap.
3435
3436 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3437 It is created by subtracting the vertical space characters (\x09, \x0a,
3438 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3439 subtraction was done in the overall bitmap for a character class, meaning
3440 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3441 be recognized. This bug has been fixed.
3442
3443 7. Patches from the folks at Google:
3444
3445 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3446 real life, but is still worth protecting against".
3447
3448 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3449 regular expressions".
3450
3451 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3452 have it.
3453
3454 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3455 "configure" and the latter not, in order to fix a problem somebody had
3456 with compiling the Arg class on HP-UX.
3457
3458 (e) Improve the error-handling of the C++ wrapper a little bit.
3459
3460 (f) New tests for checking recursion limiting.
3461
3462 8. The pcre_memmove() function, which is used only if the environment does not
3463 have a standard memmove() function (and is therefore rarely compiled),
3464 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3465 returning a result (though PCRE never actually uses the result).
3466
3467 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3468 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3469 returned instead of calling malloc() with an overflowing number that would
3470 most likely cause subsequent chaos.
3471
3472 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3473
3474 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3475 with this option is matched, the nmatch and pmatch options of regexec() are
3476 ignored.
3477
3478 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3479 provided in case anyone wants to the the POSIX interface with UTF-8
3480 strings.
3481
3482 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3483 C++ linking (needed for some HP-UX environments).
3484
3485 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3486 (unused parameter) and in the pcre_printint() function (omitted "default"
3487 switch label when the default is to do nothing).
3488
3489 15. Added some code to make it possible, when PCRE is compiled as a C++
3490 library, to replace subject pointers for pcre_exec() with a smart pointer
3491 class, thus making it possible to process discontinuous strings.
3492
3493 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3494 much the same function. They were added by different people who were trying
3495 to make PCRE easy to compile on non-Unix systems. It has been suggested
3496 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3497 for compiling on Windows systems. I have therefore replaced it with
3498 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3499 defaults to "extern" for C or "extern C" for C++, which works fine on
3500 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3501 SCOPE with something explicit in config.h. In addition:
3502
3503 (a) pcreposix.h still had just "extern" instead of either of these macros;
3504 I have replaced it with PCRE_DATA_SCOPE.
3505
3506 (b) Functions such as _pcre_xclass(), which are internal to the library,
3507 but external in the C sense, all had PCRE_EXPORT in their definitions.
3508 This is apparently wrong for the Windows case, so I have removed it.
3509 (It makes no difference on Unix-like systems.)
3510
3511 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3512 of recursive calls to match(). This is different to MATCH_LIMIT because
3513 that limits the total number of calls to match(), not all of which increase
3514 the depth of recursion. Limiting the recursion depth limits the amount of
3515 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3516 when PCRE is compiled, and changed at run time. A patch from Google adds
3517 this functionality to the C++ interface.
3518
3519 18. Changes to the handling of Unicode character properties:
3520
3521 (a) Updated the table to Unicode 4.1.0.
3522
3523 (b) Recognize characters that are not in the table as "Cn" (undefined).
3524
3525 (c) I revised the way the table is implemented to a much improved format
3526 which includes recognition of ranges. It now supports the ranges that
3527 are defined in UnicodeData.txt, and it also amalgamates other
3528 characters into ranges. This has reduced the number of entries in the
3529 table from around 16,000 to around 3,000, thus reducing its size
3530 considerably. I realized I did not need to use a tree structure after
3531 all - a binary chop search is just as efficient. Having reduced the
3532 number of entries, I extended their size from 6 bytes to 8 bytes to
3533 allow for more data.
3534
3535 (d) Added support for Unicode script names via properties such as \p{Han}.
3536
3537 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3538 matching that character.
3539
3540 20. When matching a repeated Unicode property with a minimum greater than zero,
3541 (for example \pL{2,}), PCRE could look past the end of the subject if it
3542 reached it while seeking the minimum number of characters. This could
3543 happen only if some of the characters were more than one byte long, because
3544 there is a check for at least the minimum number of bytes.
3545
3546 21. Refactored the implementation of \p and \P so as to be more general, to
3547 allow for more different types of property in future. This has changed the
3548 compiled form incompatibly. Anybody with saved compiled patterns that use
3549 \p or \P will have to recompile them.
3550
3551 22. Added "Any" and "L&" to the supported property types.
3552
3553 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3554 but give a compile time error if the value is greater than 0xff.
3555
3556 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3557 accidentally not being installed or uninstalled.
3558
3559 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3560 made were to insert the current release number. This seemed silly, because
3561 it made things harder for people building PCRE on systems that don't run
3562 "configure". I have turned pcre.h into a distributed file, no longer built
3563 by "configure", with the version identification directly included. There is
3564 no longer a pcre.h.in file.
3565
3566 However, this change necessitated a change to the pcre-config script as
3567 well. It is built from pcre-config.in, and one of the substitutions was the
3568 release number. I have updated configure.ac so that ./configure now finds
3569 the release number by grepping pcre.h.
3570
3571 26. Added the ability to run the tests under valgrind.
3572
3573
3574 Version 6.4 05-Sep-05
3575 ---------------------
3576
3577 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3578 "--" to be printed when multiple files were scanned, even when none of the
3579 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3580 consider it to be a bug, and have restored the previous behaviour.
3581
3582 2. A couple of code tidies to get rid of compiler warnings.
3583
3584 3. The pcretest program used to cheat by referring to symbols in the library
3585 whose names begin with _pcre_. These are internal symbols that are not
3586 really supposed to be visible externally, and in some environments it is
3587 possible to suppress them. The cheating is now confined to including
3588 certain files from the library's source, which is a bit cleaner.
3589
3590 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3591 file's purpose clearer.
3592
3593 5. Reorganized pcre_ucp_findchar().
3594
3595
3596 Version 6.3 15-Aug-05
3597 ---------------------
3598
3599 1. The file libpcre.pc.in did not have general read permission in the tarball.
3600
3601 2. There were some problems when building without C++ support:
3602
3603 (a) If C++ support was not built, "make install" and "make test" still
3604 tried to test it.
3605
3606 (b) There were problems when the value of CXX was explicitly set. Some
3607 changes have been made to try to fix these, and ...
3608
3609 (c) --disable-cpp can now be used to explicitly disable C++ support.
3610
3611 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3612 backslash in a target when C++ was disabled. This confuses some
3613 versions of "make", apparently. Using an intermediate variable solves
3614 this. (Same for CPP_LOBJ.)
3615
3616 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3617 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3618 necessary on certain architectures.
3619
3620 4. Added a setting of -export-symbols-regex to the link command to remove
3621 those symbols that are exported in the C sense, but actually are local
3622 within the library, and not documented. Their names all begin with
3623 "_pcre_". This is not a perfect job, because (a) we have to except some
3624 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3625 available (and never for static libraries). I have made a note to try to
3626 find a way round (a) in the future.
3627
3628
3629 Version 6.2 01-Aug-05
3630 ---------------------
3631
3632 1. There was no test for integer overflow of quantifier values. A construction
3633 such as {1111111111111111} would give undefined results. What is worse, if
3634 a minimum quantifier for a parenthesized subpattern overflowed and became
3635 negative, the calculation of the memory size went wrong. This could have
3636 led to memory overwriting.
3637
3638 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3639
3640 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3641 operating environments where this matters.
3642
3643 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3644 PCRE options from within the C++ wrapper.
3645
3646 5. Named capturing subpatterns were not being correctly counted when a pattern
3647 was compiled. This caused two problems: (a) If there were more than 100
3648 such subpatterns, the calculation of the memory needed for the whole
3649 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3650 back references of the form \12, where the number was greater than 9, were
3651 not recognized as back references, even though there were sufficient
3652 previous subpatterns.
3653
3654 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3655 versions of gcc, e.g. 2.95.4.
3656
3657
3658 Version 6.1 21-Jun-05
3659 ---------------------
3660
3661 1. There was one reference to the variable "posix" in pcretest.c that was not
3662 surrounded by "#if !defined NOPOSIX".
3663
3664 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3665 the cross-check on the old pcre_info() function, for the benefit of the
3666 cut-down version of PCRE that is currently imported into Exim.
3667
3668 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3669 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3670 patterns that start (?i) but I don't think that matters. The use of (?i) is
3671 just an example; this all applies to the other options as well.
3672
3673 4. Since libtool seems to echo the compile commands it is issuing, the output
3674 from "make" can be reduced a bit by putting "@" in front of each libtool
3675 compile command.
