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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1449 - (show annotations)
Fri Jan 17 16:59:52 2014 UTC (5 years, 8 months ago) by ph10
File size: 262622 byte(s)
Fix error in README; give an error for invalid --with-pcregrep-bufsize values 
in ./configure.
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
91 Version 8.34 15-December-2013
92 -----------------------------
93
94 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
95 executable memory. Patch inspired by Carsten Klein.
96
97 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
98 this macro is never tested and has no effect, because the work to support
99 coverage involves only compiling and linking options and special targets in
100 the Makefile. The comment in config.h implied that defining the macro would
101 enable coverage support, which is totally false. There was also support for
102 setting this macro in the CMake files (my fault, I just copied it from
103 configure). SUPPORT_GCOV has now been removed.
104
105 3. Make a small performance improvement in strlen16() and strlen32() in
106 pcretest.
107
108 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
109 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
110 these cases). There was also one in pcretest.c.
111
112 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
113
114 6. In UTF mode, the code for checking whether a group could match an empty
115 string (which is used for indefinitely repeated groups to allow for
116 breaking an infinite loop) was broken when the group contained a repeated
117 negated single-character class with a character that occupied more than one
118 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
119 UTF-8 mode). The effect was undefined: the group might or might not be
120 deemed as matching an empty string, or the program might have crashed.
121
122 7. The code for checking whether a group could match an empty string was not
123 recognizing that \h, \H, \v, \V, and \R must match a character.
124
125 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
126 an empty string. If it can, pcretest shows this in its information output.
127
128 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
129 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
130 matched by pcre_exec() without using JIT:
131
132 (a) If the rest of the pattern did not match after a maximal run of
133 grapheme clusters, the code for backing up to try with fewer of them
134 did not always back up over a full grapheme when characters that do not
135 have the modifier quality were involved, e.g. Hangul syllables.
136
137 (b) If the match point in a subject started with modifier character, and
138 there was no match, the code could incorrectly back up beyond the match
139 point, and potentially beyond the first character in the subject,
140 leading to a segfault or an incorrect match result.
141
142 10. A conditional group with an assertion condition could lead to PCRE
143 recording an incorrect first data item for a match if no other first data
144 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
145 first data item, and therefore matched "ca" after "c" instead of at the
146 start.
147
148 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
149 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
150
151 12. The source of pcregrep now includes z/OS-specific code so that it can be
152 compiled for z/OS as part of the special z/OS distribution.
153
154 13. Added the -T and -TM options to pcretest.
155
156 14. The code in pcre_compile.c for creating the table of named capturing groups
157 has been refactored. Instead of creating the table dynamically during the
158 actual compiling pass, the information is remembered during the pre-compile
159 pass (on the stack unless there are more than 20 named groups, in which
160 case malloc() is used) and the whole table is created before the actual
161 compile happens. This has simplified the code (it is now nearly 150 lines
162 shorter) and prepared the way for better handling of references to groups
163 with duplicate names.
164
165 15. A back reference to a named subpattern when there is more than one of the
166 same name now checks them in the order in which they appear in the pattern.
167 The first one that is set is used for the reference. Previously only the
168 first one was inspected. This change makes PCRE more compatible with Perl.
169
170 16. Unicode character properties were updated from Unicode 6.3.0.
171
172 17. The compile-time code for auto-possessification has been refactored, based
173 on a patch by Zoltan Herczeg. It now happens after instead of during
174 compilation. The code is cleaner, and more cases are handled. The option
175 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
176 options in pcretest are provided to set it. It can also be set by
177 (*NO_AUTO_POSSESS) at the start of a pattern.
178
179 18. The character VT has been added to the default ("C" locale) set of
180 characters that match \s and are generally treated as white space,
181 following this same change in Perl 5.18. There is now no difference between
182 "Perl space" and "POSIX space". Whether VT is treated as white space in
183 other locales depends on the locale.
184
185 19. The code for checking named groups as conditions, either for being set or
186 for being recursed, has been refactored (this is related to 14 and 15
187 above). Processing unduplicated named groups should now be as fast at
188 numerical groups, and processing duplicated groups should be faster than
189 before.
190
191 20. Two patches to the CMake build system, by Alexander Barkov:
192
193 (1) Replace the "source" command by "." in CMakeLists.txt because
194 "source" is a bash-ism.
195
196 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
197 without these the CMake build does not work on Solaris.
198
199 21. Perl has changed its handling of \8 and \9. If there is no previously
200 encountered capturing group of those numbers, they are treated as the
201 literal characters 8 and 9 instead of a binary zero followed by the
202 literals. PCRE now does the same.
203
204 22. Following Perl, added \o{} to specify codepoints in octal, making it
205 possible to specify values greater than 0777 and also making them
206 unambiguous.
207
208 23. Perl now gives an error for missing closing braces after \x{... instead of
209 treating the string as literal. PCRE now does the same.
210
211 24. RunTest used to grumble if an inappropriate test was selected explicitly,
212 but just skip it when running all tests. This make it awkward to run ranges
213 of tests when one of them was inappropriate. Now it just skips any
214 inappropriate tests, as it always did when running all tests.
215
216 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
217 character types such as \d or \w, too many callouts were inserted, and the
218 data that they returned was rubbish.
219
220 26. In UCP mode, \s was not matching two of the characters that Perl matches,
221 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
222 were matched by \h. The code has now been refactored so that the lists of
223 the horizontal and vertical whitespace characters used for \h and \v (which
224 are defined only in one place) are now also used for \s.
225
226 27. Add JIT support for the 64 bit TileGX architecture.
227 Patch by Jiong Wang (Tilera Corporation).
228
229 28. Possessive quantifiers for classes (both explicit and automatically
230 generated) now use special opcodes instead of wrapping in ONCE brackets.
231
232 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
233 (because it's meaningless), this was not happening when PCRE_CASELESS was
234 set. Not wrong, but inefficient.
235
236 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
237 properties for \w, \d, etc) is present in a test regex. Otherwise if the
238 test contains no characters greater than 255, Perl doesn't realise it
239 should be using Unicode semantics.
240
241 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
242 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
243 does in Unicode mode.
244
245 32. Added the "forbid" facility to pcretest so that putting tests into the
246 wrong test files can sometimes be quickly detected.
247
248 33. There is now a limit (default 250) on the depth of nesting of parentheses.
249 This limit is imposed to control the amount of system stack used at compile
250 time. It can be changed at build time by --with-parens-nest-limit=xxx or
251 the equivalent in CMake.
252
253 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
254 errors. Perl warns for these when in warning mode, but PCRE has no facility
255 for giving warnings.
256
257 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
258 However, this was not working for (?!) because it is optimized to (*FAIL),
259 for which PCRE does not allow quantifiers. The optimization is now disabled
260 when a quantifier follows (?!). I can't see any use for this, but it makes
261 things uniform.
262
263 36. Perl no longer allows group names to start with digits, so I have made this
264 change also in PCRE. It simplifies the code a bit.
265
266 37. In extended mode, Perl ignores spaces before a + that indicates a
267 possessive quantifier. PCRE allowed a space before the quantifier, but not
268 before the possessive +. It now does.
269
270 38. The use of \K (reset reported match start) within a repeated possessive
271 group such as (a\Kb)*+ was not working.
272
273 40. Document that the same character tables must be used at compile time and
274 run time, and that the facility to pass tables to pcre_exec() and
275 pcre_dfa_exec() is for use only with saved/restored patterns.
276
277 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
278 features for Builds with MSVC:
279
280 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
281 builds). The .rc files can be used to set FileDescription and many other
282 attributes.
283
284 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
285 This allows higher-level build scripts which want .pdb files to avoid
286 hard-coding the exact files needed."
287
288 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
289 mean "start of word" and "end of word", respectively, as a transition aid.
290
291 43. A minimizing repeat of a class containing codepoints greater than 255 in
292 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
293 compiled to use the heap for recursion.
294
295 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
296 is configured.
297
298
299 Version 8.33 28-May-2013
300 ------------------------
301
302 1. Added 'U' to some constants that are compared to unsigned integers, to
303 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
304 variables that are added to signed variables, to ensure the result is
305 signed and can be negated.
306
307 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
308 CMake config files.
309
310 3. Revise the creation of config.h.generic so that all boolean macros are
311 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
312 overriding via -D on the command line possible.
313
314 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
315 to unsigned int is reported to make a quite noticeable speed difference in
316 a specific Windows environment. Testing on Linux did also appear to show
317 some benefit (and it is clearly not harmful). Also fixed the definition of
318 Xop which should be unsigned.
319
320 5. Related to (4), changing the definition of the intermediate variable cc
321 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
322 improvements.
323
324 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
325 unnecessary spaces.
326
327 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
328 and later.
329
330 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
331
332 9. Optimizing fast_forward_start_bits in JIT.
333
334 10. Adding support for callouts in JIT, and fixing some issues revealed
335 during this work. Namely:
336
337 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
338
339 (b) Minimum length was not checked before the matching is started.
340
341 11. The value of capture_last that is passed to callouts was incorrect in some
342 cases when there was a capture on one path that was subsequently abandoned
343 after a backtrack. Also, the capture_last value is now reset after a
344 recursion, since all captures are also reset in this case.
345
346 12. The interpreter no longer returns the "too many substrings" error in the
347 case when an overflowing capture is in a branch that is subsequently
348 abandoned after a backtrack.
349
350 13. In the pathological case when an offset vector of size 2 is used, pcretest
351 now prints out the matched string after a yield of 0 or 1.
352
353 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
354 Only supported by the JIT compiler at the moment.
355
356 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
357
358 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
359 offset of the starting point of the matching process, provided the offsets
360 vector is large enough.
361
362 17. The \A escape now records a lookbehind value of 1, though its execution
363 does not actually inspect the previous character. This is to ensure that,
364 in partial multi-segment matching, at least one character from the old
365 segment is retained when a new segment is processed. Otherwise, if there
366 are no lookbehinds in the pattern, \A might match incorrectly at the start
367 of a new segment.
368
369 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
370
371 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
372 gives some modest performance improvement in 8-bit mode.
373
374 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
375 be expressed in certain programming languages using Universal Character
376 Names.
377
378 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
379 which points out that "non characters" are not "characters that may not
380 appear in Unicode strings" but rather "characters that are reserved for
381 internal use and have only local meaning".
382
383 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
384 there was a conditional group that depended on an assertion, if the
385 assertion was false, the callout that immediately followed the alternation
386 in the condition was skipped when pcre_exec() was used for matching.
387
388 23. Allow an explicit callout to be inserted before an assertion that is the
389 condition for a conditional group, for compatibility with automatic
390 callouts, which always insert a callout at this point.
391
392 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
393 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
394
395 25. (*PRUNE) is now supported by the JIT compiler.
396
397 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
398
399 27. Fix the case where there are two or more SKIPs with arguments that may be
400 ignored.
401
402 28. (*SKIP) is now supported by the JIT compiler.
403
404 29. (*THEN) is now supported by the JIT compiler.
405
406 30. Update RunTest with additional test selector options.
407
408 31. The way PCRE handles backtracking verbs has been changed in two ways.
409
410 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
411 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
412 backtracking. In some cases this makes it more Perl-compatible, but Perl's
413 rather obscure rules do not always do the same thing.
414
415 (2) Previously, backtracking verbs were confined within assertions. This is
416 no longer the case for positive assertions, except for (*ACCEPT). Again,
417 this sometimes improves Perl compatibility, and sometimes does not.
418
419 32. A number of tests that were in test 2 because Perl did things differently
420 have been moved to test 1, because either Perl or PCRE has changed, and
421 these tests are now compatible.
422
423 32. Backtracking control verbs are now handled in the same way in JIT and
424 interpreter.
425
426 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
427 contained a forward subroutine reference caused a compile error.
428
429 34. Auto-detect and optimize limited repetitions in JIT.
430
431 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
432 blocking (*UTF) etc.
433
434 36. In the interpreter, maximizing pattern repetitions for characters and
435 character types now use tail recursion, which reduces stack usage.
436
437 37. The value of the max lookbehind was not correctly preserved if a compiled
438 and saved regex was reloaded on a host of different endianness.
439
440 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
441 of the compiled pattern block, expand the flags field from 16 to 32 bits
442 because it was almost full.
443
444 39. Try madvise first before posix_madvise.
445
446 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
447 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
448 original change locked out all matches of empty strings. This has been
449 changed so that one match of an empty string per line is recognized.
450 Subsequent searches on the same line (for colouring or for --only-matching,
451 for example) do not recognize empty strings.
452
453 41. Applied a user patch to fix a number of spelling mistakes in comments.
454
455 42. Data lines longer than 65536 caused pcretest to crash.
456
457 43. Clarified the data type for length and startoffset arguments for pcre_exec
458 and pcre_dfa_exec in the function-specific man pages, where they were
459 explicitly stated to be in bytes, never having been updated. I also added
460 some clarification to the pcreapi man page.
461
462 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
463 a segmentation fault.
464
465
466 Version 8.32 30-November-2012
467 -----------------------------
468
469 1. Improved JIT compiler optimizations for first character search and single
470 character iterators.
471
472 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
473 Patch by Daniel Richard G.
474
475 3. Single character iterator optimizations in the JIT compiler.
476
477 4. Improved JIT compiler optimizations for character ranges.
478
479 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
480 Reported by Giuseppe D'Angelo.
481
482 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
483 of a line, was being set incorrectly in cases where .* appeared inside
484 atomic brackets at the start of a pattern, or where there was a subsequent
485 *PRUNE or *SKIP.
486
487 7. Improved instruction cache flush for POWER/PowerPC.
488 Patch by Daniel Richard G.
489
490 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
491 grep:
492
493 (a) There is now no limit to the number of patterns to be matched.
494
495 (b) An error is given if a pattern is too long.
496
497 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
498 are now supported.
499
500 (d) --exclude-from and --include-from (multiple use) have been added.
501
502 (e) Exclusions and inclusions now apply to all files and directories, not
503 just to those obtained from scanning a directory recursively.
504
505 (f) Multiple uses of -f and --file-list are now supported.
506
507 (g) In a Windows environment, the default for -d has been changed from
508 "read" (the GNU grep default) to "skip", because otherwise the presence
509 of a directory in the file list provokes an error.
510
511 (h) The documentation has been revised and clarified in places.
512
513 9. Improve the matching speed of capturing brackets.
514
515 10. Changed the meaning of \X so that it now matches a Unicode extended
516 grapheme cluster.
517
518 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
519 out POSIX threads when JIT support is configured.
520
521 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
522
523 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
524 structure could go wrong in environments where size_t is not the same size
525 as int.
526
527 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
528 set.
529
530 15. The EBCDIC support had decayed; later updates to the code had included
531 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
532 general tidy up of EBCDIC-related issues, and the documentation was also
533 not quite right. There is now a test that can be run on ASCII systems to
534 check some of the EBCDIC-related things (but is it not a full test).
535
536 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
537 in a small tidy to the code.
538
539 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
540
541 18. If the --only-matching (-o) option in pcregrep is specified multiple
542 times, each one causes appropriate output. For example, -o1 -o2 outputs the
543 substrings matched by the 1st and 2nd capturing parentheses. A separating
544 string can be specified by --om-separator (default empty).
545
546 19. Improving the first n character searches.
547
548 20. Turn case lists for horizontal and vertical white space into macros so that
549 they are defined only once.
550
551 21. This set of changes together give more compatible Unicode case-folding
552 behaviour for characters that have more than one other case when UCP
553 support is available.
554
555 (a) The Unicode property table now has offsets into a new table of sets of
556 three or more characters that are case-equivalent. The MultiStage2.py
557 script that generates these tables (the pcre_ucd.c file) now scans
558 CaseFolding.txt instead of UnicodeData.txt for character case
559 information.
560
561 (b) The code for adding characters or ranges of characters to a character
562 class has been abstracted into a generalized function that also handles
563 case-independence. In UTF-mode with UCP support, this uses the new data
564 to handle characters with more than one other case.
565
566 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
567 whose other case is greater than 256 are now correctly matched
568 caselessly. Previously, the high codepoint matched the low one, but not
569 vice versa.
570
571 (d) The processing of \h, \H, \v, and \ in character classes now makes use
572 of the new class addition function, using character lists defined as
573 macros alongside the case definitions of 20 above.
574
575 (e) Caseless back references now work with characters that have more than
576 one other case.
577
578 (f) General caseless matching of characters with more than one other case
579 is supported.
580
581 22. Unicode character properties were updated from Unicode 6.2.0
582
583 23. Improved CMake support under Windows. Patch by Daniel Richard G.