3676
3677 5. Patch from the folks at Google for configure.in to be a bit more thorough
3678 in checking for a suitable C++ installation before trying to compile the
3679 C++ stuff. This should fix a reported problem when a compiler was present,
3680 but no suitable headers.
3681
3682 6. The man pages all had just "PCRE" as their title. I have changed them to
3683 be the relevant file name. I have also arranged that these names are
3684 retained in the file doc/pcre.txt, which is a concatenation in text format
3685 of all the man pages except the little individual ones for each function.
3686
3687 7. The NON-UNIX-USE file had not been updated for the different set of source
3688 files that come with release 6. I also added a few comments about the C++
3689 wrapper.
3690
3691
3692 Version 6.0 07-Jun-05
3693 ---------------------
3694
3695 1. Some minor internal re-organization to help with my DFA experiments.
3696
3697 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3698 didn't matter for the library itself when fully configured, but did matter
3699 when compiling without UCP support, or within Exim, where the ucp files are
3700 not imported.
3701
3702 3. Refactoring of the library code to split up the various functions into
3703 different source modules. The addition of the new DFA matching code (see
3704 below) to a single monolithic source would have made it really too
3705 unwieldy, quite apart from causing all the code to be include in a
3706 statically linked application, when only some functions are used. This is
3707 relevant even without the DFA addition now that patterns can be compiled in
3708 one application and matched in another.
3709
3710 The downside of splitting up is that there have to be some external
3711 functions and data tables that are used internally in different modules of
3712 the library but which are not part of the API. These have all had their
3713 names changed to start with "_pcre_" so that they are unlikely to clash
3714 with other external names.
3715
3716 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3717 a different (DFA) algorithm. Although it is slower than the original
3718 function, it does have some advantages for certain types of matching
3719 problem.
3720
3721 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3722 including restarting after a partial match.
3723
3724 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3725 defined when compiling for Windows was sent to me. I have put it into the
3726 code, though I have no means of testing or verifying it.
3727
3728 7. Added the pcre_refcount() auxiliary function.
3729
3730 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3731 match before or at the first newline in the subject string. In pcretest,
3732 the /f option on a pattern can be used to set this.
3733
3734 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3735 would behave wrongly. This has been present in PCRE since release 4.0.
3736
3737 10. A number of changes to the pcregrep command:
3738
3739 (a) Refactored how -x works; insert ^(...)$ instead of setting
3740 PCRE_ANCHORED and checking the length, in preparation for adding
3741 something similar for -w.
3742
3743 (b) Added the -w (match as a word) option.
3744
3745 (c) Refactored the way lines are read and buffered so as to have more
3746 than one at a time available.
3747
3748 (d) Implemented a pcregrep test script.
3749
3750 (e) Added the -M (multiline match) option. This allows patterns to match
3751 over several lines of the subject. The buffering ensures that at least
3752 8K, or the rest of the document (whichever is the shorter) is available
3753 for matching (and similarly the previous 8K for lookbehind assertions).
3754
3755 (f) Changed the --help output so that it now says
3756
3757 -w, --word-regex(p)
3758
3759 instead of two lines, one with "regex" and the other with "regexp"
3760 because that confused at least one person since the short forms are the
3761 same. (This required a bit of code, as the output is generated
3762 automatically from a table. It wasn't just a text change.)
3763
3764 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3765 option but starts with a hyphen. Could be a pattern or a path name
3766 starting with a hyphen, for instance.
3767
3768 (h) "-" can be given as a file name to represent stdin.
3769
3770 (i) When file names are being printed, "(standard input)" is used for
3771 the standard input, for compatibility with GNU grep. Previously
3772 "<stdin>" was used.
3773
3774 (j) The option --label=xxx can be used to supply a name to be used for
3775 stdin when file names are being printed. There is no short form.
3776
3777 (k) Re-factored the options decoding logic because we are going to add
3778 two more options that take data. Such options can now be given in four
3779 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3780
3781 (l) Added the -A, -B, and -C options for requesting that lines of context
3782 around matches be printed.
3783
3784 (m) Added the -L option to print the names of files that do not contain
3785 any matching lines, that is, the complement of -l.
3786
3787 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3788 continue to scan other files.
3789
3790 (o) The -s option was incorrectly implemented. For compatibility with other
3791 greps, it now suppresses the error message for a non-existent or non-
3792 accessible file (but not the return code). There is a new option called
3793 -q that suppresses the output of matching lines, which was what -s was
3794 previously doing.
3795
3796 (p) Added --include and --exclude options to specify files for inclusion
3797 and exclusion when recursing.
3798
3799 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3800 Hopefully, it now does.
3801
3802 12. Missing cast in pcre_study().
3803
3804 13. Added an "uninstall" target to the makefile.
3805
3806 14. Replaced "extern" in the function prototypes in Makefile.in with
3807 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3808 world, but is set differently for Windows.
3809
3810 15. Added a second compiling function called pcre_compile2(). The only
3811 difference is that it has an extra argument, which is a pointer to an
3812 integer error code. When there is a compile-time failure, this is set
3813 non-zero, in addition to the error test pointer being set to point to an
3814 error message. The new argument may be NULL if no error number is required
3815 (but then you may as well call pcre_compile(), which is now just a
3816 wrapper). This facility is provided because some applications need a
3817 numeric error indication, but it has also enabled me to tidy up the way
3818 compile-time errors are handled in the POSIX wrapper.
3819
3820 16. Added VPATH=.libs to the makefile; this should help when building with one
3821 prefix path and installing with another. (Or so I'm told by someone who
3822 knows more about this stuff than I do.)
3823
3824 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3825 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3826 match everything, including newlines. This is not POSIX-compatible, but
3827 somebody wanted the feature. From pcretest it can be activated by using
3828 both the P and the s flags.
3829
3830 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3831
3832 19. libpcre.pc was being incorrectly installed as executable.
3833
3834 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3835 it now also looks for '\r' so that it will work unmodified on Windows.
3836
3837 21. Added Google's contributed C++ wrapper to the distribution.
3838
3839 22. Added some untidy missing memory free() calls in pcretest, to keep
3840 Electric Fence happy when testing.
3841
3842
3843
3844 Version 5.0 13-Sep-04
3845 ---------------------
3846
3847 1. Internal change: literal characters are no longer packed up into items
3848 containing multiple characters in a single byte-string. Each character
3849 is now matched using a separate opcode. However, there may be more than one
3850 byte in the character in UTF-8 mode.
3851
3852 2. The pcre_callout_block structure has two new fields: pattern_position and
3853 next_item_length. These contain the offset in the pattern to the next match
3854 item, and its length, respectively.
3855
3856 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3857 insertion of callouts before each pattern item. Added the /C option to
3858 pcretest to make use of this.
3859
3860 4. On the advice of a Windows user, the lines
3861
3862 #if defined(_WIN32) || defined(WIN32)
3863 _setmode( _fileno( stdout ), 0x8000 );
3864 #endif /* defined(_WIN32) || defined(WIN32) */
3865
3866 have been added to the source of pcretest. This apparently does useful
3867 magic in relation to line terminators.
3868
3869 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3870 for the benefit of those environments where the "b" makes a difference.
3871
3872 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3873 to know about it. I have put a hack into configure.in that adds in code
3874 to set GCC=yes if CC=icc. This seems to end up at a point in the
3875 generated configure script that is early enough to affect the setting of
3876 compiler options, which is what is needed, but I have no means of testing
3877 whether it really works. (The user who reported this had patched the
3878 generated configure script, which of course I cannot do.)
3879
3880 LATER: After change 22 below (new libtool files), the configure script
3881 seems to know about icc (and also ecc). Therefore, I have commented out
3882 this hack in configure.in.
3883
3884 7. Added support for pkg-config (2 patches were sent in).
3885
3886 8. Negated POSIX character classes that used a combination of internal tables
3887 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3888 [[:^ascii]]. Typically, they would match almost any characters. The other
3889 POSIX classes were not broken in this way.
3890
3891 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3892 to find the match, as PCRE was deluded into thinking that the match had to
3893 start at the start point or following a newline. The same bug applied to
3894 patterns with negative forward assertions or any backward assertions
3895 preceding ".*" at the start, unless the pattern required a fixed first
3896 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3897
3898 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3899 starting at the last subject character, bytes beyond the end of the subject
3900 string were read.
3901
3902 11. Renamed the variable "class" as "classbits" to make life easier for C++
3903 users. (Previously there was a macro definition, but it apparently wasn't
3904 enough.)
3905
3906 12. Added the new field "tables" to the extra data so that tables can be passed
3907 in at exec time, or the internal tables can be re-selected. This allows
3908 a compiled regex to be saved and re-used at a later time by a different
3909 program that might have everything at different addresses.
3910
3911 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3912 -R library as well as a -L library.
3913
3914 14. The debugging options of pcretest (-d on the command line or D on a
3915 pattern) showed incorrect output for anything following an extended class
3916 that contained multibyte characters and which was followed by a quantifier.
3917
3918 15. Added optional support for general category Unicode character properties
3919 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3920 support. It adds about 90K to the size of the library. The meanings of the
3921 inbuilt class escapes such as \d and \s have NOT been changed.
3922
3923 16. Updated pcredemo.c to include calls to free() to release the memory for the
3924 compiled pattern.
3925
3926 17. The generated file chartables.c was being created in the source directory
3927 instead of in the building directory. This caused the build to fail if the
3928 source directory was different from the building directory, and was
3929 read-only.
3930
3931 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3932 file. No doubt somebody will tell me if they don't make sense... Also added
3933 Dan Mooney's comments about building on OpenVMS.
3934
3935 19. Added support for partial matching via the PCRE_PARTIAL option for
3936 pcre_exec() and the \P data escape in pcretest.
3937
3938 20. Extended pcretest with 3 new pattern features:
3939
3940 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3941 write the compiled pattern to the file whose name is "rest-of-line".
3942 This is a straight binary dump of the data, with the saved pointer to
3943 the character tables forced to be NULL. The study data, if any, is
3944 written too. After writing, pcretest reads a new pattern.
3945
3946 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3947 compiled pattern from the given file. There must not be any
3948 occurrences of "<" in the file name (pretty unlikely); if there are,
3949 pcretest will instead treat the initial "<" as a pattern delimiter.
3950 After reading in the pattern, pcretest goes on to read data lines as
3951 usual.
3952
3953 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3954 and 16-bit fields in a compiled pattern, to simulate a pattern that
3955 was compiled on a host of opposite endianness.
3956
3957 21. The pcre-exec() function can now cope with patterns that were compiled on
3958 hosts of opposite endianness, with this restriction:
3959
3960 As for any compiled expression that is saved and used later, the tables
3961 pointer field cannot be preserved; the extra_data field in the arguments
3962 to pcre_exec() should be used to pass in a tables address if a value
3963 other than the default internal tables were used at compile time.
3964
3965 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3966 now diagnosed as an error. Previously, most of the time, a negative number
3967 would have been treated as zero, but if in addition "ovector" was passed as
3968 NULL, a crash could occur.
3969
3970 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3971 new versions from the libtool 1.5 distribution (the last one is a copy of
3972 a file called libtool.m4). This seems to have fixed the need to patch
3973 "configure" to support Darwin 1.3 (which I used to do). However, I still
3974 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3975 workstation).
3976
3977 24. Changed the PCRE licence to be the more standard "BSD" licence.
3978
3979
3980 Version 4.5 01-Dec-03
3981 ---------------------
3982
3983 1. There has been some re-arrangement of the code for the match() function so
3984 that it can be compiled in a version that does not call itself recursively.
3985 Instead, it keeps those local variables that need separate instances for
3986 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3987 needs to "recurse". Keeping track of where control must go is done by means
3988 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3989 hide most of the details from the main code, and operates only if
3990 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3991 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3992 operating.
3993
3994 To make it easier for callers to provide specially tailored get/free
3995 functions for this usage, two new functions, pcre_stack_malloc, and
3996 pcre_stack_free, are used. They are always called in strict stacking order,
3997 and the size of block requested is always the same.
3998
3999 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4000 PCRE has been compiled to use the stack or the heap for recursion. The
4001 -C option of pcretest uses this to show which version is compiled.
4002
4003 A new data escape \S, is added to pcretest; it causes the amounts of store
4004 obtained and freed by both kinds of malloc/free at match time to be added
4005 to the output.
4006
4007 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4008 what's available on my current Linux desktop machine.
4009
4010 3. When matching a UTF-8 string, the test for a valid string at the start has
4011 been extended. If start_offset is not zero, PCRE now checks that it points
4012 to a byte that is the start of a UTF-8 character. If not, it returns
4013 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4014 this is necessary because there may be backward assertions in the pattern.
4015 When matching the same subject several times, it may save resources to use
4016 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4017
4018 4. The code for checking the validity of UTF-8 strings has been tightened so
4019 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4020 containing "overlong sequences".
4021
4022 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4023 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4024 should have been "&", but it just so happened that all the cases this let
4025 through by mistake were picked up later in the function.
4026
4027 6. I had used a variable called "isblank" - this is a C99 function, causing
4028 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4029
4030 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4031 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4032 so the version is shown; (c) stop "make test" echoing "./RunTest".
4033
4034 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4035
4036 9. The prototype for memmove() for systems that don't have it was using
4037 size_t, but the inclusion of the header that defines size_t was later. I've
4038 moved the #includes for the C headers earlier to avoid this.
4039
4040 10. Added some adjustments to the code to make it easier to compiler on certain
4041 special systems:
4042
4043 (a) Some "const" qualifiers were missing.
4044 (b) Added the macro EXPORT before all exported functions; by default this
4045 is defined to be empty.
4046 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4047 that it reads its output file name as an argument instead of writing
4048 to the standard output and assuming this can be redirected.
4049
4050 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4051 class containing characters with values greater than 255, PCRE compilation
4052 went into a loop.
4053
4054 12. A recursive reference to a subpattern that was within another subpattern
4055 that had a minimum quantifier of zero caused PCRE to crash. For example,
4056 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4057 recursion. If the recursively-called subpattern itself had a zero repeat,
4058 that was OK.
4059
4060 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4061 buffer into which it was copied (for escape processing) was still set at
4062 1024, so long lines caused crashes.
4063
4064 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4065 "internal error: code overflow...". This applied to any character class
4066 that was followed by a possessive quantifier.
4067
4068 15. Modified the Makefile to add libpcre.la as a prerequisite for
4069 libpcreposix.la because I was told this is needed for a parallel build to
4070 work.
4071
4072 16. If a pattern that contained .* following optional items at the start was
4073 studied, the wrong optimizing data was generated, leading to matching
4074 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4075 matching string must start with a or b or c. The correct conclusion for
4076 this pattern is that a match can start with any character.
4077
4078
4079 Version 4.4 13-Aug-03
4080 ---------------------
4081
4082 1. In UTF-8 mode, a character class containing characters with values between
4083 127 and 255 was not handled correctly if the compiled pattern was studied.
4084 In fixing this, I have also improved the studying algorithm for such
4085 classes (slightly).
4086
4087 2. Three internal functions had redundant arguments passed to them. Removal
4088 might give a very teeny performance improvement.
4089
4090 3. Documentation bug: the value of the capture_top field in a callout is *one
4091 more than* the number of the hightest numbered captured substring.
4092
4093 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4094 in incorrectly linking with a previously installed version. They now link
4095 explicitly with libpcre.la.
4096
4097 5. configure.in no longer needs to recognize Cygwin specially.
4098
4099 6. A problem in pcre.in for Windows platforms is fixed.
4100
4101 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4102 pcretest, it used to include the size of the study block as part of its
4103 output. Unfortunately, the structure contains a field that has a different
4104 size on different hardware architectures. This meant that the tests that
4105 showed this size failed. As the block is currently always of a fixed size,
4106 this information isn't actually particularly useful in pcretest output, so
4107 I have just removed it.
4108
4109 8. Three pre-processor statements accidentally did not start in column 1.
4110 Sadly, there are *still* compilers around that complain, even though
4111 standard C has not required this for well over a decade. Sigh.
4112
4113 9. In pcretest, the code for checking callouts passed small integers in the
4114 callout_data field, which is a void * field. However, some picky compilers
4115 complained about the casts involved for this on 64-bit systems. Now
4116 pcretest passes the address of the small integer instead, which should get
4117 rid of the warnings.
4118
4119 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4120 both compile and run time, and gives an error if an invalid UTF-8 sequence
4121 is found. There is a option for disabling this check in cases where the
4122 string is known to be correct and/or the maximum performance is wanted.
4123
4124 11. In response to a bug report, I changed one line in Makefile.in from
4125
4126 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4127 to
4128 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4129
4130 to look similar to other lines, but I have no way of telling whether this
4131 is the right thing to do, as I do not use Windows. No doubt I'll get told
4132 if it's wrong...