584
585 24. Add support for 32-bit character strings, and UTF-32
586
587 25. Major JIT compiler update (code refactoring and bugfixing).
588 Experimental Sparc 32 support is added.
589
590 26. Applied a modified version of Daniel Richard G's patch to create
591 pcre.h.generic and config.h.generic by "make" instead of in the
592 PrepareRelease script.
593
594 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
595 pcre_compile.c when checking for a zero character.
596
597 28. Introducing a native interface for JIT. Through this interface, the compiled
598 machine code can be directly executed. The purpose of this interface is to
599 provide fast pattern matching, so several sanity checks are not performed.
600 However, feature tests are still performed. The new interface provides
601 1.4x speedup compared to the old one.
602
603 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
604 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
605 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
606
607 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
608 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
609 the "old" RFC 2279). Instead, it ended up passing a negative length to
610 pcre_exec().
611
612 31. Add support for GCC's visibility feature to hide internal functions.
613
614 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
615 "unknown -C option" after outputting 0 or 1.
616
617 33. There is now support for generating a code coverage report for the test
618 suite in environments where gcc is the compiler and lcov is installed. This
619 is mainly for the benefit of the developers.
620
621 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
622 unaddressable using valgrind annotations, allowing valgrind to detect
623 invalid memory accesses. This is mainly for the benefit of the developers.
624
625 25. (*UTF) can now be used to start a pattern in any of the three libraries.
626
627 26. Give configure error if --enable-cpp but no C++ compiler found.
628
629
630 Version 8.31 06-July-2012
631 -------------------------
632
633 1. Fixing a wrong JIT test case and some compiler warnings.
634
635 2. Removed a bashism from the RunTest script.
636
637 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
638 to unsigned type, result still unsigned" that was given by an MS compiler
639 on encountering the code "-sizeof(xxx)".
640
641 4. Partial matching support is added to the JIT compiler.
642
643 5. Fixed several bugs concerned with partial matching of items that consist
644 of more than one character:
645
646 (a) /^(..)\1/ did not partially match "aba" because checking references was
647 done on an "all or nothing" basis. This also applied to repeated
648 references.
649
650 (b) \R did not give a hard partial match if \r was found at the end of the
651 subject.
652
653 (c) \X did not give a hard partial match after matching one or more
654 characters at the end of the subject.
655
656 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
657 a partial match for the string "\r".
658
659 (e) When newline was set to CRLF, the metacharacter "." did not recognize
660 a partial match for a CR character at the end of the subject string.
661
662 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
663 running pcretest, the text "(JIT)" added to the output whenever JIT is
664 actually used to run the match.
665
666 7. Individual JIT compile options can be set in pcretest by following -s+[+]
667 or /S+[+] with a digit between 1 and 7.
668
669 8. OP_NOT now supports any UTF character not just single-byte ones.
670
671 9. (*MARK) control verb is now supported by the JIT compiler.
672
673 10. The command "./RunTest list" lists the available tests without actually
674 running any of them. (Because I keep forgetting what they all are.)
675
676 11. Add PCRE_INFO_MAXLOOKBEHIND.
677
678 12. Applied a (slightly modified) user-supplied patch that improves performance
679 when the heap is used for recursion (compiled with --disable-stack-for-
680 recursion). Instead of malloc and free for each heap frame each time a
681 logical recursion happens, frames are retained on a chain and re-used where
682 possible. This sometimes gives as much as 30% improvement.
683
684 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
685 call.
686
687 14. As documented, (*COMMIT) is now confined to within a positive assertion.
688
689 15. It is now possible to link pcretest with libedit as an alternative to
690 libreadline.
691
692 16. (*COMMIT) control verb is now supported by the JIT compiler.
693
694 17. The Unicode data tables have been updated to Unicode 6.1.0.
695
696 18. Added --file-list option to pcregrep.
697
698 19. Added binary file support to pcregrep, including the -a, --binary-files,
699 -I, and --text options.
700
701 20. The madvise function is renamed for posix_madvise for QNX compatibility
702 reasons. Fixed by Giuseppe D'Angelo.
703
704 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
705
706 22. Changed the option for creating symbolic links for 16-bit man pages from
707 -s to -sf so that re-installing does not cause issues.
708
709 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
710
711 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
712 matches in certain environments (the workspace was not being correctly
713 retained). Also added to pcre_dfa_exec() a simple plausibility check on
714 some of the workspace data at the beginning of a restart.
715
716 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
717 was not doing so when it should - probably a typo introduced by SVN 528
718 (change 8.10/14).
719
720 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
721 \w+ when the character tables indicated that \x{c4} was a word character.
722 There were several related cases, all because the tests for doing a table
723 lookup were testing for characters less than 127 instead of 255.
724
725 27. If a pattern contains capturing parentheses that are not used in a match,
726 their slots in the ovector are set to -1. For those that are higher than
727 any matched groups, this happens at the end of processing. In the case when
728 there were back references that the ovector was too small to contain
729 (causing temporary malloc'd memory to be used during matching), and the
730 highest capturing number was not used, memory off the end of the ovector
731 was incorrectly being set to -1. (It was using the size of the temporary
732 memory instead of the true size.)
733
734 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
735 ovector size, it uses memory at the end of the block that it has got.
736
737 29. Check for an overlong MARK name and give an error at compile time. The
738 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
739
740 30. JIT compiler update.
741
742 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
743 Rill for the patch.
744
745 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
746
747 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
748
749 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
750 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
751 was enabled.
752
753 35. Improve JIT code generation for greedy plus quantifier.
754
755 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
756 1 to "aa" instead of to an empty string. The bug affected repeated groups
757 that could potentially match an empty string.
758
759 37. Optimizing single character iterators in JIT.
760
761 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
762 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
763 codepoints that are too big for the mode are faulted, and in a UTF mode,
764 disallowed codepoints are also faulted.
765
766 39. If PCRE was compiled with UTF support, in three places in the DFA
767 matcher there was code that should only have been obeyed in UTF mode, but
768 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
769 processing when bytes with values greater than 127 were present. In 16-bit
770 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
771 both cases the values are those that cannot be the first data item in a UTF
772 character. The three items that might have provoked this were recursions,
773 possessively repeated groups, and atomic groups.
774
775 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
776 and pcregrep, because some OS require shared objects to be explicitly
777 passed to ld, causing the link step to fail if they are not.
778
779 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
780 mode, patterns that started with \h* or \R* might be incorrectly matched.
781
782
783 Version 8.30 04-February-2012
784 -----------------------------
785
786 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
787 name is defined in ctype.h.
788
789 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
790 only in quite long subpatterns.
791
792 3. Removed the function pcre_info(), which has been obsolete and deprecated
793 since it was replaced by pcre_fullinfo() in February 2000.
794
795 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
796 match a (*MARK), and the match failed at the start of the subject, a
797 reference to memory before the start of the subject could occur. This bug
798 was introduced by fix 17 of release 8.21.
799
800 5. A reference to an unset group with zero minimum repetition was giving
801 totally wrong answers (in non-JavaScript-compatibility mode). For example,
802 /(another)?(\1?)test/ matched against "hello world test". This bug was
803 introduced in release 8.13.
804
805 6. Add support for 16-bit character strings (a large amount of work involving
806 many changes and refactorings).
807
808 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
809 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
810 from a file.
811
812 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
813 rounding is not applied in this particular case).
814
815 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
816 if they appear, or are escaped, in patterns.
817
818 10. Get rid of a number of -Wunused-but-set-variable warnings.
819
820 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
821 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
822 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
823 also returns the mark "x". This bug applied to capturing parentheses,
824 non-capturing parentheses, and atomic parentheses. It also applied to some
825 assertions.
826
827 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
828 information out of configure.ac instead of relying on pcre.h.generic, which
829 is not stored in the repository.
830
831 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
832 -lreadline.
833
834 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
835
836 15. Applied Graycode's patch to put the top-level frame on the stack rather
837 than the heap when not using the stack for recursion. This gives a
838 performance improvement in many cases when recursion is not deep.
839
840 16. Experimental code added to "pcretest -C" to output the stack frame size.
841
842
843 Version 8.21 12-Dec-2011
844 ------------------------
845
846 1. Updating the JIT compiler.
847
848 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
849 are added as well.
850
851 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
852 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
853 calling _pcre_jit_exec. Some extra comments are added.
854
855 4. (*MARK) settings inside atomic groups that do not contain any capturing
856 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
857 was introduced by change 18 for 8.20.
858
859 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
860 ECMA-262 standard.
861
862 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
863 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
864 This bug was probably introduced by change 9 of 8.13.
865
866 7. While fixing 6 above, I noticed that a number of other items were being
867 incorrectly rejected as "not fixed length". This arose partly because newer
868 opcodes had not been added to the fixed-length checking code. I have (a)
869 corrected the bug and added tests for these items, and (b) arranged for an
870 error to occur if an unknown opcode is encountered while checking for fixed
871 length instead of just assuming "not fixed length". The items that were
872 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
873 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
874 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
875
876 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
877 being incorrectly compiled and would have given unpredicatble results.
878
879 9. A possessively repeated subpattern with minimum repeat count greater than
880 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
881 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
882 the first (A) could occur when it should not.
883
884 10. Add a cast and remove a redundant test from the code.
885
886 11. JIT should use pcre_malloc/pcre_free for allocation.
887
888 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
889 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
890 is anything other than /usr, -L is still shown).
891
892 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
893
894 14. Perl does not support \N without a following name in a [] class; PCRE now
895 also gives an error.
896
897 15. If a forward reference was repeated with an upper limit of around 2000,
898 it caused the error "internal error: overran compiling workspace". The
899 maximum number of forward references (including repeats) was limited by the
900 internal workspace, and dependent on the LINK_SIZE. The code has been
901 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
902 the default depends on LINK_SIZE. There is a new upper limit (for safety)
903 of around 200,000 forward references. While doing this, I also speeded up
904 the filling in of repeated forward references.
905
906 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
907 incorrectly expecting the subject to contain another "a" after the start.
908
909 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
910 in the match, the SKIP should be ignored. This was not happening; instead
911 the SKIP was being treated as NOMATCH. For patterns such as
912 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
913 tested.
914
915 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
916 now much more compatible with Perl, in particular in cases where the result
917 is a non-match for a non-anchored pattern. For example, if
918 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
919 "m", where previously it did not return a name. A side effect of this
920 change is that for partial matches, the last encountered mark name is
921 returned, as for non matches. A number of tests that were previously not
922 Perl-compatible have been moved into the Perl-compatible test files. The
923 refactoring has had the pleasing side effect of removing one argument from
924 the match() function, thus reducing its stack requirements.
925
926 19. If the /S+ option was used in pcretest to study a pattern using JIT,
927 subsequent uses of /S (without +) incorrectly behaved like /S+.
928
929 21. Retrieve executable code size support for the JIT compiler and fixing
930 some warnings.
931
932 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
933 not work when the shorter character appeared right at the end of the
934 subject string.
935
936 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
937 systems.
938
939 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
940 output it when the /M option is used in pcretest.
941
942 25. The CheckMan script was not being included in the distribution. Also, added
943 an explicit "perl" to run Perl scripts from the PrepareRelease script
944 because this is reportedly needed in Windows.
945
946 26. If study data was being save in a file and studying had not found a set of
947 "starts with" bytes for the pattern, the data written to the file (though
948 never used) was taken from uninitialized memory and so caused valgrind to
949 complain.
950
951 27. Updated RunTest.bat as provided by Sheri Pierce.
952
953 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
954
955 29. Computation of memory usage for the table of capturing group names was
956 giving an unnecessarily large value.
957
958
959 Version 8.20 21-Oct-2011
960 ------------------------
961
962 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
963 a POSIX class. After further experiments with Perl, which convinced me that
964 Perl has bugs and confusions, a closing square bracket is no longer allowed
965 in a POSIX name. This bug also affected patterns with classes that started
966 with full stops.
967
968 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
969 captured substring, but while checking the failing first alternative,
970 substring 1 is temporarily captured. If the output vector supplied to
971 pcre_exec() was not big enough for this capture, the yield of the function
972 was still zero ("insufficient space for captured substrings"). This cannot
973 be totally fixed without adding another stack variable, which seems a lot
974 of expense for a edge case. However, I have improved the situation in cases
975 such as /(a)(b)x|abc/ matched against "abc", where the return code
976 indicates that fewer than the maximum number of slots in the ovector have
977 been set.
978
979 3. Related to (2) above: when there are more back references in a pattern than
980 slots in the output vector, pcre_exec() uses temporary memory during
981 matching, and copies in the captures as far as possible afterwards. It was
982 using the entire output vector, but this conflicts with the specification
983 that only 2/3 is used for passing back captured substrings. Now it uses
984 only the first 2/3, for compatibility. This is, of course, another edge
985 case.
986
987 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
988 main code base, and can be used by building with --enable-jit. When this is
989 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
990 runtime --no-jit option is given.
991
992 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
993 ovector, the return from the function was zero, implying that there were
994 other matches that did not fit. The correct "exactly full" value is now
995 returned.
996
997 6. If a subpattern that was called recursively or as a subroutine contained
998 (*PRUNE) or any other control that caused it to give a non-standard return,
999 invalid errors such as "Error -26 (nested recursion at the same subject
1000 position)" or even infinite loops could occur.
1001
1002 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1003 computing the minimum length on reaching *ACCEPT, and so ended up with the
1004 wrong value of 1 rather than 0. Further investigation indicates that
1005 computing a minimum subject length in the presence of *ACCEPT is difficult
1006 (think back references, subroutine calls), and so I have changed the code
1007 so that no minimum is registered for a pattern that contains *ACCEPT.
1008
1009 8. If (*THEN) was present in the first (true) branch of a conditional group,
1010 it was not handled as intended. [But see 16 below.]
1011
1012 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1013 Sheri Pierce.
1014
1015 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1016 the first byte in a match must be "a".
1017
1018 11. Change 17 for 8.13 increased the recursion depth for patterns like
1019 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1020 pattern contains any instances of (*THEN). If it does not, the old
1021 optimizations are restored. It would be nice to do this on a per-group
1022 basis, but at the moment that is not feasible.
1023
1024 12. In some environments, the output of pcretest -C is CRLF terminated. This
1025 broke RunTest's code that checks for the link size. A single white space
1026 character after the value is now allowed for.
1027
1028 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1029 For "fr", it uses the Windows-specific input and output files.
1030
1031 14. If (*THEN) appeared in a group that was called recursively or as a
1032 subroutine, it did not work as intended. [But see next item.]
1033
1034 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1035 pattern fragments (but not containing any | characters). If A and B are
1036 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1037 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1038 D. In other words, Perl considers parentheses that do not contain any |
1039 characters to be part of a surrounding alternative, whereas PCRE was
1040 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1041 differently. PCRE now behaves in the same way as Perl, except in the case
1042 of subroutine/recursion calls such as (?1) which have in any case always
1043 been different (but PCRE had them first :-).
1044
1045 16. Related to 15 above: Perl does not treat the | in a conditional group as
1046 creating alternatives. Such a group is treated in the same way as an
1047 ordinary group without any | characters when processing (*THEN). PCRE has
1048 been changed to match Perl's behaviour.
1049
1050 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1051 RunGrepTest script failed.
1052
1053 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1054 inevitable for groups that contain captures, but it can lead to a lot of
1055 stack use in large patterns. The old behaviour has been restored for atomic
1056 groups that do not contain any capturing parentheses.
1057
1058 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1059 suppress the check for a minimum subject length at run time. (If it was
1060 given to pcre_exec() or pcre_dfa_exec() it did work.)
1061
1062 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1063 fail to work when decoding hex characters in data strings in EBCDIC
1064 environments.
1065
1066 21. It appears that in at least one Mac OS environment, the isxdigit() function
1067 is implemented as a macro that evaluates to its argument more than once,
1068 contravening the C 90 Standard (I haven't checked a later standard). There
1069 was an instance in pcretest which caused it to go wrong when processing
1070 \x{...} escapes in subject strings. The has been rewritten to avoid using
1071 things like p++ in the argument of isxdigit().
1072
1073
1074 Version 8.13 16-Aug-2011
1075 ------------------------
1076
1077 1. The Unicode data tables have been updated to Unicode 6.0.0.
1078
1079 2. Two minor typos in pcre_internal.h have been fixed.
1080
1081 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1082 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1083 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1084
1085 4. There were a number of related bugs in the code for matching backrefences
1086 caselessly in UTF-8 mode when codes for the characters concerned were
1087 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1088 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1089 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1090 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1091 2-byte code at the end of the subject (it thought there wasn't enough data
1092 left).