4133
4134
4135 Version 4.3 21-May-03
4136 ---------------------
4137
4138 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4139 Makefile.
4140
4141 2. Some refactoring to improve the quality of the code:
4142
4143 (i) The utf8_table... variables are now declared "const".
4144
4145 (ii) The code for \cx, which used the "case flipping" table to upper case
4146 lower case letters, now just substracts 32. This is ASCII-specific,
4147 but the whole concept of \cx is ASCII-specific, so it seems
4148 reasonable.
4149
4150 (iii) PCRE was using its character types table to recognize decimal and
4151 hexadecimal digits in the pattern. This is silly, because it handles
4152 only 0-9, a-f, and A-F, but the character types table is locale-
4153 specific, which means strange things might happen. A private
4154 table is now used for this - though it costs 256 bytes, a table is
4155 much faster than multiple explicit tests. Of course, the standard
4156 character types table is still used for matching digits in subject
4157 strings against \d.
4158
4159 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4160 ending in _t are). So I've renamed it as ESC_tee.
4161
4162 3. The first argument for regexec() in the POSIX wrapper should have been
4163 defined as "const".
4164
4165 4. Changed pcretest to use malloc() for its buffers so that they can be
4166 Electric Fenced for debugging.
4167
4168 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4169 to read one or more bytes before the start of the subject string. Often this
4170 had no effect on PCRE's behaviour, but in some circumstances it could
4171 provoke a segmentation fault.
4172
4173 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4174 to try to read one or more bytes before the start of the subject string.
4175
4176 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4177 UTF-8 support could misbehave in various ways if the subject string
4178 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4179 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4180 back over UTF-8 characters.)
4181
4182
4183 Version 4.2 14-Apr-03
4184 ---------------------
4185
4186 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4187
4188 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4189 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4190 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4191 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4192 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4193 and BUILD_EXEEXT
4194 Note: automatic setting of the BUILD variables is not yet working
4195 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4196 compile-time but not at link-time
4197 [LINK]: use for linking executables only
4198 make different versions for Windows and non-Windows
4199 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4200 libraries
4201 [LINK_FOR_BUILD]: new variable
4202 [OBJEXT]: use throughout
4203 [EXEEXT]: use throughout
4204 <winshared>: new target
4205 <wininstall>: new target
4206 <dftables.o>: use native compiler
4207 <dftables>: use native linker
4208 <install>: handle Windows platform correctly
4209 <clean>: ditto
4210 <check>: ditto
4211 copy DLL to top builddir before testing
4212
4213 As part of these changes, -no-undefined was removed again. This was reported
4214 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4215 in any case.
4216
4217 3. Some tidies to get rid of compiler warnings:
4218
4219 . In the match_data structure, match_limit was an unsigned long int, whereas
4220 match_call_count was an int. I've made them both unsigned long ints.
4221
4222 . In pcretest the fact that a const uschar * doesn't automatically cast to
4223 a void * provoked a warning.
4224
4225 . Turning on some more compiler warnings threw up some "shadow" variables
4226 and a few more missing casts.
4227
4228 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4229 option, a class that contained a single character with a value between 128
4230 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4231
4232 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4233 option, a class that contained several characters, but with at least one
4234 whose value was between 128 and 255 caused PCRE to crash.
4235
4236
4237 Version 4.1 12-Mar-03
4238 ---------------------
4239
4240 1. Compiling with gcc -pedantic found a couple of places where casts were
4241 needed, and a string in dftables.c that was longer than standard compilers are
4242 required to support.
4243
4244 2. Compiling with Sun's compiler found a few more places where the code could
4245 be tidied up in order to avoid warnings.
4246
4247 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4248 first of these names is deprecated in the latest Autoconf in favour of the name
4249 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4250 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4251 analogy I have changed it to CFLAGS_FOR_BUILD.
4252
4253 4. Added -no-undefined to the linking command in the Makefile, because this is
4254 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4255 linking step for the pcreposix library.
4256
4257 5. PCRE was failing to diagnose the case of two named groups with the same
4258 name.
4259
4260 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4261 literal character that is needed in the subject for a match, and scans along to
4262 ensure that it is present before embarking on the full matching process. This
4263 saves time in cases of nested unlimited repeats that are never going to match.
4264 Problem: the scan can take a lot of time if the subject is very long (e.g.
4265 megabytes), thus penalizing straightforward matches. It is now done only if the
4266 amount of subject to be scanned is less than 1000 bytes.
4267
4268 7. A lesser problem with the same optimization is that it was recording the
4269 first character of an anchored pattern as "needed", thus provoking a search
4270 right along the subject, even when the first match of the pattern was going to
4271 fail. The "needed" character is now not set for anchored patterns, unless it
4272 follows something in the pattern that is of non-fixed length. Thus, it still
4273 fulfils its original purpose of finding quick non-matches in cases of nested
4274 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4275
4276
4277 Version 4.0 17-Feb-03
4278 ---------------------
4279
4280 1. If a comment in an extended regex that started immediately after a meta-item
4281 extended to the end of string, PCRE compiled incorrect data. This could lead to
4282 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4283
4284 2. Moved to autoconf 2.53 and libtool 1.4.2.
4285
4286 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4287 the special perltest8 script is no longer needed - all the tests can be run
4288 from a single perltest script.
4289
4290 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4291 by \s. It has now been removed in PCRE. This means it isn't recognized as
4292 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4293 class [:space:] *does* include VT, thereby creating a mess.
4294
4295 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4296 space and tab.
4297
4298 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4299 its new features into the main test script, reducing the number of scripts.
4300
4301 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4302 were backward compatible, and made the (?i) apply to the whole pattern, as if
4303 /i were given. Now it behaves more logically, and applies the option setting
4304 only to what follows. PCRE has been changed to follow suit. However, if it
4305 finds options settings right at the start of the pattern, it extracts them into
4306 the global options, as before. Thus, they show up in the info data.
4307
4308 8. Added support for the \Q...\E escape sequence. Characters in between are
4309 treated as literals. This is slightly different from Perl in that $ and @ are
4310 also handled as literals inside the quotes. In Perl, they will cause variable
4311 interpolation. Note the following examples:
4312
4313 Pattern PCRE matches Perl matches
4314
4315 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4316 \Qabc\$xyz\E abc\$xyz abc\$xyz
4317 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4318
4319 For compatibility with Perl, \Q...\E sequences are recognized inside character
4320 classes as well as outside them.
4321
4322 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4323 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4324 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4325 signed/unsigned warnings.
4326
4327 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4328 option for pcretest, so I've replaced it by a simple function that does just
4329 that job.
4330
4331 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4332 "pcregrep -".
4333
4334 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4335 Java package. This provides some syntactic sugar for simple cases of what my
4336 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4337 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4338 item, you can use this simplified notation. Note that only makes sense with
4339 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4340 greediness, whatever the setting of the PCRE_UNGREEDY option.
4341
4342 13. A change of greediness default within a pattern was not taking effect at
4343 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4344 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4345 was abstracted outside.
4346
4347 14. PCRE now supports the \G assertion. It is true when the current matching
4348 position is at the start point of the match. This differs from \A when the
4349 starting offset is non-zero. Used with the /g option of pcretest (or similar
4350 code), it works in the same way as it does for Perl's /g option. If all
4351 alternatives of a regex begin with \G, the expression is anchored to the start
4352 match position, and the "anchored" flag is set in the compiled expression.
4353
4354 15. Some bugs concerning the handling of certain option changes within patterns
4355 have been fixed. These applied to options other than (?ims). For example,
4356 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4357 been the other way round. Some of this was related to change 7 above.
4358
4359 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4360 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4361 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4362 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4363
4364 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4365 mode. Unlike ".", it always matches newline, whatever the setting of
4366 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4367 assertions. Perl allows it, but it doesn't (in general) work because it can't
4368 calculate the length of the lookbehind. At least, that's the case for Perl
4369 5.8.0 - I've been told they are going to document that it doesn't work in
4370 future.
4371
4372 18. Added an error diagnosis for escapes that PCRE does not support: these are
4373 \L, \l, \N, \P, \p, \U, \u, and \X.
4374
4375 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4376 reading past the end of the pattern in cases such as /[abcd/.
4377
4378 20. PCRE was getting more memory than necessary for patterns with classes that
4379 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4380
4381 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4382 compiling PCRE for use with Virtual Pascal.
4383
4384 22. Small fix to the Makefile to make it work properly if the build is done
4385 outside the source tree.
4386
4387 23. Added a new extension: a condition to go with recursion. If a conditional
4388 subpattern starts with (?(R) the "true" branch is used if recursion has
4389 happened, whereas the "false" branch is used only at the top level.
4390
4391 24. When there was a very long string of literal characters (over 255 bytes
4392 without UTF support, over 250 bytes with UTF support), the computation of how
4393 much memory was required could be incorrect, leading to segfaults or other
4394 strange effects.
4395
4396 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4397 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4398 there was a subsequent back reference to those brackets. This meant that, for
4399 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4400 possible to check for precisely this case. All we can do is abandon the
4401 optimization if .* occurs inside capturing brackets when there are any back
4402 references whatsoever. (See below for a better fix that came later.)
4403
4404 26. The handling of the optimization for finding the first character of a
4405 non-anchored pattern, and for finding a character that is required later in the
4406 match were failing in some cases. This didn't break the matching; it just
4407 failed to optimize when it could. The way this is done has been re-implemented.
4408
4409 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4410
4411 28. Added a new feature that provides some of the functionality that Perl
4412 provides with (?{...}). The facility is termed a "callout". The way it is done
4413 in PCRE is for the caller to provide an optional function, by setting
4414 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4415 global variable. By default it is unset, which disables all calling out. To get
4416 the function called, the regex must include (?C) at appropriate points. This
4417 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4418 This provides a means of identifying different callout points. When PCRE
4419 reaches such a point in the regex, if pcre_callout has been set, the external
4420 function is called. It is provided with data in a structure called
4421 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4422 matching continues; if it returns a non-zero value, the match at the current
4423 point fails. However, backtracking will occur if possible. [This was changed
4424 later and other features added - see item 49 below.]
4425
4426 29. pcretest is upgraded to test the callout functionality. It provides a
4427 callout function that displays information. By default, it shows the start of
4428 the match and the current position in the text. There are some new data escapes
4429 to vary what happens:
4430
4431 \C+ in addition, show current contents of captured substrings
4432 \C- do not supply a callout function
4433 \C!n return 1 when callout number n is reached
4434 \C!n!m return 1 when callout number n is reached for the mth time
4435
4436 30. If pcregrep was called with the -l option and just a single file name, it
4437 output "<stdin>" if a match was found, instead of the file name.
4438
4439 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4440 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4441 pcre_exec(). This saves a malloc/free per call. The default value of
4442 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4443 when configuring.
4444
4445 32. The default maximum size of a compiled pattern is 64K. There have been a
4446 few cases of people hitting this limit. The code now uses macros to handle the
4447 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4448 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4449 configuring. Tests 2 and 5 work only with 2-byte links because they output
4450 debugging information about compiled patterns.
4451
4452 33. Internal code re-arrangements:
4453
4454 (a) Moved the debugging function for printing out a compiled regex into
4455 its own source file (printint.c) and used #include to pull it into
4456 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4457 separate copies.
4458
4459 (b) Defined the list of op-code names for debugging as a macro in
4460 internal.h so that it is next to the definition of the opcodes.
4461
4462 (c) Defined a table of op-code lengths for simpler skipping along compiled
4463 code. This is again a macro in internal.h so that it is next to the
4464 definition of the opcodes.
4465
4466 34. Added support for recursive calls to individual subpatterns, along the
4467 lines of Robin Houston's patch (but implemented somewhat differently).
4468
4469 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4470 allow it to read and process whole directories in Win32. This code was
4471 contributed by Lionel Fourquaux; it has not been tested by me.
4472
4473 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4474 used to name a group. Names consist of alphanumerics and underscores, and must
4475 be unique. Back references use the syntax (?P=name) and recursive calls use
4476 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4477 numbers. The function pcre_fullinfo() can be used after compilation to extract
4478 a name/number map. There are three relevant calls:
4479
4480 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4481 PCRE_INFO_NAMECOUNT yields the number of entries
4482 PCRE_INFO_NAMETABLE yields a pointer to the map.
4483
4484 The map is a vector of fixed-size entries. The size of each entry depends on
4485 the length of the longest name used. The first two bytes of each entry are the
4486 group number, most significant byte first. There follows the corresponding
4487 name, zero terminated. The names are in alphabetical order.
4488
4489 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4490 case instead of 255. Making it the same both with and without UTF-8 support
4491 means that the same test output works with both.
4492
4493 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4494 calling malloc() with a zero argument.
4495
4496 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4497 optimization. I've improved things by keeping a bitmap of backreferences with
4498 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4499 fact referenced, the optimization can be applied. It is unlikely that a
4500 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4501 the match to follow \n) will appear inside brackets with a number greater than
4502 31, but if it does, any back reference > 31 suppresses the optimization.
4503
4504 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4505 of disabling numbered capturing parentheses. Any opening parenthesis that is
4506 not followed by ? behaves as if it were followed by ?: but named parentheses
4507 can still be used for capturing (and they will acquire numbers in the usual
4508 way).
4509
4510 41. Redesigned the return codes from the match() function into yes/no/error so
4511 that errors can be passed back from deep inside the nested calls. A malloc
4512 failure while inside a recursive subpattern call now causes the
4513 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4514
4515 42. It is now possible to set a limit on the number of times the match()
4516 function is called in a call to pcre_exec(). This facility makes it possible to
4517 limit the amount of recursion and backtracking, though not in a directly
4518 obvious way, because the match() function is used in a number of different
4519 circumstances. The count starts from zero for each position in the subject
4520 string (for non-anchored patterns). The default limit is, for compatibility, a
4521 large number, namely 10 000 000. You can change this in two ways:
4522
4523 (a) When configuring PCRE before making, you can use --with-match-limit=n
4524 to set a default value for the compiled library.
4525
4526 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4527 a different value is set. See 45 below.
4528
4529 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4530
4531 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4532 of things that can be changed at compile time. The first argument specifies
4533 what is wanted and the second points to where the information is to be placed.
4534 The current list of available information is:
4535
4536 PCRE_CONFIG_UTF8
4537
4538 The output is an integer that is set to one if UTF-8 support is available;
4539 otherwise it is set to zero.
4540
4541 PCRE_CONFIG_NEWLINE
4542
4543 The output is an integer that it set to the value of the code that is used for
4544 newline. It is either LF (10) or CR (13).
4545
4546 PCRE_CONFIG_LINK_SIZE
4547
4548 The output is an integer that contains the number of bytes used for internal
4549 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4550
4551 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4552
4553 The output is an integer that contains the threshold above which the POSIX
4554 interface uses malloc() for output vectors. See item 31 above.
4555
4556 PCRE_CONFIG_MATCH_LIMIT
4557
4558 The output is an unsigned integer that contains the default limit of the number
4559 of match() calls in a pcre_exec() execution. See 42 above.
4560
4561 44. pcretest has been upgraded by the addition of the -C option. This causes it
4562 to extract all the available output from the new pcre_config() function, and to
4563 output it. The program then exits immediately.
4564
4565 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4566 order to support additional features. One way would have been to define
4567 pcre_exec2() (for example) with extra arguments, but this would not have been
4568 extensible, and would also have required all calls to the original function to
4569 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4570 is used for passing in "extra" data from pcre_study().
4571
4572 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4573 contains the following fields:
4574
4575 flags a bitmap indicating which of the following fields are set
4576 study_data opaque data from pcre_study()
4577 match_limit a way of specifying a limit on match() calls for a specific
4578 call to pcre_exec()
4579 callout_data data for callouts (see 49 below)
4580
4581 The flag bits are also defined in pcre.h, and are
4582
4583 PCRE_EXTRA_STUDY_DATA
4584 PCRE_EXTRA_MATCH_LIMIT
4585 PCRE_EXTRA_CALLOUT_DATA
4586
4587 The pcre_study() function now returns one of these new pcre_extra blocks, with
4588 the actual study data pointed to by the study_data field, and the
4589 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4590 before. That is, this change is entirely upwards-compatible and requires no
4591 change to existing code.
4592
4593 If you want to pass in additional data to pcre_exec(), you can either place it
4594 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4595 block.
4596
4597 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4598 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4599 times with different match limits, until it finds the minimum value needed for
4600 pcre_exec() to complete. The value is then output. This can be instructive; for
4601 most simple matches the number is quite small, but for pathological cases it
4602 gets very large very quickly.
4603
4604 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4605 returns the size of the data block pointed to by the study_data field in a
4606 pcre_extra block, that is, the value that was passed as the argument to
4607 pcre_malloc() when PCRE was getting memory in which to place the information
4608 created by pcre_study(). The fourth argument should point to a size_t variable.