1093
1094 5. Comprehensive information about what went wrong is now returned by
1095 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1096 as the output vector has at least 2 elements. The offset of the start of
1097 the failing character and a reason code are placed in the vector.
1098
1099 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1100 now returned is for the first byte of the failing character, instead of the
1101 last byte inspected. This is an incompatible change, but I hope it is small
1102 enough not to be a problem. It makes the returned offset consistent with
1103 pcre_exec() and pcre_dfa_exec().
1104
1105 7. pcretest now gives a text phrase as well as the error number when
1106 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1107 failure, the offset and reason code are output.
1108
1109 8. When \R was used with a maximizing quantifier it failed to skip backwards
1110 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1111 back over a single character (\n). This seems wrong (because it treated the
1112 two characters as a single entity when going forwards), conflicts with the
1113 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1114 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1115 has been changed.
1116
1117 9. Some internal refactoring has changed the processing so that the handling
1118 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1119 time (the PCRE_DOTALL option was changed this way some time ago: version
1120 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1121 which was always a bit of a fudge. It also means that there is one less
1122 argument for the match() function, which reduces its stack requirements
1123 slightly. This change also fixes an incompatibility with Perl: the pattern
1124 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1125
1126 10. More internal refactoring has drastically reduced the number of recursive
1127 calls to match() for possessively repeated groups such as (abc)++ when
1128 using pcre_exec().
1129
1130 11. While implementing 10, a number of bugs in the handling of groups were
1131 discovered and fixed:
1132
1133 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1134 (a|)*(?1) gave a compile-time internal error.
1135 ((a|)+)+ did not notice that the outer group could match an empty string.
1136 (^a|^)+ was not marked as anchored.
1137 (.*a|.*)+ was not marked as matching at start or after a newline.
1138
1139 12. Yet more internal refactoring has removed another argument from the match()
1140 function. Special calls to this function are now indicated by setting a
1141 value in a variable in the "match data" data block.
1142
1143 13. Be more explicit in pcre_study() instead of relying on "default" for
1144 opcodes that mean there is no starting character; this means that when new
1145 ones are added and accidentally left out of pcre_study(), testing should
1146 pick them up.
1147
1148 14. The -s option of pcretest has been documented for ages as being an old
1149 synonym of -m (show memory usage). I have changed it to mean "force study
1150 for every regex", that is, assume /S for every regex. This is similar to -i
1151 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1152 using it. It makes it easier to run collections of tests with and without
1153 study enabled, and thereby test pcre_study() more easily. All the standard
1154 tests are now run with and without -s (but some patterns can be marked as
1155 "never study" - see 20 below).
1156
1157 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1158 restoration of the capturing data to the outer values was not happening
1159 correctly.
1160
1161 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1162 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1163 pattern had matched an empty string, and so incorrectly returned a no
1164 match.
1165
1166 17. There was optimizing code for the last branch of non-capturing parentheses,
1167 and also for the obeyed branch of a conditional subexpression, which used
1168 tail recursion to cut down on stack usage. Unfortunately, now that there is
1169 the possibility of (*THEN) occurring in these branches, tail recursion is
1170 no longer possible because the return has to be checked for (*THEN). These
1171 two optimizations have therefore been removed. [But see 8.20/11 above.]
1172
1173 18. If a pattern containing \R was studied, it was assumed that \R always
1174 matched two bytes, thus causing the minimum subject length to be
1175 incorrectly computed because \R can also match just one byte.
1176
1177 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1178 was incorrectly computed.
1179
1180 20. If /S is present twice on a test pattern in pcretest input, it now
1181 *disables* studying, thereby overriding the use of -s on the command line
1182 (see 14 above). This is necessary for one or two tests to keep the output
1183 identical in both cases.
1184
1185 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1186 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1187
1188 22. When an atomic group that contained a capturing parenthesis was
1189 successfully matched, but the branch in which it appeared failed, the
1190 capturing was not being forgotten if a higher numbered group was later
1191 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1192 group 1 to "a", when in fact it should be unset. This applied to multi-
1193 branched capturing and non-capturing groups, repeated or not, and also to
1194 positive assertions (capturing in negative assertions does not happen
1195 in PCRE) and also to nested atomic groups.
1196
1197 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1198 subject after a captured substring, to make it easier to tell which of a
1199 number of identical substrings has been captured.
1200
1201 24. The way atomic groups are processed by pcre_exec() has been changed so that
1202 if they are repeated, backtracking one repetition now resets captured
1203 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1204 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1205 "aaa". Previously, it would have been "a". As part of this code
1206 refactoring, the way recursive calls are handled has also been changed.
1207
1208 25. If an assertion condition captured any substrings, they were not passed
1209 back unless some other capturing happened later. For example, if
1210 (?(?=(a))a) was matched against "a", no capturing was returned.
1211
1212 26. When studying a pattern that contained subroutine calls or assertions,
1213 the code for finding the minimum length of a possible match was handling
1214 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1215 group 1 called group 2 while simultaneously a separate group 2 called group
1216 1). A stack overflow occurred in this case. I have fixed this by limiting
1217 the recursion depth to 10.
1218
1219 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1220 Fortmann. This supports explicit test numbers on the command line, and has
1221 argument validation and error reporting.
1222
1223 28. An instance of \X with an unlimited repeat could fail if at any point the
1224 first character it looked at was a mark character.
1225
1226 29. Some minor code refactoring concerning Unicode properties and scripts
1227 should reduce the stack requirement of match() slightly.
1228
1229 30. Added the '=' option to pcretest to check the setting of unused capturing
1230 slots at the end of the pattern, which are documented as being -1, but are
1231 not included in the return count.
1232
1233 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1234 compiled something random. Now it gives a compile-time error (as does
1235 Perl).
1236
1237 32. A *MARK encountered during the processing of a positive assertion is now
1238 recorded and passed back (compatible with Perl).
1239
1240 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1241 had alternative anchored branches, the search for a second match in a line
1242 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1243 matched the line "0102" twice. The same bug affected patterns that started
1244 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1245 twice.
1246
1247 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1248 does, and because of capturing effects, quantifying parenthesized
1249 assertions may at times be useful. Quantifiers are now allowed for
1250 parenthesized assertions.
1251
1252 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1253
1254 36. \g was being checked for fancy things in a character class, when it should
1255 just be a literal "g".
1256
1257 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1258 appearance of a nested POSIX class supersedes an apparent external class.
1259 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1260 unescaped square brackets may also appear as part of class names. For
1261 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1262 more like Perl. (But see 8.20/1 above.)
1263
1264 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1265 was because it thought it was \N{name}, which is not supported).
1266
1267 39. Add minix to OS list not supporting the -S option in pcretest.
1268
1269 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1270 cannot analyze patterns in sufficient detail to catch mutual recursions
1271 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1272 subgroup is called recursively as a subpattern for a second time at the
1273 same position in the subject string. In previous releases this might have
1274 been caught by the recursion limit, or it might have run out of stack.
1275
1276 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1277 happen only once. PCRE was, however incorrectly giving a compile time error
1278 "recursive call could loop indefinitely" because it cannot analyze the
1279 pattern in sufficient detail. The compile time test no longer happens when
1280 PCRE is compiling a conditional subpattern, but actual runaway loops are
1281 now caught at runtime (see 40 above).
1282
1283 42. It seems that Perl allows any characters other than a closing parenthesis
1284 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1285 has been changed to be the same.
1286
1287 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1288 as not to get warnings when autogen.sh is called. Also changed
1289 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1290
1291 44. To help people who use pcregrep to scan files containing exceedingly long
1292 lines, the following changes have been made:
1293
1294 (a) The default value of the buffer size parameter has been increased from
1295 8K to 20K. (The actual buffer used is three times this size.)
1296
1297 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1298 PCRE is built.
1299
1300 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1301 to be set at run time.
1302
1303 (d) Numerical values in pcregrep options can be followed by K or M, for
1304 example --buffer-size=50K.
1305
1306 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1307 given and the return code is set to 2.
1308
1309 45. Add a pointer to the latest mark to the callout data block.
1310
1311 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1312 partial match of an empty string instead of no match. This was specific to
1313 the use of ".".
1314
1315 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1316 complete match instead of a partial match. This bug was dependent on both
1317 the PCRE_UTF8 and PCRE_DOTALL options being set.
1318
1319 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1320 starting byte set, because \b was not being ignored.
1321
1322
1323 Version 8.12 15-Jan-2011
1324 ------------------------
1325
1326 1. Fixed some typos in the markup of the man pages, and wrote a script that
1327 checks for such things as part of the documentation building process.
1328
1329 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1330 --match-limit and --recursion-limit options (added for 8.11). In
1331 particular, this made one of the standard tests fail. (The integer value
1332 went into the wrong half of a long int.)
1333
1334 3. If the --colour option was given to pcregrep with -v (invert match), it
1335 did strange things, either producing crazy output, or crashing. It should,
1336 of course, ignore a request for colour when reporting lines that do not
1337 match.
1338
1339 4. Another pcregrep bug caused similar problems if --colour was specified with
1340 -M (multiline) and the pattern match finished with a line ending.
1341
1342 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1343 matched in multiline mode, the following line was shown as part of the
1344 match. This seems wrong, so I have changed it.
1345
1346 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1347 the check for further matches in the same line (so they could be coloured)
1348 to overrun the end of the current line. If another match was found, it was
1349 incorrectly shown (and then shown again when found in the next line).
1350
1351 7. If pcregrep was compiled under Windows, there was a reference to the
1352 function pcregrep_exit() before it was defined. I am assuming this was
1353 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1354 reported by a user. I've moved the definition above the reference.
1355
1356
1357 Version 8.11 10-Dec-2010
1358 ------------------------
1359
1360 1. (*THEN) was not working properly if there were untried alternatives prior
1361 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1362 backtracked to try for "b" instead of moving to the next alternative branch
1363 at the same level (in this case, to look for "c"). The Perl documentation
1364 is clear that when (*THEN) is backtracked onto, it goes to the "next
1365 alternative in the innermost enclosing group".
1366
1367 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1368 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1369 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1370 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1371 (*THEN).
1372
1373 3. If \s appeared in a character class, it removed the VT character from
1374 the class, even if it had been included by some previous item, for example
1375 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1376 of \s, but is part of the POSIX "space" class.)
1377
1378 4. A partial match never returns an empty string (because you can always
1379 match an empty string at the end of the subject); however the checking for
1380 an empty string was starting at the "start of match" point. This has been
1381 changed to the "earliest inspected character" point, because the returned
1382 data for a partial match starts at this character. This means that, for
1383 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1384 (previously it gave "no match").
1385
1386 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1387 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1388 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1389 has an implication that the given string is incomplete (because a partial
1390 match is preferred over a full match). For this reason, these items now
1391 give a partial match in this situation. [Aside: previously, the one case
1392 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1393 match rather than a full match, which was wrong by the old rules, but is
1394 now correct.]
1395
1396 6. There was a bug in the handling of #-introduced comments, recognized when
1397 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1398 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1399 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1400 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1401 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1402 places in pcre_compile().
1403
1404 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1405 comments when looking ahead for named forward references to subpatterns,
1406 the only newline sequence it recognized was NL. It now handles newlines
1407 according to the set newline convention.
1408
1409 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1410 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1411 cater for a lack of strerror(). These oversights have been fixed.
1412
1413 9. Added --match-limit and --recursion-limit to pcregrep.
1414
1415 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1416
1417 11. When the -o option was used, pcregrep was setting a return code of 1, even
1418 when matches were found, and --line-buffered was not being honoured.
1419
1420 12. Added an optional parentheses number to the -o and --only-matching options
1421 of pcregrep.
1422
1423 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1424 can match an empty string. The code to do it in pcretest and pcredemo
1425 needed fixing:
1426
1427 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1428 only one byte after an empty string match just before CRLF (this case
1429 just got forgotten; "any" and "anycrlf" were OK).
1430
1431 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1432 mode when an empty string match preceded an ASCII character followed by
1433 a non-ASCII character. (The code for advancing by one character rather
1434 than one byte was nonsense.)
1435
1436 (c) The pcredemo.c sample program did not have any code at all to handle
1437 the cases when CRLF is a valid newline sequence.
1438
1439 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1440 as a starting offset was within the subject string. There is now a new
1441 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1442 negative or greater than the length of the string. In order to test this,
1443 pcretest is extended to allow the setting of negative starting offsets.
1444
1445 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1446 starting offset points to the beginning of a UTF-8 character was
1447 unnecessarily clumsy. I tidied it up.
1448
1449 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1450 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1451
1452 17. Nobody had reported that the --include_dir option, which was added in
1453 release 7.7 should have been called --include-dir (hyphen, not underscore)
1454 for compatibility with GNU grep. I have changed it to --include-dir, but
1455 left --include_dir as an undocumented synonym, and the same for
1456 --exclude-dir, though that is not available in GNU grep, at least as of
1457 release 2.5.4.
1458
1459 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1460 characters from a string of bytes) have been redefined so as not to use
1461 loops, in order to improve performance in some environments. At the same
1462 time, I abstracted some of the common code into auxiliary macros to save
1463 repetition (this should not affect the compiled code).
1464
1465 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1466 compile-time error is now given if \c is not followed by an ASCII
1467 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1468 different, and any byte value is allowed.)
1469
1470 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1471 START_OPTIMIZE option, which is now allowed at compile time - but just
1472 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1473 to pcregrep and other applications that have no direct access to PCRE
1474 options. The new /Y option in pcretest sets this option when calling
1475 pcre_compile().
1476
1477 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1478 back references. Groups containing recursive back references were forced to
1479 be atomic by that change, but in the case of named groups, the amount of
1480 memory required was incorrectly computed, leading to "Failed: internal
1481 error: code overflow". This has been fixed.
1482
1483 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1484 pcretest.c, to avoid build problems in some Borland environments.
1485
1486
1487 Version 8.10 25-Jun-2010
1488 ------------------------
1489
1490 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1491 THEN.
1492
1493 2. (*ACCEPT) was not working when inside an atomic group.
1494
1495 3. Inside a character class, \B is treated as a literal by default, but
1496 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1497 causes the error). The code is unchanged, but I tidied the documentation.
1498
1499 4. Inside a character class, PCRE always treated \R and \X as literals,
1500 whereas Perl faults them if its -w option is set. I have changed PCRE so
1501 that it faults them when PCRE_EXTRA is set.
1502
1503 5. Added support for \N, which always matches any character other than
1504 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1505
1506 6. When compiling pcregrep with newer versions of gcc which may have
1507 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1508 declared with attribute warn_unused_result" were given. Just casting the
1509 result to (void) does not stop the warnings; a more elaborate fudge is
1510 needed. I've used a macro to implement this.
1511
1512 7. Minor change to pcretest.c to avoid a compiler warning.
1513
1514 8. Added four artifical Unicode properties to help with an option to make
1515 \s etc use properties (see next item). The new properties are: Xan
1516 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1517
1518 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1519 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1520 this option. Modified pcretest to add /W to test this facility. Added
1521 REG_UCP to make it available via the POSIX interface.
1522
1523 10. Added --line-buffered to pcregrep.
1524
1525 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1526 studied, and the match started with a letter with a code point greater than
1527 127 whose first byte was different to the first byte of the other case of
1528 the letter, the other case of this starting letter was not recognized
1529 (#976).
1530
1531 12. If a pattern that was studied started with a repeated Unicode property
1532 test, for example, \p{Nd}+, there was the theoretical possibility of
1533 setting up an incorrect bitmap of starting bytes, but fortunately it could
1534 not have actually happened in practice until change 8 above was made (it
1535 added property types that matched character-matching opcodes).
1536
1537 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1538 possible starting bytes for non-anchored patterns.
1539
1540 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1541 \R, and also a number of cases that involve Unicode properties, both
1542 explicit and implicit when PCRE_UCP is set.
1543
1544 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1545 input, it could crash or give wrong results if characters with values
1546 greater than 0xc0 were present in the subject string. (Detail: it assumed
1547 UTF-8 input when processing these items.)
1548
1549 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1550 size_t is 64-bit (#991).
1551
1552 17. Added a check for running out of memory when PCRE is compiled with
1553 --disable-stack-for-recursion (#990).
1554
1555 18. If the last data line in a file for pcretest does not have a newline on
1556 the end, a newline was missing in the output.