4609 pcretest has been extended so that this information is shown after a successful
4610 pcre_study() call when information about the compiled regex is being displayed.
4611
4612 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4613 because what follows is always an absolute path. (Later: it turns out that this
4614 is more than cosmetic for MinGW, because it doesn't like empty path
4615 components.)
4616
4617 49. Some changes have been made to the callout feature (see 28 above):
4618
4619 (i) A callout function now has three choices for what it returns:
4620
4621 0 => success, carry on matching
4622 > 0 => failure at this point, but backtrack if possible
4623 < 0 => serious error, return this value from pcre_exec()
4624
4625 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4626 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4627 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4628 use by callout functions. It will never be used by PCRE itself.
4629
4630 (ii) The pcre_extra structure (see 45 above) has a void * field called
4631 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4632 pcre_callout_block structure has a field of the same name. The contents of
4633 the field passed in the pcre_extra structure are passed to the callout
4634 function in the corresponding field in the callout block. This makes it
4635 easier to use the same callout-containing regex from multiple threads. For
4636 testing, the pcretest program has a new data escape
4637
4638 \C*n pass the number n (may be negative) as callout_data
4639
4640 If the callout function in pcretest receives a non-zero value as
4641 callout_data, it returns that value.
4642
4643 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4644 there were some redundant $(CFLAGS) in commands that are now specified as
4645 $(LINK), which already includes $(CFLAGS).
4646
4647 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4648 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4649 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4650 one-byte characters throughout. Note that case-insensitive matching applies
4651 only to characters whose values are less than 256. PCRE doesn't support the
4652 notion of cases for higher-valued characters.
4653
4654 (i) A character class whose characters are all within 0-255 is handled as
4655 a bit map, and the map is inverted for negative classes. Previously, a
4656 character > 255 always failed to match such a class; however it should
4657 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4658
4659 (ii) A negated character class with a single character < 255 is coded as
4660 "not this character" (OP_NOT). This wasn't working properly when the test
4661 character was multibyte, either singly or repeated.
4662
4663 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4664 mode, for example: \x{100}{2,3}.
4665
4666 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4667 singly or repeated) now correctly test multibyte characters. However,
4668 PCRE doesn't recognize any characters with values greater than 255 as
4669 digits, spaces, or word characters. Such characters always match \D, \S,
4670 and \W, and never match \d, \s, or \w.
4671
4672 (v) Classes may now contain characters and character ranges with values
4673 greater than 255. For example: [ab\x{100}-\x{400}].
4674
4675 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4676 PCRE in UTF-8 mode.
4677
4678 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4679 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4680 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4681 value.)
4682
4683 53. The single man page has become too large. I have therefore split it up into
4684 a number of separate man pages. These also give rise to individual HTML pages;
4685 these are now put in a separate directory, and there is an index.html page that
4686 lists them all. Some hyperlinking between the pages has been installed.
4687
4688 54. Added convenience functions for handling named capturing parentheses.
4689
4690 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4691 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4692 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4693 are faulted.
4694
4695 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4696 calling configure. These values are used when compiling the dftables.c program
4697 which is run to generate the source of the default character tables. They
4698 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4699 you will need to set these values.
4700
4701 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4702
4703
4704 Version 3.9 02-Jan-02
4705 ---------------------
4706
4707 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4708
4709 2. If --disable-static was given, the building process failed when trying to
4710 build pcretest and pcregrep. (For some reason it was using libtool to compile
4711 them, which is not right, as they aren't part of the library.)
4712
4713
4714 Version 3.8 18-Dec-01
4715 ---------------------
4716
4717 1. The experimental UTF-8 code was completely screwed up. It was packing the
4718 bytes in the wrong order. How dumb can you get?
4719
4720
4721 Version 3.7 29-Oct-01
4722 ---------------------
4723
4724 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4725 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4726 this didn't happen under Solaris 8, where I normally test things.
4727
4728 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4729 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4730 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4731 this fix an hour or so after the initial 3.7 release.)
4732
4733
4734 Version 3.6 23-Oct-01
4735 ---------------------
4736
4737 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4738 offsets passed as NULL with zero offset count.
4739
4740 2. The config.guess and config.sub files had not been updated when I moved to
4741 the latest autoconf.
4742
4743
4744 Version 3.5 15-Aug-01
4745 ---------------------
4746
4747 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4748 had been forgotten.
4749
4750 2. By using declared but undefined structures, we can avoid using "void"
4751 definitions in pcre.h while keeping the internal definitions of the structures
4752 private.
4753
4754 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4755 user point of view, this means that both static and shared libraries are built
4756 by default, but this can be individually controlled. More of the work of
4757 handling this static/shared cases is now inside libtool instead of PCRE's make
4758 file.
4759
4760 4. The pcretest utility is now installed along with pcregrep because it is
4761 useful for users (to test regexs) and by doing this, it automatically gets
4762 relinked by libtool. The documentation has been turned into a man page, so
4763 there are now .1, .txt, and .html versions in /doc.
4764
4765 5. Upgrades to pcregrep:
4766 (i) Added long-form option names like gnu grep.
4767 (ii) Added --help to list all options with an explanatory phrase.
4768 (iii) Added -r, --recursive to recurse into sub-directories.
4769 (iv) Added -f, --file to read patterns from a file.
4770
4771 6. pcre_exec() was referring to its "code" argument before testing that
4772 argument for NULL (and giving an error if it was NULL).
4773
4774 7. Upgraded Makefile.in to allow for compiling in a different directory from
4775 the source directory.
4776
4777 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4778 options bits, the pointer it was passed was to an int instead of to an unsigned
4779 long int. This mattered only on 64-bit systems.
4780
4781 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4782 generated) instead of pcre.in, which it its source. Also made the same change
4783 in several of the .c files.
4784
4785 10. A new release of gcc defines printf() as a macro, which broke pcretest
4786 because it had an ifdef in the middle of a string argument for printf(). Fixed
4787 by using separate calls to printf().
4788
4789 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4790 script, to force use of CR or LF instead of \n in the source. On non-Unix
4791 systems, the value can be set in config.h.
4792
4793 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4794 absolute limit. Changed the text of the error message to make this clear, and
4795 likewise updated the man page.
4796
4797 13. The limit of 99 on the number of capturing subpatterns has been removed.
4798 The new limit is 65535, which I hope will not be a "real" limit.
4799
4800
4801 Version 3.4 22-Aug-00
4802 ---------------------
4803
4804 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4805
4806 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4807
4808
4809 Version 3.3 01-Aug-00
4810 ---------------------
4811
4812 1. If an octal character was given, but the value was greater than \377, it
4813 was not getting masked to the least significant bits, as documented. This could
4814 lead to crashes in some systems.
4815
4816 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4817 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4818
4819 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4820 These just pass their arguments on to (pcre_free)(), but they are provided
4821 because some uses of PCRE bind it to non-C systems that can call its functions,
4822 but cannot call free() or pcre_free() directly.
4823
4824 4. Add "make test" as a synonym for "make check". Corrected some comments in
4825 the Makefile.
4826
4827 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4828 Makefile.
4829
4830 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4831 command called pgrep for grepping around the active processes.
4832
4833 7. Added the beginnings of support for UTF-8 character strings.
4834
4835 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4836 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4837 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4838 out for the ar command.)
4839
4840
4841 Version 3.2 12-May-00
4842 ---------------------
4843
4844 This is purely a bug fixing release.
4845
4846 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4847 of ZA. This was just one example of several cases that could provoke this bug,
4848 which was introduced by change 9 of version 2.00. The code for breaking
4849 infinite loops after an iteration that matches an empty string was't working
4850 correctly.
4851
4852 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4853 when matched against abbab (for example). After matching an empty string, it
4854 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4855 caused it to match further down the string than it should.
4856
4857 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4858 was there because it doesn't seem to be needed, and it causes trouble on some
4859 systems, as it is not a Standard C header. It has been removed.
4860
4861 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4862 were reported on the Macintosh. The changes were from
4863
4864 while ((c = *(++ptr)) != 0 && c != '\n');
4865 to
4866 while ((c = *(++ptr)) != 0 && c != '\n') ;
4867
4868 Totally extraordinary, but if that's what it takes...
4869
4870 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4871 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4872 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4873 assumes the way PCRE uses memmove() (always moving upwards).
4874
4875 6. PCRE is being used in one environment where strchr() is not available. There
4876 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4877 faster code anyway.