1557
1558 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1559 less than 128) in its various bitmaps. However, there is a facility for
1560 generating tables according to the current locale when PCRE is compiled. It
1561 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1562 characters, are recognized by isspace() and therefore were getting set in
1563 these tables, and indeed these tables seem to approximate to ISO 8859. This
1564 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1565 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1566 which of course cannot start UTF-8 characters. I have changed the code so
1567 that only real ASCII characters (less than 128) and the correct starting
1568 bytes for UTF-8 encodings are set for characters greater than 127 when in
1569 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1570 altogether.)
1571
1572 20. Added the /T option to pcretest so as to be able to run tests with non-
1573 standard character tables, thus making it possible to include the tests
1574 used for 19 above in the standard set of tests.
1575
1576 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1577 reference to a subpattern the other side of a comment that contains an
1578 opening parenthesis caused either an internal compiling error, or a
1579 reference to the wrong subpattern.
1580
1581
1582 Version 8.02 19-Mar-2010
1583 ------------------------
1584
1585 1. The Unicode data tables have been updated to Unicode 5.2.0.
1586
1587 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1588 configured.
1589
1590 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1591 original author of that file, following a query about its status.
1592
1593 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1594 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1595
1596 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1597 quantifier applied to a forward-referencing subroutine call, could compile
1598 incorrect code or give the error "internal error: previously-checked
1599 referenced subpattern not found".
1600
1601 6. Both MS Visual Studio and Symbian OS have problems with initializing
1602 variables to point to external functions. For these systems, therefore,
1603 pcre_malloc etc. are now initialized to local functions that call the
1604 relevant global functions.
1605
1606 7. There were two entries missing in the vectors called coptable and poptable
1607 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1608 I've fixed the data, and added a kludgy way of testing at compile time that
1609 the lengths are correct (equal to the number of opcodes).
1610
1611 8. Following on from 7, I added a similar kludge to check the length of the
1612 eint vector in pcreposix.c.
1613
1614 9. Error texts for pcre_compile() are held as one long string to avoid too
1615 much relocation at load time. To find a text, the string is searched,
1616 counting zeros. There was no check for running off the end of the string,
1617 which could happen if a new error number was added without updating the
1618 string.
1619
1620 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1621
1622 11. \K was not working if it appeared in an atomic group or in a group that
1623 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1624 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1625 the assertion is positive, but not if it is negative.
1626
1627 12. Change 11 fortuitously reduced the size of the stack frame used in the
1628 "match()" function of pcre_exec.c by one pointer. Forthcoming
1629 implementation of support for (*MARK) will need an extra pointer on the
1630 stack; I have reserved it now, so that the stack frame size does not
1631 decrease.
1632
1633 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1634 item in branch that calls a recursion is a subroutine call - as in the
1635 second branch in the above example - was incorrectly given the compile-
1636 time error "recursive call could loop indefinitely" because pcre_compile()
1637 was not correctly checking the subroutine for matching a non-empty string.
1638
1639 14. The checks for overrunning compiling workspace could trigger after an
1640 overrun had occurred. This is a "should never occur" error, but it can be
1641 triggered by pathological patterns such as hundreds of nested parentheses.
1642 The checks now trigger 100 bytes before the end of the workspace.
1643
1644 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1645
1646
1647 Version 8.01 19-Jan-2010
1648 ------------------------
1649
1650 1. If a pattern contained a conditional subpattern with only one branch (in
1651 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1652 computed the wrong minimum data length (which is of course zero for such
1653 subpatterns). This could cause incorrect "no match" results.
1654
1655 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1656 the pattern is reset in the first branch, pcre_compile() failed with
1657 "internal error: code overflow at offset...". This happened only when
1658 the reset was to the original external option setting. (An optimization
1659 abstracts leading options settings into an external setting, which was the
1660 cause of this.)
1661
1662 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1663 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1664 assertion pattern did not match (meaning that the assertion was true), it
1665 was incorrectly treated as false if the SKIP had been reached during the
1666 matching. This also applied to assertions used as conditions.
1667
1668 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1669 assertion subpattern, including such a pattern used as a condition,
1670 unpredictable results occurred, instead of the error return
1671 PCRE_ERROR_DFA_UITEM.
1672
1673 5. The C++ GlobalReplace function was not working like Perl for the special
1674 situation when an empty string is matched. It now does the fancy magic
1675 stuff that is necessary.
1676
1677 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1678 removed. (These were left over from very, very early versions of PCRE.)
1679
1680 7. Some cosmetic changes to the code to make life easier when compiling it
1681 as part of something else:
1682
1683 (a) Change DEBUG to PCRE_DEBUG.
1684
1685 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1686 called "current" as "current_branch", to prevent a collision with the
1687 Linux macro when compiled as a kernel module.
1688
1689 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1690 prevent a collision with the Linux macro when compiled as a kernel
1691 module.
1692
1693 8. In pcre_compile() there are some checks for integer overflows that used to
1694 cast potentially large values to (double). This has been changed to that
1695 when building, a check for int64_t is made, and if it is found, it is used
1696 instead, thus avoiding the use of floating point arithmetic. (There is no
1697 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1698 double.
1699
1700 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1701 2005 (difference between two addresses compared to an unsigned value).
1702
1703 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1704 custom one, because of the following reported problem in Windows:
1705
1706 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1707 under Win32.
1708 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1709 therefore missing the function definition.
1710 - The compiler thus generates a "C" signature for the test function.
1711 - The linker fails to find the "C" function.
1712 - PCRE fails to configure if asked to do so against libbz2.
1713
1714 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1715 messages were output:
1716
1717 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1718 rerunning libtoolize, to keep the correct libtool macros in-tree.
1719 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1720
1721 I have done both of these things.
1722
1723 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1724 most of the time, it *can* run out if it is given a pattern that contains a
1725 runaway infinite recursion. I updated the discussion in the pcrestack man
1726 page.
1727
1728 13. Now that we have gone to the x.xx style of version numbers, the minor
1729 version may start with zero. Using 08 or 09 is a bad idea because users
1730 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1731 interpreted as invalid octal numbers. I've updated the previous comment in
1732 configure.ac, and also added a check that gives an error if 08 or 09 are
1733 used.
1734
1735 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1736 causing partial matching to fail when the end of the subject matched \W
1737 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1738
1739 15. There were some discrepancies between the declarations in pcre_internal.h
1740 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1741 their definitions. The declarations used "const uschar *" and the
1742 definitions used USPTR. Even though USPTR is normally defined as "const
1743 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1744 reported that: "This difference in casting confuses some C++ compilers, for
1745 example, SunCC recognizes above declarations as different functions and
1746 generates broken code for hbpcre." I have changed the declarations to use
1747 USPTR.
1748
1749 16. GNU libtool is named differently on some systems. The autogen.sh script now
1750 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1751 (FreeBSD).
1752
1753 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1754 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1755 comment: "Figure out how to create a longlong from a string: strtoll and
1756 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1757 instance, but it only takes 2 args instead of 3!"
1758
1759 18. A subtle bug concerned with back references has been fixed by a change of
1760 specification, with a corresponding code fix. A pattern such as
1761 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1762 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1763 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1764 same bug. Such groups have to be quantified to be useful, or contained
1765 inside another quantified group. (If there's no repetition, the reference
1766 can never match.) The problem arises because, having left the group and
1767 moved on to the rest of the pattern, a later failure that backtracks into
1768 the group uses the captured value from the final iteration of the group
1769 rather than the correct earlier one. I have fixed this in PCRE by forcing
1770 any group that contains a reference to itself to be an atomic group; that
1771 is, there cannot be any backtracking into it once it has completed. This is
1772 similar to recursive and subroutine calls.
1773
1774
1775 Version 8.00 19-Oct-09
1776 ----------------------
1777
1778 1. The table for translating pcre_compile() error codes into POSIX error codes
1779 was out-of-date, and there was no check on the pcre_compile() error code
1780 being within the table. This could lead to an OK return being given in
1781 error.
1782
1783 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1784 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1785 in a Windows environment.
1786
1787 3. The pcregrep --count option prints the count for each file even when it is
1788 zero, as does GNU grep. However, pcregrep was also printing all files when
1789 --files-with-matches was added. Now, when both options are given, it prints
1790 counts only for those files that have at least one match. (GNU grep just
1791 prints the file name in this circumstance, but including the count seems
1792 more useful - otherwise, why use --count?) Also ensured that the
1793 combination -clh just lists non-zero counts, with no names.
1794
1795 4. The long form of the pcregrep -F option was incorrectly implemented as
1796 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1797 but it seems right to fix it, and I didn't think it was worth preserving
1798 the old behaviour.
1799
1800 5. The command line items --regex=pattern and --regexp=pattern were not
1801 recognized by pcregrep, which required --regex pattern or --regexp pattern
1802 (with a space rather than an '='). The man page documented the '=' forms,
1803 which are compatible with GNU grep; these now work.
1804
1805 6. No libpcreposix.pc file was created for pkg-config; there was just
1806 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1807
1808 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1809 when UCP support is not needed, by modifying the Python script that
1810 generates it from Unicode data files. This should not matter if the module
1811 is correctly used as a library, but I received one complaint about 50K of
1812 unwanted data. My guess is that the person linked everything into his
1813 program rather than using a library. Anyway, it does no harm.
1814
1815 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1816 was a minimum greater than 1 for a wide character in a possessive
1817 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1818 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1819 character. Chaos in the form of incorrect output or a compiling loop could
1820 result.
1821
1822 9. The restrictions on what a pattern can contain when partial matching is
1823 requested for pcre_exec() have been removed. All patterns can now be
1824 partially matched by this function. In addition, if there are at least two
1825 slots in the offset vector, the offset of the earliest inspected character
1826 for the match and the offset of the end of the subject are set in them when
1827 PCRE_ERROR_PARTIAL is returned.
1828
1829 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1830 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1831 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1832 and may be more useful for multi-segment matching.
1833
1834 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1835 used to be given if ever the end of the subject was reached; now it is
1836 given only if matching could not proceed because another character was
1837 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1838 string "Z", which now yields "no match" instead of "partial match". In the
1839 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1840 final character ended with (*FAIL).
1841
1842 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1843 if the pattern had a "must contain" character that was already found in the
1844 earlier partial match, unless partial matching was again requested. For
1845 example, with the pattern /dog.(body)?/, the "must contain" character is
1846 "g". If the first part-match was for the string "dog", restarting with
1847 "sbody" failed. This bug has been fixed.
1848
1849 13. The string returned by pcre_dfa_exec() after a partial match has been
1850 changed so that it starts at the first inspected character rather than the
1851 first character of the match. This makes a difference only if the pattern
1852 starts with a lookbehind assertion or \b or \B (\K is not supported by
1853 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1854 matching functions compatible, and I think it's the right thing to do.
1855
1856 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1857 so that the demonstration program is easily available in environments where
1858 PCRE has not been installed from source.
1859
1860 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1861 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1862 library.
1863
1864 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1865 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1866 is not the first non-POSIX option to be added. Clearly some people find
1867 these options useful.
1868
1869 17. If a caller to the POSIX matching function regexec() passes a non-zero
1870 value for nmatch with a NULL value for pmatch, the value of
1871 nmatch is forced to zero.
1872
1873 18. RunGrepTest did not have a test for the availability of the -u option of
1874 the diff command, as RunTest does. It now checks in the same way as
1875 RunTest, and also checks for the -b option.
1876
1877 19. If an odd number of negated classes containing just a single character
1878 interposed, within parentheses, between a forward reference to a named
1879 subpattern and the definition of the subpattern, compilation crashed with
1880 an internal error, complaining that it could not find the referenced
1881 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1882 [The bug was that it was starting one character too far in when skipping
1883 over the character class, thus treating the ] as data rather than
1884 terminating the class. This meant it could skip too much.]
1885
1886 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1887 /g option in pcretest when the pattern contains \K, which makes it possible
1888 to have an empty string match not at the start, even when the pattern is
1889 anchored. Updated pcretest and pcredemo to use this option.
1890
1891 21. If the maximum number of capturing subpatterns in a recursion was greater
1892 than the maximum at the outer level, the higher number was returned, but
1893 with unset values at the outer level. The correct (outer level) value is
1894 now given.
1895
1896 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1897 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1898 make it do so. The string so far is captured, making this feature
1899 compatible with Perl.
1900
1901 23. The tests have been re-organized, adding tests 11 and 12, to make it
1902 possible to check the Perl 5.10 features against Perl 5.10.
1903
1904 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1905 pattern matches a fixed length string. PCRE did not allow this; now it
1906 does. Neither allows recursion.
1907
1908 25. I finally figured out how to implement a request to provide the minimum
1909 length of subject string that was needed in order to match a given pattern.
1910 (It was back references and recursion that I had previously got hung up
1911 on.) This code has now been added to pcre_study(); it finds a lower bound
1912 to the length of subject needed. It is not necessarily the greatest lower
1913 bound, but using it to avoid searching strings that are too short does give
1914 some useful speed-ups. The value is available to calling programs via
1915 pcre_fullinfo().
1916
1917 26. While implementing 25, I discovered to my embarrassment that pcretest had
1918 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1919 study optimizations had never been tested with that matching function.
1920 Oops. What is worse, even when it was passed study data, there was a bug in
1921 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1922 were also very few tests of studied patterns with pcre_dfa_exec().
1923
1924 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1925 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1926 on the other side of the coin, they are no longer allowed to have different
1927 names, because these cannot be distinguished in PCRE, and this has caused
1928 confusion. (This is a difference from Perl.)
1929
1930 28. When duplicate subpattern names are present (necessarily with different
1931 numbers, as required by 27 above), and a test is made by name in a
1932 conditional pattern, either for a subpattern having been matched, or for
1933 recursion in such a pattern, all the associated numbered subpatterns are
1934 tested, and the overall condition is true if the condition is true for any
1935 one of them. This is the way Perl works, and is also more like the way
1936 testing by number works.
1937
1938
1939 Version 7.9 11-Apr-09
1940 ---------------------
1941
1942 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1943 (pcretest), all targets were linked against these libraries. This included
1944 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1945 libraries. This caused unwanted dependencies to be created. This problem
1946 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1947 pcretest is linked with readline.
1948
1949 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1950 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1951 moved outside it again, because FALSE and TRUE are already defined in AIX,
1952 but BOOL is not.
1953
1954 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1955 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1956
1957 4. The pcregrep documentation said spaces were inserted as well as colons (or
1958 hyphens) following file names and line numbers when outputting matching
1959 lines. This is not true; no spaces are inserted. I have also clarified the
1960 wording for the --colour (or --color) option.
1961
1962 5. In pcregrep, when --colour was used with -o, the list of matching strings
1963 was not coloured; this is different to GNU grep, so I have changed it to be
1964 the same.
1965
1966 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1967 each matching line was coloured. Now it goes on to look for further matches
1968 of any of the test patterns, which is the same behaviour as GNU grep.
1969
1970 7. A pattern that could match an empty string could cause pcregrep to loop; it
1971 doesn't make sense to accept an empty string match in pcregrep, so I have
1972 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1973 seems to be how GNU grep behaves. [But see later change 40 for release
1974 8.33.]
1975
1976 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1977 start or after a newline", because the conditional assertion was not being
1978 correctly handled. The rule now is that both the assertion and what follows
1979 in the first alternative must satisfy the test.
1980
1981 9. If auto-callout was enabled in a pattern with a conditional group whose
1982 condition was an assertion, PCRE could crash during matching, both with
1983 pcre_exec() and pcre_dfa_exec().
1984
1985 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1986 used for matching.
1987
1988 11. Unicode property support in character classes was not working for
1989 characters (bytes) greater than 127 when not in UTF-8 mode.
1990
1991 12. Added the -M command line option to pcretest.
1992
1993 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1994
1995 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1996
1997 16. Added comments and documentation about mis-use of no_arg in the C++
1998 wrapper.
1999
2000 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2001 from Martin Jerabek that uses macro names for all relevant character and
2002 string constants.
2003
2004 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2005 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2006 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2007 these, but not everybody uses configure.
2008
2009 19. A conditional group that had only one branch was not being correctly
2010 recognized as an item that could match an empty string. This meant that an
2011 enclosing group might also not be so recognized, causing infinite looping
2012 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2013 with the subject "ab", where knowledge that the repeated group can match
2014 nothing is needed in order to break the loop.
2015
2016 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2017 exec(), but without supplying a callout function, matching went wrong.
2018
2019 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2020 leak if the size of the offset vector was greater than 30. When the vector
2021 is smaller, the saved offsets during recursion go onto a local stack
2022 vector, but for larger vectors malloc() is used. It was failing to free
2023 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2024 error, in fact).
2025
2026 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2027 heapframe that is used only when UTF-8 support is enabled. This caused no
2028 problem, but was untidy.
2029
2030 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2031 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2032 included within another project.
2033
2034 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2035 slightly modified by me:
2036
2037 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2038 not building pcregrep.
2039
2040 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2041 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2042
2043 25. Forward references, both numeric and by name, in patterns that made use of
2044 duplicate group numbers, could behave incorrectly or give incorrect errors,
2045 because when scanning forward to find the reference group, PCRE was not
2046 taking into account the duplicate group numbers. A pattern such as
2047 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2048
2049 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2050 the feature of a custom pointer more persuasive (as requested by a user).
2051
2052 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2053 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2054 pre-defined.
2055
2056 28. Added support for (*UTF8) at the start of a pattern.
2057
2058 29. Arrange for flags added by the "release type" setting in CMake to be shown
2059 in the configuration summary.
2060
2061
2062 Version 7.8 05-Sep-08
2063 ---------------------
2064
2065 1. Replaced UCP searching code with optimized version as implemented for Ad
2066 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2067 stage table and inline lookup instead of a function, giving speed ups of 2
2068 to 5 times on some simple patterns that I tested. Permission was given to
2069 distribute the MultiStage2.py script that generates the tables (it's not in
2070 the tarball, but is in the Subversion repository).
2071
2072 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2073 scripts.
2074
2075 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2076 a group with a zero qualifier. The result of the study could be incorrect,
2077 or the function might crash, depending on the pattern.
2078
2079 4. Caseless matching was not working for non-ASCII characters in back
2080 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2081 It now works when Unicode Property Support is available.
2082
2083 5. In pcretest, an escape such as \x{de} in the data was always generating
2084 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2085 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2086 truncation.
2087
2088 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2089
2090 7. Added two (int) casts to pcregrep when printing the difference of two
2091 pointers, in case they are 64-bit values.
2092
2093 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2094 test 2 if it fails.
2095
2096 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2097 and a #define of that name to empty if it is not externally set. This is to
2098 allow users of MSVC to set it if necessary.
2099
2100 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2101 the convenience functions in the pcre_get.c source file.
2102
2103 11. An option change at the start of a pattern that had top-level alternatives
2104 could cause overwriting and/or a crash. This command provoked a crash in
2105 some environments:
2106
2107 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2108
2109 This potential security problem was recorded as CVE-2008-2371.
2110
2111 12. For a pattern where the match had to start at the beginning or immediately
2112 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2113 pcre_dfa_exec() could read past the end of the passed subject if there was
2114 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2115 pcretest so that it places the subject at the end of its malloc-ed buffer.
2116
2117 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2118 exec() might read past the end of the data buffer in UTF-8 mode.
2119
2120 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2121 the data contained the byte 0x85 as part of a UTF-8 character within its
2122 first line. This applied both to normal and DFA matching.
2123
2124 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2125 /^[^d]*?$/8 failed to match "abc".
2126
2127 16. Added a missing copyright notice to pcrecpp_internal.h.
2128
2129 17. Make it more clear in the documentation that values returned from
2130 pcre_exec() in ovector are byte offsets, not character counts.
2131
2132 18. Tidied a few places to stop certain compilers from issuing warnings.
2133
2134 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2135 supplied by Stefan Weber. I made a further small update for 7.8 because
2136 there is a change of source arrangements: the pcre_searchfuncs.c module is
2137 replaced by pcre_ucd.c.
2138
2139
2140 Version 7.7 07-May-08
2141 ---------------------
2142
2143 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2144 a string to a long long, pretend we don't even have a long long." This is
2145 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2146
2147 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2148 pre-7.6 versions, which defined a global no_arg variable instead of putting
2149 it in the RE class. (See also #8 below.)
2150
2151 3. Remove a line of dead code, identified by coverity and reported by Nuno
2152 Lopes.
2153
2154 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2155
2156 (1) The include/exclude patterns were being applied to the whole pathnames
2157 of files, instead of just to the final components.
2158
2159 (2) If there was more than one level of directory, the subdirectories were
2160 skipped unless they satisfied the include/exclude conditions. This is
2161 inconsistent with GNU grep (and could even be seen as contrary to the
2162 pcregrep specification - which I improved to make it absolutely clear).
2163 The action now is always to scan all levels of directory, and just
2164 apply the include/exclude patterns to regular files.
2165
2166 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2167 --exclude_dir in the tests to avoid scanning .svn directories.
2168
2169 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2170 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2171 doesn't support NULs in patterns.
2172
2173 7. Added some missing "const"s to declarations of static tables in
2174 pcre_compile.c and pcre_dfa_exec.c.
2175
2176 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2177 caused by fix #2 above. (Subsequently also a second patch to fix the
2178 first patch. And a third patch - this was a messy problem.)
2179
2180 9. Applied Craig's patch to remove the use of push_back().
2181
2182 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2183 matching function regexec().
2184
2185 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2186 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2187 references. PCRE supports relative numbers with this syntax (I don't think
2188 Oniguruma does).
2189
2190 12. Previously, a group with a zero repeat such as (...){0} was completely
2191 omitted from the compiled regex. However, this means that if the group
2192 was called as a subroutine from elsewhere in the pattern, things went wrong
2193 (an internal error was given). Such groups are now left in the compiled
2194 pattern, with a new opcode that causes them to be skipped at execution
2195 time.
2196
2197 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2198 to the way PCRE behaves:
2199
2200 (a) A lone ] character is dis-allowed (Perl treats it as data).
2201
2202 (b) A back reference to an unmatched subpattern matches an empty string
2203 (Perl fails the current match path).
2204
2205 (c) A data ] in a character class must be notated as \] because if the
2206 first data character in a class is ], it defines an empty class. (In
2207 Perl it is not possible to have an empty class.) The empty class []
2208 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2209 The negative empty class [^] matches any one character, independently
2210 of the DOTALL setting.
2211
2212 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2213 non-existent subpattern following a character class starting with ']' and
2214 containing () gave an internal compiling error instead of "reference to
2215 non-existent subpattern". Fortunately, when the pattern did exist, the
2216 compiled code was correct. (When scanning forwards to check for the
2217 existence of the subpattern, it was treating the data ']' as terminating
2218 the class, so got the count wrong. When actually compiling, the reference
2219 was subsequently set up correctly.)
2220
2221 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2222 it was being rejected as not supported by pcre_dfa_exec(), even though
2223 other assertions are supported. I have made pcre_dfa_exec() support
2224 (*FAIL).
2225
2226 16. The implementation of 13c above involved the invention of a new opcode,
2227 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2228 cannot be changed at match time, I realized I could make a small
2229 improvement to matching performance by compiling OP_ALLANY instead of
2230 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2231 on the OP_ANY path.
2232
2233 17. Compiling pcretest on Windows with readline support failed without the
2234 following two fixes: (1) Make the unistd.h include conditional on
2235 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2236
2237 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2238 ncurses library to be included for pcretest when ReadLine support is
2239 requested, but also to allow for it to be overridden. This patch came from
2240 Daniel Bergström.
2241
2242 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2243 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2244 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2245 spotting this.
2246
2247
2248 Version 7.6 28-Jan-08
2249 ---------------------
2250
2251 1. A character class containing a very large number of characters with
2252 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2253 overflow.
2254
2255 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2256 HAVE_LONG_LONG is not defined.
2257
2258 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2259 bring them up to date and include new features. This patch includes:
2260
2261 - Fixed PH's badly added libz and libbz2 support.
2262 - Fixed a problem with static linking.
2263 - Added pcredemo. [But later removed - see 7 below.]
2264 - Fixed dftables problem and added an option.
2265 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2266 HAVE_LONG_LONG.
2267 - Added readline support for pcretest.
2268 - Added an listing of the option settings after cmake has run.
2269
2270 4. A user submitted a patch to Makefile that makes it easy to create
2271 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2272 Makefile.am that cause it to include this special target, without
2273 affecting anything else. Note that the same mingw target plus all
2274 the other distribution libraries and programs are now supported
2275 when configuring with CMake (see 6 below) instead of with
2276 Configure/Make.
2277
2278 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2279 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2280 exported in the Windows port". It has not yet been confirmed that the patch
2281 solves the problem, but it does no harm.
2282
2283 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2284 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2285 with CMake, and also correct the comment about stack recursion.
2286
2287 7. Remove the automatic building of pcredemo from the ./configure system and
2288 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2289 of a program that users should build themselves after PCRE is installed, so
2290 building it automatically is not really right. What is more, it gave
2291 trouble in some build environments.
2292
2293 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2294
2295
2296 Version 7.5 10-Jan-08
2297 ---------------------
2298
2299 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2300 values in parens when parsing an RE using the C++ wrapper."
2301
2302 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2303 Characters greater than 255 were excluded from the class instead of being
2304 included.
2305
2306 3. The same bug as (2) above applied to negated POSIX classes such as
2307 [:^space:].
2308
2309 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2310 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2311 I have changed it.
2312
2313 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2314 first named subpattern) and a construct such as (?&a) would reference the
2315 first named subpattern whose name started with "a" (in other words, the
2316 length check was missing). Both these problems are fixed. "Subpattern name
2317 expected" is now given for (?&) (a zero-length name), and this patch also
2318 makes it give the same error for \k'' (previously it complained that that
2319 was a reference to a non-existent subpattern).
2320
2321 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2322 this is right because (?- can be followed by option settings as well as by
2323 digits. I have, however, made the messages clearer.
2324
2325 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2326 than the number used in the conditional) now cause a compile-time error.
2327 This is actually not compatible with Perl, which accepts such patterns, but
2328 treats the conditional as always being FALSE (as PCRE used to), but it
2329 seems to me that giving a diagnostic is better.
2330
2331 8. Change "alphameric" to the more common word "alphanumeric" in comments
2332 and messages.
2333
2334 9. Fix two occurrences of "backslash" in comments that should have been
2335 "backspace".
2336
2337 10. Remove two redundant lines of code that can never be obeyed (their function
2338 was moved elsewhere).
2339
2340 11. The program that makes PCRE's Unicode character property table had a bug
2341 which caused it to generate incorrect table entries for sequences of
2342 characters that have the same character type, but are in different scripts.
2343 It amalgamated them into a single range, with the script of the first of
2344 them. In other words, some characters were in the wrong script. There were
2345 thirteen such cases, affecting characters in the following ranges:
2346
2347 U+002b0 - U+002c1
2348 U+0060c - U+0060d
2349 U+0061e - U+00612
2350 U+0064b - U+0065e
2351 U+0074d - U+0076d
2352 U+01800 - U+01805
2353 U+01d00 - U+01d77
2354 U+01d9b - U+01dbf
2355 U+0200b - U+0200f
2356 U+030fc - U+030fe
2357 U+03260 - U+0327f
2358 U+0fb46 - U+0fbb1
2359 U+10450 - U+1049d
2360
2361 12. The -o option (show only the matching part of a line) for pcregrep was not
2362 compatible with GNU grep in that, if there was more than one match in a
2363 line, it showed only the first of them. It now behaves in the same way as
2364 GNU grep.
2365
2366 13. If the -o and -v options were combined for pcregrep, it printed a blank
2367 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2368 does the same. The return code can be used to tell if there were any
2369 non-matching lines.
2370
2371 14. Added --file-offsets and --line-offsets to pcregrep.
2372
2373 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2374 infinitely looping recursion. The bug was that positive lookaheads were not
2375 being skipped when checking for a possible empty match (negative lookaheads
2376 and both kinds of lookbehind were skipped).
2377
2378 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2379 inclusion of <windows.h> to before rather than after the definition of
2380 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2381
2382 17. Specifying a possessive quantifier with a specific limit for a Unicode
2383 character property caused pcre_compile() to compile bad code, which led at
2384 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2385 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2386 caused the error; without that there was no problem.
2387
2388 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2389
2390 19. Added --enable-pcretest-libreadline.
2391
2392 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2393 RE::GlobalReplace(). As a result, the number of replacements returned was
2394 double what it should be. I removed one of the increments, but Craig sent a
2395 later patch that removed the other one (the right fix) and added unit tests
2396 that check the return values (which was not done before).
2397
2398 21. Several CMake things:
2399
2400 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2401 the names libpcre and libpcreposix, not just pcre and pcreposix.
2402
2403 (2) The above change means that pcretest and pcregrep are now correctly
2404 linked with the newly-built libraries, not previously installed ones.
2405
2406 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2407
2408 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2409 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2410 UTF-8 newline character). The key issue is that the pattern starts .*;
2411 this means that the match must be either at the beginning, or after a
2412 newline. The bug was in the code for advancing after a failed match and
2413 checking that the new position followed a newline. It was not taking
2414 account of UTF-8 characters correctly.
2415
2416 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2417 character classes. PCRE was not treating the sequence [:...:] as a
2418 character class unless the ... were all letters. Perl, however, seems to
2419 allow any characters between [: and :], though of course it rejects as
2420 unknown any "names" that contain non-letters, because all the known class
2421 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2422 for example, whereas PCRE did not - it did not recognize a POSIX character
2423 class. This seemed a bit dangerous, so the code has been changed to be
2424 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2425 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2426 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2427 Perl does, and where it didn't before.
2428
2429 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2430 Windows environments %n is disabled by default.
2431
2432
2433 Version 7.4 21-Sep-07
2434 ---------------------
2435
2436 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2437 means that a class such as [\s] counted as "explicit reference to CR or
2438 LF". That isn't really right - the whole point of the change was to try to
2439 help when there was an actual mention of one of the two characters. So now
2440 the change happens only if \r or \n (or a literal CR or LF) character is
2441 encountered.
2442
2443 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2444 of both had grown to the point where there were only 3 bits left.
2445 Fortunately, there was spare space in the data structure, and so I have
2446 moved the internal flags into a new 16-bit field to free up more option
2447 bits.
2448
2449 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2450 but did not set the internal JCHANGED flag - either of these is enough to
2451 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2452 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2453 start sets both bits.
2454
2455 4. Added options (at build time, compile time, exec time) to change \R from
2456 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2457
2458 5. doc/pcresyntax.html was missing from the distribution.
2459
2460 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2461 compatibility, even though it is no longer used.
2462
2463 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2464 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2465 windows.h file is present (where different names are used). [This was
2466 reversed later after testing - see 16 below.]
2467
2468 8. Changed all #include <config.h> to #include "config.h". There were also
2469 some further <pcre.h> cases that I changed to "pcre.h".
2470
2471 9. When pcregrep was used with the --colour option, it missed the line ending
2472 sequence off the lines that it output.
2473
2474 10. It was pointed out to me that arrays of string pointers cause lots of
2475 relocations when a shared library is dynamically loaded. A technique of
2476 using a single long string with a table of offsets can drastically reduce
2477 these. I have refactored PCRE in four places to do this. The result is
2478 dramatic:
2479
2480 Originally: 290
2481 After changing UCP table: 187
2482 After changing error message table: 43
2483 After changing table of "verbs" 36
2484 After changing table of Posix names 22
2485
2486 Thanks to the folks working on Gregex for glib for this insight.
2487
2488 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2489 unicode-properties was also set.
2490
2491 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2492
2493 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2494 checked only for CRLF.
2495
2496 14. Added casts to pcretest.c to avoid compiler warnings.
2497
2498 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2499
2500 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2501 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2502 entirely. This removes changes made in 7 above.
2503
2504 17. The CMake files have been updated, and there is now more information about
2505 building with CMake in the NON-UNIX-USE document.
2506
2507
2508 Version 7.3 28-Aug-07
2509 ---------------------
2510
2511 1. In the rejigging of the build system that eventually resulted in 7.1, the
2512 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2513 brackets there is not right, since it causes compilers to look for an
2514 installed pcre.h, not the version that is in the source that is being
2515 compiled (which of course may be different). I have changed it back to:
2516
2517 #include "pcre.h"
2518
2519 I have a vague recollection that the change was concerned with compiling in
2520 different directories, but in the new build system, that is taken care of
2521 by the VPATH setting the Makefile.
2522
2523 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2524 when the subject happened to end in the byte 0x85 (e.g. if the last
2525 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2526 characters but of course it shouldn't be taken as a newline when it is part
2527 of another character. The bug was that, for an unlimited repeat of . in
2528 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2529 characters when looking for a newline.
2530
2531 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2532
2533 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2534 in debug output.
2535
2536 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2537 long printing in the pcrecpp unittest when running under MinGW.
2538
2539 6. ESC_K was left out of the EBCDIC table.