4878
4879
4880 Version 3.1 09-Feb-00
4881 ---------------------
4882
4883 The only change in this release is the fixing of some bugs in Makefile.in for
4884 the "install" target:
4885
4886 (1) It was failing to install pcreposix.h.
4887
4888 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4889
4890
4891 Version 3.0 01-Feb-00
4892 ---------------------
4893
4894 1. Add support for the /+ modifier to perltest (to output $` like it does in
4895 pcretest).
4896
4897 2. Add support for the /g modifier to perltest.
4898
4899 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4900 matches null strings.
4901
4902 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4903 pattern. Perl treats empty patterns specially - it reuses the most recent
4904 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4905 effect.
4906
4907 5. The POSIX interface was broken in that it was just handing over the POSIX
4908 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4909 required a bigger vector, with some working space on the end. This means that
4910 the POSIX wrapper now has to get and free some memory, and copy the results.
4911
4912 6. Added some simple autoconf support, placing the test data and the
4913 documentation in separate directories, re-organizing some of the
4914 information files, and making it build pcre-config (a GNU standard). Also added
4915 libtool support for building PCRE as a shared library, which is now the
4916 default.
4917
4918 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4919 09 are not valid octal constants. Single digits will be used for minor values
4920 less than 10.
4921
4922 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4923 existing programs that set these in the POSIX interface can use PCRE without
4924 modification.
4925
4926 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4927 return all that pcre_info() returns, plus additional data. The pcre_info()
4928 function is retained for compatibility, but is considered to be obsolete.
4929
4930 10. Added experimental recursion feature (?R) to handle one common case that
4931 Perl 5.6 will be able to do with (?p{...}).
4932
4933 11. Added support for POSIX character classes like [:alpha:], which Perl is
4934 adopting.
4935
4936
4937 Version 2.08 31-Aug-99
4938 ----------------------
4939
4940 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4941 trying to match at the startoffset position, but instead was moving forward to
4942 the next newline as if a previous match had failed.
4943
4944 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4945 and could get into a loop if a null string was matched other than at the start
4946 of the subject.
4947
4948 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4949 be distinguished at compile time, and for completeness also added PCRE_DATE.
4950
4951 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4952 in GnuWin32 environments.
4953
4954
4955 Version 2.07 29-Jul-99
4956 ----------------------
4957
4958 1. The documentation is now supplied in plain text form and HTML as well as in
4959 the form of man page sources.
4960
4961 2. C++ compilers don't like assigning (void *) values to other pointer types.
4962 In particular this affects malloc(). Although there is no problem in Standard
4963 C, I've put in casts to keep C++ compilers happy.
4964
4965 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4966 should be (const char *).
4967
4968 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4969 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4970 However, I haven't made this a standard facility. The documentation doesn't
4971 mention it, and the Makefile doesn't support it.
4972
4973 5. The Makefile now contains an "install" target, with editable destinations at
4974 the top of the file. The pcretest program is not installed.
4975
4976 6. pgrep -V now gives the PCRE version number and date.
4977
4978 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4979 causing the entire string to be ignored, instead of just the last character.
4980
4981 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4982 non-matching string, it can take a very, very long time, even for strings of
4983 quite modest length, because of the nested recursion. PCRE now does better in
4984 some of these cases. It does this by remembering the last required literal
4985 character in the pattern, and pre-searching the subject to ensure it is present
4986 before running the real match. In other words, it applies a heuristic to detect
4987 some types of certain failure quickly, and in the above example, if presented
4988 with a string that has no trailing " it gives "no match" very quickly.
4989
4990 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4991 other alternatives are tried instead.
4992
4993
4994 Version 2.06 09-Jun-99
4995 ----------------------
4996
4997 1. Change pcretest's output for amount of store used to show just the code
4998 space, because the remainder (the data block) varies in size between 32-bit and
4999 64-bit systems.
5000
5001 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5002 start matching at. This allows lookbehinds to work when searching for multiple
5003 occurrences in a string.
5004
5005 3. Added additional options to pcretest for testing multiple occurrences:
5006
5007 /+ outputs the rest of the string that follows a match
5008 /g loops for multiple occurrences, using the new startoffset argument
5009 /G loops for multiple occurrences by passing an incremented pointer
5010
5011 4. PCRE wasn't doing the "first character" optimization for patterns starting
5012 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5013 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5014 the letter 'x'. On long subject strings, this gives a significant speed-up.
5015
5016
5017 Version 2.05 21-Apr-99
5018 ----------------------
5019
5020 1. Changed the type of magic_number from int to long int so that it works
5021 properly on 16-bit systems.
5022
5023 2. Fixed a bug which caused patterns starting with .* not to work correctly
5024 when the subject string contained newline characters. PCRE was assuming
5025 anchoring for such patterns in all cases, which is not correct because .* will
5026 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5027 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5028 must be retried after every newline in the subject.
5029
5030
5031 Version 2.04 18-Feb-99
5032 ----------------------
5033
5034 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5035 computation of the store needed to hold the pattern was incorrect (too large).
5036 If such patterns were nested a few deep, this could multiply and become a real
5037 problem.
5038
5039 2. Added /M option to pcretest to show the memory requirement of a specific
5040 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5041
5042 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5043 compiled in such a way that the backtracking after subsequent failure was
5044 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5045 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5046
5047
5048 Version 2.03 02-Feb-99
5049 ----------------------
5050
5051 1. Fixed typo and small mistake in man page.
5052
5053 2. Added 4th condition (GPL supersedes if conflict) and created separate
5054 LICENCE file containing the conditions.
5055
5056 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5057 Perl, that is the internal \ allows the delimiter to be included in the
5058 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5059 the final delimiter, add \ to the end of the pattern (to test the error).
5060
5061 4. Added the convenience functions for extracting substrings after a successful
5062 match. Updated pcretest to make it able to test these functions.
5063
5064
5065 Version 2.02 14-Jan-99
5066 ----------------------
5067
5068 1. Initialized the working variables associated with each extraction so that
5069 their saving and restoring doesn't refer to uninitialized store.
5070
5071 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5072 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5073 fix the problem.
5074
5075 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5076 calls, and wasn't printing the correct value for compiling calls. Increased the
5077 default value of LOOPREPEAT, and the number of significant figures in the
5078 times.
5079
5080 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5081
5082 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5083 a building problem on Windows NT with a FAT file system.
5084
5085
5086 Version 2.01 21-Oct-98
5087 ----------------------
5088
5089 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5090 to character tables built by pcre_maketables() in the current locale. If NULL
5091 is passed, the default tables are used.
5092
5093
5094 Version 2.00 24-Sep-98
5095 ----------------------
5096
5097 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5098 it any more.
5099
5100 2. Allow quantification of (?>) groups, and make it work correctly.
5101
5102 3. The first character computation wasn't working for (?>) groups.
5103
5104 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5105 end of the subject) and add 5.005's \z, which really does match only at the
5106 very end of the subject.
5107
5108 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5109
5110 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5111 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5112 localized options. All options to pcre_study() were also removed.
5113
5114 7. Add other new features from 5.005:
5115
5116 $(?<= positive lookbehind
5117 $(?<! negative lookbehind
5118 (?imsx-imsx) added the unsetting capability
5119 such a setting is global if at outer level; local otherwise
5120 (?imsx-imsx:) non-capturing groups with option setting
5121 (?(cond)re|re) conditional pattern matching
5122
5123 A backreference to itself in a repeated group matches the previous
5124 captured string.
5125
5126 8. General tidying up of studying (both automatic and via "study")
5127 consequential on the addition of new assertions.
5128
5129 9. As in 5.005, unlimited repeated groups that could match an empty substring
5130 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5131 runtime if any iteration does actually match an empty substring.
5132
5133 10. Include the RunTest script in the distribution.
5134
5135 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5136 discrepancies, some of which were old and were also with respect to 5.004. They
5137 have now been fixed.
5138
5139
5140 Version 1.09 28-Apr-98
5141 ----------------------
5142
5143 1. A negated single character class followed by a quantifier with a minimum
5144 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5145 program crashes, or just wrong answers. This did not apply to negated classes
5146 containing more than one character, or to minima other than one.
5147
5148
5149 Version 1.08 27-Mar-98
5150 ----------------------
5151
5152 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5153
5154 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5155 latter must appear before anything that relies on it in the pattern.
5156
5157
5158 Version 1.07 16-Feb-98
5159 ----------------------
5160
5161 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5162 repeat of a potentially empty string).
5163
5164
5165 Version 1.06 23-Jan-98
5166 ----------------------
5167
5168 1. Added Markus Oberhumer's little patches for C++.