2540
2541 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2542 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2543 limit also applies to "virtual nesting" when a pattern is recursive, and in
2544 this case 1000 isn't so big. I have been able to remove this limit at the
2545 expense of backing off one optimization in certain circumstances. Normally,
2546 when pcre_exec() would call its internal match() function recursively and
2547 immediately return the result unconditionally, it uses a "tail recursion"
2548 feature to save stack. However, when a subpattern that can match an empty
2549 string has an unlimited repetition quantifier, it no longer makes this
2550 optimization. That gives it a stack frame in which to save the data for
2551 checking that an empty string has been matched. Previously this was taken
2552 from the 1000-entry workspace that had been reserved. So now there is no
2553 explicit limit, but more stack is used.
2554
2555 8. Applied Daniel's patches to solve problems with the import/export magic
2556 syntax that is required for Windows, and which was going wrong for the
2557 pcreposix and pcrecpp parts of the library. These were overlooked when this
2558 problem was solved for the main library.
2559
2560 9. There were some crude static tests to avoid integer overflow when computing
2561 the size of patterns that contain repeated groups with explicit upper
2562 limits. As the maximum quantifier is 65535, the maximum group length was
2563 set at 30,000 so that the product of these two numbers did not overflow a
2564 32-bit integer. However, it turns out that people want to use groups that
2565 are longer than 30,000 bytes (though not repeat them that many times).
2566 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2567 made it possible to implement the integer overflow checks in a much more
2568 dynamic way, which I have now done. The artificial limitation on group
2569 length has been removed - we now have only the limit on the total length of
2570 the compiled pattern, which depends on the LINK_SIZE setting.
2571
2572 10. Fixed a bug in the documentation for get/copy named substring when
2573 duplicate names are permitted. If none of the named substrings are set, the
2574 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2575 empty string.
2576
2577 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2578 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2579 because the ] is interpreted as the first data character and the
2580 terminating ] is not found. PCRE has been made compatible with Perl in this
2581 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2582 cause memory overwriting.
2583
2584 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2585 string has been matched (to stop an infinite loop). It was not recognizing
2586 a conditional subpattern that could match an empty string if that
2587 subpattern was within another subpattern. For example, it looped when
2588 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2589 condition was not nested. This bug has been fixed.
2590
2591 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2592 past the start of the subject in the presence of bytes with the top bit
2593 set, for example "\x8aBCD".
2594
2595 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2596 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2597
2598 14. Optimized (?!) to (*FAIL).
2599
2600 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2601 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2602 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2603 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2604 does: it's just the validity check that is more restrictive.
2605
2606 16. Inserted checks for integer overflows during escape sequence (backslash)
2607 processing, and also fixed erroneous offset values for syntax errors during
2608 backslash processing.
2609
2610 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2611 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2612
2613 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2614 caused an overrun.
2615
2616 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2617 something other than just ASCII characters) inside a group that had an
2618 unlimited repeat caused a loop at compile time (while checking to see
2619 whether the group could match an empty string).
2620
2621 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2622 [\P{Any}] did so. (Error in the code for printing property names.)
2623
2624 21. An orphan \E inside a character class could cause a crash.
2625
2626 22. A repeated capturing bracket such as (A)? could cause a wild memory
2627 reference during compilation.
2628
2629 23. There are several functions in pcre_compile() that scan along a compiled
2630 expression for various reasons (e.g. to see if it's fixed length for look
2631 behind). There were bugs in these functions when a repeated \p or \P was
2632 present in the pattern. These operators have additional parameters compared
2633 with \d, etc, and these were not being taken into account when moving along
2634 the compiled data. Specifically:
2635
2636 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2637 length.
2638
2639 (b) An item such as \pL+ within a repeated group could cause crashes or
2640 loops.
2641
2642 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2643 "reference to non-existent subpattern" error.
2644
2645 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2646
2647 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2648 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2649
2650 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2651
2652 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2653 character were causing crashes (broken optimization).
2654
2655 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2656 \p or \P) caused a compile-time loop.
2657
2658 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2659 break. For example, the unstudied pattern [\r\n]A does not match the string
2660 "\r\nA" because change 7.0/46 below moves the current point on by two
2661 characters after failing to match at the start. However, the pattern \nA
2662 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2663 the same is true. There doesn't seem any very clean way out of this, but
2664 what I have chosen to do makes the common cases work: PCRE now takes note
2665 of whether there can be an explicit match for \r or \n anywhere in the
2666 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2667 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2668 pattern has explicit CR or LF references.
2669
2670 29. Added (*CR) etc for changing newline setting at start of pattern.
2671
2672
2673 Version 7.2 19-Jun-07
2674 ---------------------
2675
2676 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2677 which is apparently normally available under Windows.
2678
2679 2. Re-jig the pcregrep tests with different newline settings in an attempt
2680 to make them independent of the local environment's newline setting.
2681
2682 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2683
2684 4. Some of the "internals" tests were previously cut out when the link size
2685 was not 2, because the output contained actual offsets. The recent new
2686 "Z" feature of pcretest means that these can be cut out, making the tests
2687 usable with all link sizes.
2688
2689 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2690 stack recursion. This gives a massive performance boost under BSD, but just
2691 a small improvement under Linux. However, it saves one field in the frame
2692 in all cases.
2693
2694 6. Added more features from the forthcoming Perl 5.10:
2695
2696 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2697 recursion call. It refers to the nth most recently opened parentheses.
2698
2699 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2700 to be opened parentheses.
2701
2702 (c) Conditions that refer to capturing parentheses can be specified
2703 relatively, for example, (?(-2)... or (?(+3)...
2704
2705 (d) \K resets the start of the current match so that everything before
2706 is not part of it.
2707
2708 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2709
2710 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2711 reference syntax.
2712
2713 (g) (?| introduces a group in which the numbering of parentheses in each
2714 alternative starts with the same number.
2715
2716 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2717
2718 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2719 PCRE_INFO_JCHANGED.
2720
2721 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2722 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2723 for detecting groups that can match an empty string.
2724
2725 9. A pattern with a very large number of alternatives (more than several
2726 hundred) was running out of internal workspace during the pre-compile
2727 phase, where pcre_compile() figures out how much memory will be needed. A
2728 bit of new cunning has reduced the workspace needed for groups with
2729 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2730 workspace instead of running out of the 4096 that are available.
2731
2732 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2733
2734 11. Applied patch from Google to remove an optimization that didn't quite work.
2735 The report of the bug said:
2736
2737 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2738 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2739 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2740
2741 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2742 it matched the wrong number of bytes.
2743
2744
2745 Version 7.1 24-Apr-07
2746 ---------------------
2747
2748 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2749 that is more "standard", making use of automake and other Autotools. There
2750 is some re-arrangement of the files and adjustment of comments consequent
2751 on this.
2752
2753 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2754 for recursive directory scanning broke on some systems because the files
2755 are not scanned in any specific order and on different systems the order
2756 was different. A call to "sort" has been inserted into RunGrepTest for the
2757 approprate test as a short-term fix. In the longer term there may be an
2758 alternative.
2759
2760 3. I had an email from Eric Raymond about problems translating some of PCRE's
2761 man pages to HTML (despite the fact that I distribute HTML pages, some
2762 people do their own conversions for various reasons). The problems
2763 concerned the use of low-level troff macros .br and .in. I have therefore
2764 removed all such uses from the man pages (some were redundant, some could
2765 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2766 HTML has been updated to handle .nf/.fi and to complain if it encounters
2767 .br or .in.
2768
2769 4. Updated comments in configure.ac that get placed in config.h.in and also
2770 arranged for config.h to be included in the distribution, with the name
2771 config.h.generic, for the benefit of those who have to compile without
2772 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2773
2774 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2775 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2776 makevp.bat for the current PCRE, using the additional files
2777 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2778
2779 6. A Windows user reported a minor discrepancy with test 2, which turned out
2780 to be caused by a trailing space on an input line that had got lost in his
2781 copy. The trailing space was an accident, so I've just removed it.
2782
2783 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2784 that is needed.
2785
2786 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2787 as "const" (a) because they are and (b) because it helps the PHP
2788 maintainers who have recently made a script to detect big data structures
2789 in the php code that should be moved to the .rodata section. I remembered
2790 to update Builducptable as well, so it won't revert if ucptable.h is ever
2791 re-created.
2792
2793 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2794 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2795 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2796 support is not required. This saves 1.5-2K of code, which is important in
2797 some applications.
2798
2799 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2800 so as not to refer to the tables, even though these functions will never be
2801 called when UTF-8 support is disabled. Otherwise there are problems with a
2802 shared library.
2803
2804 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2805
2806 (a) It was defining its arguments as char * instead of void *.
2807
2808 (b) It was assuming that all moves were upwards in memory; this was true
2809 a long time ago when I wrote it, but is no longer the case.
2810
2811 The emulated memove() is provided for those environments that have neither
2812 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2813 is clearly not the case, as these two bugs were recently reported.
2814
2815 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2816 and Detrail to create the HTML documentation, the .txt form of the man
2817 pages, and it removes trailing spaces from listed files. It also creates
2818 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2819 case, it wraps all the #defines with #ifndefs. This script should be run
2820 before "make dist".
2821
2822 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2823 with Unicode property support.
2824
2825 (a) For a maximizing quantifier, if the two different cases of the
2826 character were of different lengths in their UTF-8 codings (there are
2827 some cases like this - I found 11), and the matching function had to
2828 back up over a mixture of the two cases, it incorrectly assumed they
2829 were both the same length.
2830
2831 (b) When PCRE was configured to use the heap rather than the stack for
2832 recursion during matching, it was not correctly preserving the data for
2833 the other case of a UTF-8 character when checking ahead for a match
2834 while processing a minimizing repeat. If the check also involved
2835 matching a wide character, but failed, corruption could cause an
2836 erroneous result when trying to check for a repeat of the original
2837 character.
2838
2839 13. Some tidying changes to the testing mechanism:
2840
2841 (a) The RunTest script now detects the internal link size and whether there
2842 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2843 values substituted by "configure". (The RunGrepTest script already did
2844 this for UTF-8.) The configure.ac script no longer substitutes the
2845 relevant variables.
2846
2847 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2848 with length and offset values. This means that the output is different
2849 for different internal link sizes. Test 2 is skipped for link sizes
2850 other than 2 because of this, bypassing the problem. Unfortunately,
2851 there was also a test in test 3 (the locale tests) that used /B and
2852 failed for link sizes other than 2. Rather than cut the whole test out,
2853 I have added a new /Z option to pcretest that replaces the length and
2854 offset values with spaces. This is now used to make test 3 independent
2855 of link size. (Test 2 will be tidied up later.)
2856
2857 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2858 segmentation fault instead of returning the appropriate error message.
2859
2860 15. In multiline mode when the newline sequence was set to "any", the pattern
2861 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2862 This doesn't seem right; it now treats the CRLF combination as the line
2863 ending, and so does not match in that case. It's only a pattern such as ^$
2864 that would hit this one: something like ^ABC$ would have failed after \r
2865 and then tried again after \r\n.
2866
2867 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2868 in an attempt to make files that differ only in their line terminators
2869 compare equal. This works on Linux.
2870
2871 17. Under certain error circumstances pcregrep might try to free random memory
2872 as it exited. This is now fixed, thanks to valgrind.
2873
2874 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2875 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2876 was because its rules for how to advance for /g after matching an empty
2877 string at the end of a line did not allow for this case. They now check for
2878 it specially.
2879
2880 20. pcretest is supposed to handle patterns and data of any length, by
2881 extending its buffers when necessary. It was getting this wrong when the
2882 buffer for a data line had to be extended.
2883
2884 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2885 CRLF as a newline sequence.
2886
2887 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2888 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2889 I have nevertheless tidied it up.
2890
2891 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2892
2893 24. Added a man page for pcre-config.
2894
2895
2896 Version 7.0 19-Dec-06
2897 ---------------------
2898
2899 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2900 moving to gcc 4.1.1.
2901
2902 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2903 sys/time.h, which is documented as needed for this function. It doesn't
2904 seem to matter on Linux, but it showed up on some releases of OS X.
2905
2906 3. It seems that there are systems where bytes whose values are greater than
2907 127 match isprint() in the "C" locale. The "C" locale should be the
2908 default when a C program starts up. In most systems, only ASCII printing
2909 characters match isprint(). This difference caused the output from pcretest
2910 to vary, making some of the tests fail. I have changed pcretest so that:
2911
2912 (a) When it is outputting text in the compiled version of a pattern, bytes
2913 other than 32-126 are always shown as hex escapes.
2914
2915 (b) When it is outputting text that is a matched part of a subject string,
2916 it does the same, unless a different locale has been set for the match
2917 (using the /L modifier). In this case, it uses isprint() to decide.
2918
2919 4. Fixed a major bug that caused incorrect computation of the amount of memory
2920 required for a compiled pattern when options that changed within the
2921 pattern affected the logic of the preliminary scan that determines the
2922 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2923 that the computed length was too small. The symptoms of this bug were
2924 either the PCRE error "internal error: code overflow" from pcre_compile(),
2925 or a glibc crash with a message such as "pcretest: free(): invalid next
2926 size (fast)". Examples of patterns that provoked this bug (shown in
2927 pcretest format) are:
2928
2929 /(?-x: )/x
2930 /(?x)(?-x: \s*#\s*)/
2931 /((?i)[\x{c0}])/8
2932 /(?i:[\x{c0}])/8
2933
2934 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2935 is now done differently.
2936
2937 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2938 wrapper classes; (b) implement a new function in the C++ scanner that is
2939 more efficient than the old way of doing things because it avoids levels of
2940 recursion in the regex matching; (c) add a paragraph to the documentation
2941 for the FullMatch() function.
2942
2943 6. The escape sequence \n was being treated as whatever was defined as
2944 "newline". Not only was this contrary to the documentation, which states
2945 that \n is character 10 (hex 0A), but it also went horribly wrong when
2946 "newline" was defined as CRLF. This has been fixed.
2947
2948 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2949 was being set to -1 for the "end of line" case (supposedly a value that no
2950 character can have). Though this value is never used (the check for end of
2951 line is "zero bytes in current character"), it caused compiler complaints.
2952 I've changed it to 0xffffffff.
2953
2954 8. In pcre_version.c, the version string was being built by a sequence of
2955 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2956 string (as it is for production releases) called a macro with an empty
2957 argument. The C standard says the result of this is undefined. The gcc
2958 compiler treats it as an empty string (which was what was wanted) but it is
2959 reported that Visual C gives an error. The source has been hacked around to
2960 avoid this problem.
2961
2962 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2963 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2964 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2965 of them did).
2966
2967 10. Originally, pcretest opened its input and output without "b"; then I was
2968 told that "b" was needed in some environments, so it was added for release
2969 5.0 to both the input and output. (It makes no difference on Unix-like
2970 systems.) Later I was told that it is wrong for the input on Windows. I've
2971 now abstracted the modes into two macros, to make it easier to fiddle with
2972 them, and removed "b" from the input mode under Windows.
2973
2974 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2975
2976 12. Added -help and --help to pcretest as an official way of being reminded
2977 of the options.
2978
2979 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2980 and pcrecpp.cc because they annoy compilers at high warning levels.
2981
2982 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2983
2984 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2985 scripts are not C programs :-) and which was not noticed because it works
2986 on Linux.
2987
2988 16. pcretest is supposed to handle any length of pattern and data line (as one
2989 line or as a continued sequence of lines) by extending its input buffer if
2990 necessary. This feature was broken for very long pattern lines, leading to
2991 a string of junk being passed to pcre_compile() if the pattern was longer
2992 than about 50K.
2993
2994 17. I have done a major re-factoring of the way pcre_compile() computes the
2995 amount of memory needed for a compiled pattern. Previously, there was code
2996 that made a preliminary scan of the pattern in order to do this. That was
2997 OK when PCRE was new, but as the facilities have expanded, it has become
2998 harder and harder to keep it in step with the real compile phase, and there
2999 have been a number of bugs (see for example, 4 above). I have now found a
3000 cunning way of running the real compile function in a "fake" mode that
3001 enables it to compute how much memory it would need, while actually only
3002 ever using a few hundred bytes of working memory and without too many
3003 tests of the mode. This should make future maintenance and development
3004 easier. A side effect of this work is that the limit of 200 on the nesting
3005 depth of parentheses has been removed (though this was never a serious
3006 limitation, I suspect). However, there is a downside: pcre_compile() now
3007 runs more slowly than before (30% or more, depending on the pattern). I
3008 hope this isn't a big issue. There is no effect on runtime performance.
3009
3010 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3011 newline (only possible for the last line of a file) and it was a
3012 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3013
3014 19. Added additional timing features to pcretest. (1) The -tm option now times
3015 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3016 separate command line item, by a number that specifies the number of
3017 repeats to use when timing. The default is 50000; this gives better
3018 precision, but takes uncomfortably long for very large patterns.
3019
3020 20. Extended pcre_study() to be more clever in cases where a branch of a
3021 subpattern has no definite first character. For example, (a*|b*)[cd] would
3022 previously give no result from pcre_study(). Now it recognizes that the
3023 first character must be a, b, c, or d.
3024
3025 21. There was an incorrect error "recursive call could loop indefinitely" if
3026 a subpattern (or the entire pattern) that was being tested for matching an
3027 empty string contained only one non-empty item after a nested subpattern.
3028 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3029 incorrectly, because the \d was being skipped in the check.
3030
3031 22. The pcretest program now has a new pattern option /B and a command line
3032 option -b, which is equivalent to adding /B to every pattern. This causes
3033 it to show the compiled bytecode, without the additional information that
3034 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3035 is the same as /B/I).
3036
3037 23. A new optimization is now able automatically to treat some sequences such
3038 as a*b as a*+b. More specifically, if something simple (such as a character
3039 or a simple class like \d) has an unlimited quantifier, and is followed by
3040 something that cannot possibly match the quantified thing, the quantifier
3041 is automatically "possessified".
3042
3043 24. A recursive reference to a subpattern whose number was greater than 39
3044 went wrong under certain circumstances in UTF-8 mode. This bug could also
3045 have affected the operation of pcre_study().
3046
3047 25. Realized that a little bit of performance could be had by replacing
3048 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3049
3050 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3051
3052 27. Possessive quantifiers such as a++ were previously implemented by turning
3053 them into atomic groups such as ($>a+). Now they have their own opcodes,
3054 which improves performance. This includes the automatically created ones
3055 from 23 above.
3056
3057 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3058 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3059 the first matched character to be a colon. This applied both to named and
3060 numbered groups.
3061
3062 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3063
3064 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3065 building PCRE on Windows easier, so I have included it in the distribution.
3066
3067 31. There is now a check in pcretest against a ridiculously large number being
3068 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3069 loop, the loop is abandoned.
3070
3071 32. Forward references to subpatterns in conditions such as (?(2)...) where
3072 subpattern 2 is defined later cause pcre_compile() to search forwards in
3073 the pattern for the relevant set of parentheses. This search went wrong
3074 when there were unescaped parentheses in a character class, parentheses
3075 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3076
3077 33. "Subroutine" calls and backreferences were previously restricted to
3078 referencing subpatterns earlier in the regex. This restriction has now
3079 been removed.
3080
3081 34. Added a number of extra features that are going to be in Perl 5.10. On the
3082 whole, these are just syntactic alternatives for features that PCRE had
3083 previously implemented using the Python syntax or my own invention. The
3084 other formats are all retained for compatibility.
3085
3086 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3087 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3088 also .NET compatible.
3089
3090 (b) A recursion or subroutine call to a named group can now be defined as
3091 (?&name) as well as (?P>name).
3092
3093 (c) A backreference to a named group can now be defined as \k<name> or
3094 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3095 5.10, are also .NET compatible.
3096
3097 (d) A conditional reference to a named group can now use the syntax
3098 (?(<name>) or (?('name') as well as (?(name).
3099
3100 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3101 groups (named and numbered) that are never evaluated inline, but can be
3102 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3103 is always false. There may be only one alternative in such a group.
3104
3105 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3106 as the simple (?(R). The condition is true only if the most recent
3107 recursion is that of the given number or name. It does not search out
3108 through the entire recursion stack.
3109
3110 (g) The escape \gN or \g{N} has been added, where N is a positive or
3111 negative number, specifying an absolute or relative reference.
3112
3113 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3114 some "unreachable code" warnings.
3115
3116 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3117 things, this adds five new scripts.
3118
3119 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3120 There were also incompatibilities regarding the handling of \Q..\E inside
3121 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3122 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3123
3124 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3125 matches an empty string, and forcibly breaks the loop. There were bugs in
3126 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3127 against aaaa the result was just "a" rather than "aaaa", for example. Two
3128 separate and independent bugs (that affected different cases) have been
3129 fixed.
3130
3131 39. Refactored the code to abolish the use of different opcodes for small
3132 capturing bracket numbers. This is a tidy that I avoided doing when I
3133 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3134 The new approach is not only tidier, it makes it possible to reduce the
3135 memory needed to fix the previous bug (38).
3136
3137 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3138 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3139 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3140 mode.
3141
3142 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3143 report.
3144
3145 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3146 copy construction and assignment in the C++ wrapper.
3147
3148 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3149 couple of bugs that could have given wrong results in the "--newline=crlf"
3150 case.
3151
3152 44. Added a number of casts and did some reorganization of signed/unsigned int
3153 variables following suggestions from Dair Grant. Also renamed the variable
3154 "this" as "item" because it is a C++ keyword.
3155
3156 45. Arranged for dftables to add
3157
3158 #include "pcre_internal.h"
3159
3160 to pcre_chartables.c because without it, gcc 4.x may remove the array
3161 definition from the final binary if PCRE is built into a static library and
3162 dead code stripping is activated.
3163
3164 46. For an unanchored pattern, if a match attempt fails at the start of a
3165 newline sequence, and the newline setting is CRLF or ANY, and the next two
3166 characters are CRLF, advance by two characters instead of one.
3167
3168
3169 Version 6.7 04-Jul-06
3170 ---------------------
3171
3172 1. In order to handle tests when input lines are enormously long, pcretest has
3173 been re-factored so that it automatically extends its buffers when
3174 necessary. The code is crude, but this _is_ just a test program. The
3175 default size has been increased from 32K to 50K.
3176
3177 2. The code in pcre_study() was using the value of the re argument before
3178 testing it for NULL. (Of course, in any sensible call of the function, it
3179 won't be NULL.)
3180
3181 3. The memmove() emulation function in pcre_internal.h, which is used on
3182 systems that lack both memmove() and bcopy() - that is, hardly ever -
3183 was missing a "static" storage class specifier.
3184
3185 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3186 containing an extended class (one that cannot be represented by a bitmap
3187 because it contains high-valued characters or Unicode property items, e.g.
3188 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3189 pattern, but PCRE should not loop if you do not (it no longer does).
3190 [Detail: two cases were found: (a) a repeated subpattern containing an
3191 extended class; (b) a recursive reference to a subpattern that followed a
3192 previous extended class. It wasn't skipping over the extended class
3193 correctly when UTF-8 mode was not set.]
3194
3195 5. A negated single-character class was not being recognized as fixed-length
3196 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3197 compile error "lookbehind assertion is not fixed length".
3198
3199 6. The RunPerlTest auxiliary script was showing an unexpected difference
3200 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3201 write a Perl script that can interpret lines of an input file either as
3202 byte characters or as UTF-8, which is what "perltest" was being required to
3203 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3204 can't do is switch easily at run time between having the "use utf8;" pragma
3205 or not. In the end, I fudged it by using the RunPerlTest script to insert
3206 "use utf8;" explicitly for the UTF-8 tests.
3207
3208 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3209 the end of the subject string, contrary to the documentation and to what
3210 Perl does. This was true of both matching functions. Now it matches only at
3211 the start of the subject and immediately after *internal* newlines.
3212
3213 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3214 a pointer to an int instead of a pointer to an unsigned long int. This
3215 caused problems on 64-bit systems.
3216
3217 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3218 instance of the 'standard' template library not being so standard".
3219
3220 10. There was no check on the number of named subpatterns nor the maximum
3221 length of a subpattern name. The product of these values is used to compute
3222 the size of the memory block for a compiled pattern. By supplying a very
3223 long subpattern name and a large number of named subpatterns, the size
3224 computation could be caused to overflow. This is now prevented by limiting
3225 the length of names to 32 characters, and the number of named subpatterns
3226 to 10,000.
3227
3228 11. Subpatterns that are repeated with specific counts have to be replicated in
3229 the compiled pattern. The size of memory for this was computed from the
3230 length of the subpattern and the repeat count. The latter is limited to
3231 65535, but there was no limit on the former, meaning that integer overflow
3232 could in principle occur. The compiled length of a repeated subpattern is
3233 now limited to 30,000 bytes in order to prevent this.
3234
3235 12. Added the optional facility to have named substrings with the same name.
3236
3237 13. Added the ability to use a named substring as a condition, using the
3238 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3239 are numbers (not recommended). Forward references are permitted.
3240
3241 14. Added forward references in named backreferences (if you see what I mean).
3242
3243 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3244 pattern could run off the end of the subject. For example, the pattern
3245 "(?s)(.{1,5})"8 did this with the subject "ab".
3246
3247 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3248 PCRE_CASELESS was set when matching characters that were quantified with ?
3249 or *.
3250
3251 17. A character class other than a single negated character that had a minimum
3252 but no maximum quantifier - for example [ab]{6,} - was not handled
3253 correctly by pce_dfa_exec(). It would match only one character.
3254
3255 18. A valid (though odd) pattern that looked like a POSIX character
3256 class but used an invalid character after [ (for example [[,abc,]]) caused
3257 pcre_compile() to give the error "Failed: internal error: code overflow" or
3258 in some cases to crash with a glibc free() error. This could even happen if
3259 the pattern terminated after [[ but there just happened to be a sequence of
3260 letters, a binary zero, and a closing ] in the memory that followed.
3261
3262 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3263 over the years. Originally (before any Unicode support), just the bottom 8
3264 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3265 output from "man perlunicode" includes this:
3266
3267 The regular expression compiler produces polymorphic opcodes. That
3268 is, the pattern adapts to the data and automatically switches to
3269 the Unicode character scheme when presented with Unicode data--or
3270 instead uses a traditional byte scheme when presented with byte
3271 data.
3272
3273 Sadly, a wide octal escape does not cause a switch, and in a string with
3274 no other multibyte characters, these octal escapes are treated as before.
3275 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3276 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3277 Unicode string.
3278
3279 I have not perpetrated such confusion in PCRE. Up till now, it took just
3280 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3281 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3282 translate to the appropriate multibyte character.
3283
3284 29. Applied some refactoring to reduce the number of warnings from Microsoft
3285 and Borland compilers. This has included removing the fudge introduced
3286 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3287 a warning about an unused variable.
3288
3289 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3290 characters since release 4.0, because Perl (from release 5.004) does not.
3291 [Or at least, is documented not to: some releases seem to be in conflict
3292 with the documentation.] However, when a pattern was studied with
3293 pcre_study() and all its branches started with \s, PCRE still included VT
3294 as a possible starting character. Of course, this did no harm; it just
3295 caused an unnecessary match attempt.
3296
3297 22. Removed a now-redundant internal flag bit that recorded the fact that case
3298 dependency changed within the pattern. This was once needed for "required
3299 byte" processing, but is no longer used. This recovers a now-scarce options
3300 bit. Also moved the least significant internal flag bit to the most-
3301 significant bit of the word, which was not previously used (hangover from
3302 the days when it was an int rather than a uint) to free up another bit for
3303 the future.
3304
3305 23. Added support for CRLF line endings as well as CR and LF. As well as the
3306 default being selectable at build time, it can now be changed at runtime
3307 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3308 specify that it is scanning data with non-default line endings.
3309
3310 24. Changed the definition of CXXLINK to make it agree with the definition of
3311 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3312
3313 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3314 recursions. This makes a big different to stack usage for some patterns.
3315
3316 26. If a subpattern containing a named recursion or subroutine reference such
3317 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3318 the space required for the compiled pattern went wrong and gave too small a
3319 value. Depending on the environment, this could lead to "Failed: internal
3320 error: code overflow at offset 49" or "glibc detected double free or
3321 corruption" errors.
3322
3323 27. Applied patches from Google (a) to support the new newline modes and (b) to
3324 advance over multibyte UTF-8 characters in GlobalReplace.
3325
3326 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3327 difference for some implementation of PCRE in some Windows version.
3328
3329 29. Added some extra testing facilities to pcretest:
3330
3331 \q<number> in a data line sets the "match limit" value
3332 \Q<number> in a data line sets the "match recursion limt" value
3333 -S <number> sets the stack size, where <number> is in megabytes
3334
3335 The -S option isn't available for Windows.
3336
3337
3338 Version 6.6 06-Feb-06
3339 ---------------------
3340
3341 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3342 in pcreposix.h. I have copied the definition from pcre.h.
3343
3344 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3345 because pcre.h is no longer a built file.
3346
3347 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3348 not normally included in the compiled code.
3349
3350
3351 Version 6.5 01-Feb-06
3352 ---------------------
3353
3354 1. When using the partial match feature with pcre_dfa_exec(), it was not
3355 anchoring the second and subsequent partial matches at the new starting
3356 point. This could lead to incorrect results. For example, with the pattern
3357 /1234/, partially matching against "123" and then "a4" gave a match.
3358
3359 2. Changes to pcregrep:
3360
3361 (a) All non-match returns from pcre_exec() were being treated as failures
3362 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3363 error message is output. Some extra information is given for the
3364 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3365 probably the only errors that are likely to be caused by users (by
3366 specifying a regex that has nested indefinite repeats, for instance).
3367 If there are more than 20 of these errors, pcregrep is abandoned.
3368
3369 (b) A binary zero was treated as data while matching, but terminated the
3370 output line if it was written out. This has been fixed: binary zeroes
3371 are now no different to any other data bytes.
3372
3373 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3374 used to set a locale for matching. The --locale=xxxx long option has
3375 been added (no short equivalent) to specify a locale explicitly on the
3376 pcregrep command, overriding the environment variables.
3377
3378 (d) When -B was used with -n, some line numbers in the output were one less
3379 than they should have been.
3380
3381 (e) Added the -o (--only-matching) option.
3382
3383 (f) If -A or -C was used with -c (count only), some lines of context were
3384 accidentally printed for the final match.
3385
3386 (g) Added the -H (--with-filename) option.
3387
3388 (h) The combination of options -rh failed to suppress file names for files
3389 that were found from directory arguments.
3390
3391 (i) Added the -D (--devices) and -d (--directories) options.
3392
3393 (j) Added the -F (--fixed-strings) option.
3394
3395 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3396
3397 (l) Added the --colo(u)r option.
3398
3399 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3400 is not present by default.
3401
3402 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3403 items such as (?R) or (?1), when the recursion could match a number of
3404 alternatives. If it matched one of the alternatives, but subsequently,
3405 outside the recursion, there was a failure, the code tried to back up into
3406 the recursion. However, because of the way PCRE is implemented, this is not
3407 possible, and the result was an incorrect result from the match.
3408
3409 In order to prevent this happening, the specification of recursion has
3410 been changed so that all such subpatterns are automatically treated as
3411 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3412
3413 4. I had overlooked the fact that, in some locales, there are characters for
3414 which isalpha() is true but neither isupper() nor islower() are true. In
3415 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3416 and ordfeminine) are like this. This affected the treatment of \w and \W
3417 when they appeared in character classes, but not when they appeared outside
3418 a character class. The bit map for "word" characters is now created
3419 separately from the results of isalnum() instead of just taking it from the
3420 upper, lower, and digit maps. (Plus the underscore character, of course.)
3421
3422 5. The above bug also affected the handling of POSIX character classes such as
3423 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3424 permanent tables. Instead, the bit maps for such a class were previously
3425 created as the appropriate unions of the upper, lower, and digit bitmaps.
3426 Now they are created by subtraction from the [[:word:]] class, which has
3427 its own bitmap.
3428
3429 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3430 It is created by subtracting the vertical space characters (\x09, \x0a,
3431 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3432 subtraction was done in the overall bitmap for a character class, meaning
3433 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3434 be recognized. This bug has been fixed.
3435
3436 7. Patches from the folks at Google:
3437
3438 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3439 real life, but is still worth protecting against".
3440
3441 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3442 regular expressions".
3443
3444 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3445 have it.
3446
3447 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3448 "configure" and the latter not, in order to fix a problem somebody had
3449 with compiling the Arg class on HP-UX.
3450
3451 (e) Improve the error-handling of the C++ wrapper a little bit.
3452
3453 (f) New tests for checking recursion limiting.
3454
3455 8. The pcre_memmove() function, which is used only if the environment does not
3456 have a standard memmove() function (and is therefore rarely compiled),
3457 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3458 returning a result (though PCRE never actually uses the result).
3459
3460 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3461 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3462 returned instead of calling malloc() with an overflowing number that would
3463 most likely cause subsequent chaos.
3464
3465 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3466
3467 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3468 with this option is matched, the nmatch and pmatch options of regexec() are
3469 ignored.
3470
3471 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3472 provided in case anyone wants to the the POSIX interface with UTF-8
3473 strings.