5169
5170 2. Literal strings longer than 255 characters were broken.
5171
5172
5173 Version 1.05 23-Dec-97
5174 ----------------------
5175
5176 1. Negated character classes containing more than one character were failing if
5177 PCRE_CASELESS was set at run time.
5178
5179
5180 Version 1.04 19-Dec-97
5181 ----------------------
5182
5183 1. Corrected the man page, where some "const" qualifiers had been omitted.
5184
5185 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5186 input syntax.
5187
5188 3. Fixed memory leak which occurred when a regex with back references was
5189 matched with an offsets vector that wasn't big enough. The temporary memory
5190 that is used in this case wasn't being freed if the match failed.
5191
5192 4. Tidied pcretest to ensure it frees memory that it gets.
5193
5194 5. Temporary memory was being obtained in the case where the passed offsets
5195 vector was exactly big enough.
5196
5197 6. Corrected definition of offsetof() from change 5 below.
5198
5199 7. I had screwed up change 6 below and broken the rules for the use of
5200 setjmp(). Now fixed.
5201
5202
5203 Version 1.03 18-Dec-97
5204 ----------------------
5205
5206 1. A erroneous regex with a missing opening parenthesis was correctly
5207 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5208 on some systems.
5209
5210 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5211 it was reported that one broken compiler failed on the former because "code" is
5212 also an independent variable.
5213
5214 3. The erroneous regex a[]b caused an array overrun reference.
5215
5216 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5217 fail on data ending with that character. (It was going on too far, and checking
5218 the next character, typically a binary zero.) This was specific to the
5219 optimized code for single-character negative classes.
5220
5221 5. Added a contributed patch from the TIN world which does the following:
5222
5223 + Add an undef for memmove, in case the the system defines a macro for it.
5224
5225 + Add a definition of offsetof(), in case there isn't one. (I don't know
5226 the reason behind this - offsetof() is part of the ANSI standard - but
5227 it does no harm).
5228
5229 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5230 most of the places where whitespace preceded '#'. I have given up and
5231 allowed the remaining 2 cases to be at the margin.
5232
5233 + Rename some variables in pcre to eliminate shadowing. This seems very
5234 pedantic, but does no harm, of course.
5235
5236 6. Moved the call to setjmp() into its own function, to get rid of warnings
5237 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5238
5239 7. Constructs such as \d{8,} were compiling into the equivalent of
5240 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5241 outcome, but in this particular case used more store than had been allocated,
5242 which caused the bug to be discovered because it threw up an internal error.
5243
5244 8. The debugging code in both pcre and pcretest for outputting the compiled
5245 form of a regex was going wrong in the case of back references followed by
5246 curly-bracketed repeats.
5247
5248
5249 Version 1.02 12-Dec-97
5250 ----------------------
5251
5252 1. Typos in pcre.3 and comments in the source fixed.
5253
5254 2. Applied a contributed patch to get rid of places where it used to remove
5255 'const' from variables, and fixed some signed/unsigned and uninitialized
5256 variable warnings.
5257
5258 3. Added the "runtest" target to Makefile.
5259
5260 4. Set default compiler flag to -O2 rather than just -O.
5261
5262
5263 Version 1.01 19-Nov-97
5264 ----------------------
5265
5266 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5267 like /([ab]*)*/, that is, for classes with more than one character in them.
5268
5269 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5270 as /((?>a*))*/ (a PCRE_EXTRA facility).
5271
5272
5273 Version 1.00 18-Nov-97
5274 ----------------------
5275
5276 1. Added compile-time macros to support systems such as SunOS4 which don't have
5277 memmove() or strerror() but have other things that can be used instead.
5278
5279 2. Arranged that "make clean" removes the executables.
5280
5281
5282 Version 0.99 27-Oct-97
5283 ----------------------
5284
5285 1. Fixed bug in code for optimizing classes with only one character. It was
5286 initializing a 32-byte map regardless, which could cause it to run off the end
5287 of the memory it had got.
5288
5289 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5290
5291
5292 Version 0.98 22-Oct-97
5293 ----------------------
5294
5295 1. Fixed bug in code for handling temporary memory usage when there are more
5296 back references than supplied space in the ovector. This could cause segfaults.
5297
5298
5299 Version 0.97 21-Oct-97
5300 ----------------------
5301
5302 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5303
5304 2. Optimized negated single characters not to use a bit map.
5305
5306 3. Brought error texts together as macro definitions; clarified some of them;
5307 fixed one that was wrong - it said "range out of order" when it meant "invalid
5308 escape sequence".
5309
5310 4. Changed some char * arguments to const char *.
5311
5312 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5313
5314 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5315 pcretest.
5316
5317
5318 Version 0.96 16-Oct-97
5319 ----------------------
5320
5321 1. Added a simple "pgrep" utility to the distribution.
5322
5323 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5324 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5325 where "ddd" means "one or more decimal digits".
5326
5327 3. Fixed serious bug. If a pattern had a back reference, but the call to
5328 pcre_exec() didn't supply a large enough ovector to record the related
5329 identifying subpattern, the match always failed. PCRE now remembers the number
5330 of the largest back reference, and gets some temporary memory in which to save
5331 the offsets during matching if necessary, in order to ensure that
5332 backreferences always work.
5333
5334 4. Increased the compatibility with Perl in a number of ways:
5335
5336 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5337 to request this handling. The option can be set at compile or exec time.
5338
5339 (b) $ matches before a terminating newline by default; an option
5340 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5341 mode). The option can be set at compile or exec time.
5342
5343 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5344 the same as Perl's. If the decimal number it represents is less than 10
5345 or there aren't that many previous left capturing parentheses, an octal
5346 escape is read. Inside a character class, it's always an octal escape,
5347 even if it is a single digit.
5348
5349 (d) An escaped but undefined alphabetic character is taken as a literal,
5350 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5351 escapes.
5352
5353 (e) {0} is now permitted. (The previous item is removed from the compiled
5354 pattern).
5355
5356 5. Changed all the names of code files so that the basic parts are no longer
5357 than 10 characters, and abolished the teeny "globals.c" file.
5358
5359 6. Changed the handling of character classes; they are now done with a 32-byte
5360 bit map always.
5361
5362 7. Added the -d and /D options to pcretest to make it possible to look at the
5363 internals of compilation without having to recompile pcre.
5364
5365
5366 Version 0.95 23-Sep-97
5367 ----------------------
5368
5369 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5370 \x20 at the start of a run of normal characters. These were being treated as
5371 real characters, instead of the source characters being re-checked.
5372
5373
5374 Version 0.94 18-Sep-97
5375 ----------------------
5376
5377 1. The functions are now thread-safe, with the caveat that the global variables
5378 containing pointers to malloc() and free() or alternative functions are the
5379 same for all threads.
5380
5381 2. Get pcre_study() to generate a bitmap of initial characters for non-
5382 anchored patterns when this is possible, and use it if passed to pcre_exec().
5383
5384
5385 Version 0.93 15-Sep-97
5386 ----------------------
5387
5388 1. /(b)|(:+)/ was computing an incorrect first character.
5389
5390 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5391 but not actually doing anything yet.
5392
5393 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5394 as Perl does (e.g. [-az] or [az-]).
5395
5396 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5397 all possible positions.
5398
5399 5. Split up into different modules to avoid including unneeded functions in a
5400 compiled binary. However, compile and exec are still in one module. The "study"
5401 function is split off.
5402
5403 6. The character tables are now in a separate module whose source is generated
5404 by an auxiliary program - but can then be edited by hand if required. There are
5405 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5406 toupper() in the code.
5407
5408 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5409 make them global. Abolish the function for setting them, as the caller can now
5410 set them directly.
5411
5412
5413 Version 0.92 11-Sep-97
5414 ----------------------
5415
5416 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5417 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5418
5419 2. Caseless matching was not working in character classes if the characters in
5420 the pattern were in upper case.
5421
5422 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5423
5424 4. Make PCRE_ANCHORED public and accept as a compile option.
5425
5426 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5427 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5428 pass them.
5429
5430 6. Give an error if bad option bits passed at compile or run time.
5431
5432 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5433 pcretest to cause it to pass that flag.
5434
5435 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5436 options, and the first character, if set.
5437
5438 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5439
5440
5441 Version 0.91 10-Sep-97
5442 ----------------------
5443
5444 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5445 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5446
5447 2. PCRE was looping on encountering an indefinitely repeated back reference to
5448 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5449 Perl does - treats the match as successful.
5450
5451 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5