3474
3475 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3476 C++ linking (needed for some HP-UX environments).
3477
3478 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3479 (unused parameter) and in the pcre_printint() function (omitted "default"
3480 switch label when the default is to do nothing).
3481
3482 15. Added some code to make it possible, when PCRE is compiled as a C++
3483 library, to replace subject pointers for pcre_exec() with a smart pointer
3484 class, thus making it possible to process discontinuous strings.
3485
3486 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3487 much the same function. They were added by different people who were trying
3488 to make PCRE easy to compile on non-Unix systems. It has been suggested
3489 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3490 for compiling on Windows systems. I have therefore replaced it with
3491 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3492 defaults to "extern" for C or "extern C" for C++, which works fine on
3493 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3494 SCOPE with something explicit in config.h. In addition:
3495
3496 (a) pcreposix.h still had just "extern" instead of either of these macros;
3497 I have replaced it with PCRE_DATA_SCOPE.
3498
3499 (b) Functions such as _pcre_xclass(), which are internal to the library,
3500 but external in the C sense, all had PCRE_EXPORT in their definitions.
3501 This is apparently wrong for the Windows case, so I have removed it.
3502 (It makes no difference on Unix-like systems.)
3503
3504 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3505 of recursive calls to match(). This is different to MATCH_LIMIT because
3506 that limits the total number of calls to match(), not all of which increase
3507 the depth of recursion. Limiting the recursion depth limits the amount of
3508 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3509 when PCRE is compiled, and changed at run time. A patch from Google adds
3510 this functionality to the C++ interface.
3511
3512 18. Changes to the handling of Unicode character properties:
3513
3514 (a) Updated the table to Unicode 4.1.0.
3515
3516 (b) Recognize characters that are not in the table as "Cn" (undefined).
3517
3518 (c) I revised the way the table is implemented to a much improved format
3519 which includes recognition of ranges. It now supports the ranges that
3520 are defined in UnicodeData.txt, and it also amalgamates other
3521 characters into ranges. This has reduced the number of entries in the
3522 table from around 16,000 to around 3,000, thus reducing its size
3523 considerably. I realized I did not need to use a tree structure after
3524 all - a binary chop search is just as efficient. Having reduced the
3525 number of entries, I extended their size from 6 bytes to 8 bytes to
3526 allow for more data.
3527
3528 (d) Added support for Unicode script names via properties such as \p{Han}.
3529
3530 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3531 matching that character.
3532
3533 20. When matching a repeated Unicode property with a minimum greater than zero,
3534 (for example \pL{2,}), PCRE could look past the end of the subject if it
3535 reached it while seeking the minimum number of characters. This could
3536 happen only if some of the characters were more than one byte long, because
3537 there is a check for at least the minimum number of bytes.
3538
3539 21. Refactored the implementation of \p and \P so as to be more general, to
3540 allow for more different types of property in future. This has changed the
3541 compiled form incompatibly. Anybody with saved compiled patterns that use
3542 \p or \P will have to recompile them.
3543
3544 22. Added "Any" and "L&" to the supported property types.
3545
3546 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3547 but give a compile time error if the value is greater than 0xff.
3548
3549 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3550 accidentally not being installed or uninstalled.
3551
3552 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3553 made were to insert the current release number. This seemed silly, because
3554 it made things harder for people building PCRE on systems that don't run
3555 "configure". I have turned pcre.h into a distributed file, no longer built
3556 by "configure", with the version identification directly included. There is
3557 no longer a pcre.h.in file.
3558
3559 However, this change necessitated a change to the pcre-config script as
3560 well. It is built from pcre-config.in, and one of the substitutions was the
3561 release number. I have updated configure.ac so that ./configure now finds
3562 the release number by grepping pcre.h.
3563
3564 26. Added the ability to run the tests under valgrind.
3565
3566
3567 Version 6.4 05-Sep-05
3568 ---------------------
3569
3570 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3571 "--" to be printed when multiple files were scanned, even when none of the
3572 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3573 consider it to be a bug, and have restored the previous behaviour.
3574
3575 2. A couple of code tidies to get rid of compiler warnings.
3576
3577 3. The pcretest program used to cheat by referring to symbols in the library
3578 whose names begin with _pcre_. These are internal symbols that are not
3579 really supposed to be visible externally, and in some environments it is
3580 possible to suppress them. The cheating is now confined to including
3581 certain files from the library's source, which is a bit cleaner.
3582
3583 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3584 file's purpose clearer.
3585
3586 5. Reorganized pcre_ucp_findchar().
3587
3588
3589 Version 6.3 15-Aug-05
3590 ---------------------
3591
3592 1. The file libpcre.pc.in did not have general read permission in the tarball.
3593
3594 2. There were some problems when building without C++ support:
3595
3596 (a) If C++ support was not built, "make install" and "make test" still
3597 tried to test it.
3598
3599 (b) There were problems when the value of CXX was explicitly set. Some
3600 changes have been made to try to fix these, and ...
3601
3602 (c) --disable-cpp can now be used to explicitly disable C++ support.
3603
3604 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3605 backslash in a target when C++ was disabled. This confuses some
3606 versions of "make", apparently. Using an intermediate variable solves
3607 this. (Same for CPP_LOBJ.)
3608
3609 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3610 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3611 necessary on certain architectures.
3612
3613 4. Added a setting of -export-symbols-regex to the link command to remove
3614 those symbols that are exported in the C sense, but actually are local
3615 within the library, and not documented. Their names all begin with
3616 "_pcre_". This is not a perfect job, because (a) we have to except some
3617 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3618 available (and never for static libraries). I have made a note to try to
3619 find a way round (a) in the future.
3620
3621
3622 Version 6.2 01-Aug-05
3623 ---------------------
3624
3625 1. There was no test for integer overflow of quantifier values. A construction
3626 such as {1111111111111111} would give undefined results. What is worse, if
3627 a minimum quantifier for a parenthesized subpattern overflowed and became
3628 negative, the calculation of the memory size went wrong. This could have
3629 led to memory overwriting.
3630
3631 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3632
3633 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3634 operating environments where this matters.
3635
3636 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3637 PCRE options from within the C++ wrapper.
3638
3639 5. Named capturing subpatterns were not being correctly counted when a pattern
3640 was compiled. This caused two problems: (a) If there were more than 100
3641 such subpatterns, the calculation of the memory needed for the whole
3642 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3643 back references of the form \12, where the number was greater than 9, were
3644 not recognized as back references, even though there were sufficient
3645 previous subpatterns.
3646
3647 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3648 versions of gcc, e.g. 2.95.4.
3649
3650
3651 Version 6.1 21-Jun-05
3652 ---------------------
3653
3654 1. There was one reference to the variable "posix" in pcretest.c that was not
3655 surrounded by "#if !defined NOPOSIX".
3656
3657 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3658 the cross-check on the old pcre_info() function, for the benefit of the
3659 cut-down version of PCRE that is currently imported into Exim.
3660
3661 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3662 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3663 patterns that start (?i) but I don't think that matters. The use of (?i) is
3664 just an example; this all applies to the other options as well.
3665
3666 4. Since libtool seems to echo the compile commands it is issuing, the output
3667 from "make" can be reduced a bit by putting "@" in front of each libtool
3668 compile command.
3669
3670 5. Patch from the folks at Google for configure.in to be a bit more thorough
3671 in checking for a suitable C++ installation before trying to compile the
3672 C++ stuff. This should fix a reported problem when a compiler was present,
3673 but no suitable headers.
3674
3675 6. The man pages all had just "PCRE" as their title. I have changed them to
3676 be the relevant file name. I have also arranged that these names are
3677 retained in the file doc/pcre.txt, which is a concatenation in text format
3678 of all the man pages except the little individual ones for each function.
3679
3680 7. The NON-UNIX-USE file had not been updated for the different set of source
3681 files that come with release 6. I also added a few comments about the C++
3682 wrapper.
3683
3684
3685 Version 6.0 07-Jun-05
3686 ---------------------
3687
3688 1. Some minor internal re-organization to help with my DFA experiments.
3689
3690 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3691 didn't matter for the library itself when fully configured, but did matter
3692 when compiling without UCP support, or within Exim, where the ucp files are
3693 not imported.
3694
3695 3. Refactoring of the library code to split up the various functions into
3696 different source modules. The addition of the new DFA matching code (see
3697 below) to a single monolithic source would have made it really too
3698 unwieldy, quite apart from causing all the code to be include in a
3699 statically linked application, when only some functions are used. This is
3700 relevant even without the DFA addition now that patterns can be compiled in
3701 one application and matched in another.
3702
3703 The downside of splitting up is that there have to be some external
3704 functions and data tables that are used internally in different modules of
3705 the library but which are not part of the API. These have all had their
3706 names changed to start with "_pcre_" so that they are unlikely to clash
3707 with other external names.
3708
3709 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3710 a different (DFA) algorithm. Although it is slower than the original
3711 function, it does have some advantages for certain types of matching
3712 problem.
3713
3714 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3715 including restarting after a partial match.
3716
3717 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3718 defined when compiling for Windows was sent to me. I have put it into the
3719 code, though I have no means of testing or verifying it.
3720
3721 7. Added the pcre_refcount() auxiliary function.
3722
3723 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3724 match before or at the first newline in the subject string. In pcretest,
3725 the /f option on a pattern can be used to set this.
3726
3727 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3728 would behave wrongly. This has been present in PCRE since release 4.0.
3729
3730 10. A number of changes to the pcregrep command:
3731
3732 (a) Refactored how -x works; insert ^(...)$ instead of setting
3733 PCRE_ANCHORED and checking the length, in preparation for adding
3734 something similar for -w.
3735
3736 (b) Added the -w (match as a word) option.
3737
3738 (c) Refactored the way lines are read and buffered so as to have more
3739 than one at a time available.
3740
3741 (d) Implemented a pcregrep test script.
3742
3743 (e) Added the -M (multiline match) option. This allows patterns to match
3744 over several lines of the subject. The buffering ensures that at least
3745 8K, or the rest of the document (whichever is the shorter) is available
3746 for matching (and similarly the previous 8K for lookbehind assertions).
3747
3748 (f) Changed the --help output so that it now says
3749
3750 -w, --word-regex(p)
3751
3752 instead of two lines, one with "regex" and the other with "regexp"
3753 because that confused at least one person since the short forms are the
3754 same. (This required a bit of code, as the output is generated
3755 automatically from a table. It wasn't just a text change.)
3756
3757 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3758 option but starts with a hyphen. Could be a pattern or a path name
3759 starting with a hyphen, for instance.
3760
3761 (h) "-" can be given as a file name to represent stdin.
3762
3763 (i) When file names are being printed, "(standard input)" is used for
3764 the standard input, for compatibility with GNU grep. Previously
3765 "<stdin>" was used.
3766
3767 (j) The option --label=xxx can be used to supply a name to be used for
3768 stdin when file names are being printed. There is no short form.
3769
3770 (k) Re-factored the options decoding logic because we are going to add
3771 two more options that take data. Such options can now be given in four
3772 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3773
3774 (l) Added the -A, -B, and -C options for requesting that lines of context
3775 around matches be printed.
3776
3777 (m) Added the -L option to print the names of files that do not contain
3778 any matching lines, that is, the complement of -l.
3779
3780 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3781 continue to scan other files.
3782
3783 (o) The -s option was incorrectly implemented. For compatibility with other
3784 greps, it now suppresses the error message for a non-existent or non-
3785 accessible file (but not the return code). There is a new option called
3786 -q that suppresses the output of matching lines, which was what -s was
3787 previously doing.
3788
3789 (p) Added --include and --exclude options to specify files for inclusion
3790 and exclusion when recursing.
3791
3792 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3793 Hopefully, it now does.
3794
3795 12. Missing cast in pcre_study().
3796
3797 13. Added an "uninstall" target to the makefile.
3798
3799 14. Replaced "extern" in the function prototypes in Makefile.in with
3800 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3801 world, but is set differently for Windows.
3802
3803 15. Added a second compiling function called pcre_compile2(). The only
3804 difference is that it has an extra argument, which is a pointer to an
3805 integer error code. When there is a compile-time failure, this is set
3806 non-zero, in addition to the error test pointer being set to point to an
3807 error message. The new argument may be NULL if no error number is required
3808 (but then you may as well call pcre_compile(), which is now just a
3809 wrapper). This facility is provided because some applications need a
3810 numeric error indication, but it has also enabled me to tidy up the way
3811 compile-time errors are handled in the POSIX wrapper.
3812
3813 16. Added VPATH=.libs to the makefile; this should help when building with one
3814 prefix path and installing with another. (Or so I'm told by someone who
3815 knows more about this stuff than I do.)
3816
3817 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3818 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3819 match everything, including newlines. This is not POSIX-compatible, but
3820 somebody wanted the feature. From pcretest it can be activated by using
3821 both the P and the s flags.
3822
3823 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3824
3825 19. libpcre.pc was being incorrectly installed as executable.
3826
3827 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3828 it now also looks for '\r' so that it will work unmodified on Windows.
3829
3830 21. Added Google's contributed C++ wrapper to the distribution.
3831
3832 22. Added some untidy missing memory free() calls in pcretest, to keep
3833 Electric Fence happy when testing.
3834
3835
3836
3837 Version 5.0 13-Sep-04
3838 ---------------------
3839
3840 1. Internal change: literal characters are no longer packed up into items
3841 containing multiple characters in a single byte-string. Each character
3842 is now matched using a separate opcode. However, there may be more than one
3843 byte in the character in UTF-8 mode.
3844
3845 2. The pcre_callout_block structure has two new fields: pattern_position and
3846 next_item_length. These contain the offset in the pattern to the next match
3847 item, and its length, respectively.
3848
3849 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3850 insertion of callouts before each pattern item. Added the /C option to
3851 pcretest to make use of this.
3852
3853 4. On the advice of a Windows user, the lines
3854
3855 #if defined(_WIN32) || defined(WIN32)
3856 _setmode( _fileno( stdout ), 0x8000 );
3857 #endif /* defined(_WIN32) || defined(WIN32) */
3858
3859 have been added to the source of pcretest. This apparently does useful
3860 magic in relation to line terminators.
3861
3862 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3863 for the benefit of those environments where the "b" makes a difference.
3864
3865 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3866 to know about it. I have put a hack into configure.in that adds in code
3867 to set GCC=yes if CC=icc. This seems to end up at a point in the
3868 generated configure script that is early enough to affect the setting of
3869 compiler options, which is what is needed, but I have no means of testing
3870 whether it really works. (The user who reported this had patched the
3871 generated configure script, which of course I cannot do.)
3872
3873 LATER: After change 22 below (new libtool files), the configure script
3874 seems to know about icc (and also ecc). Therefore, I have commented out
3875 this hack in configure.in.
3876
3877 7. Added support for pkg-config (2 patches were sent in).
3878
3879 8. Negated POSIX character classes that used a combination of internal tables
3880 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3881 [[:^ascii]]. Typically, they would match almost any characters. The other
3882 POSIX classes were not broken in this way.
3883
3884 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3885 to find the match, as PCRE was deluded into thinking that the match had to
3886 start at the start point or following a newline. The same bug applied to
3887 patterns with negative forward assertions or any backward assertions
3888 preceding ".*" at the start, unless the pattern required a fixed first
3889 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3890
3891 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3892 starting at the last subject character, bytes beyond the end of the subject
3893 string were read.
3894
3895 11. Renamed the variable "class" as "classbits" to make life easier for C++
3896 users. (Previously there was a macro definition, but it apparently wasn't
3897 enough.)
3898
3899 12. Added the new field "tables" to the extra data so that tables can be passed
3900 in at exec time, or the internal tables can be re-selected. This allows
3901 a compiled regex to be saved and re-used at a later time by a different
3902 program that might have everything at different addresses.
3903
3904 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3905 -R library as well as a -L library.
3906
3907 14. The debugging options of pcretest (-d on the command line or D on a
3908 pattern) showed incorrect output for anything following an extended class
3909 that contained multibyte characters and which was followed by a quantifier.
3910
3911 15. Added optional support for general category Unicode character properties
3912 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3913 support. It adds about 90K to the size of the library. The meanings of the
3914 inbuilt class escapes such as \d and \s have NOT been changed.
3915
3916 16. Updated pcredemo.c to include calls to free() to release the memory for the
3917 compiled pattern.
3918
3919 17. The generated file chartables.c was being created in the source directory
3920 instead of in the building directory. This caused the build to fail if the
3921 source directory was different from the building directory, and was
3922 read-only.
3923
3924 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3925 file. No doubt somebody will tell me if they don't make sense... Also added
3926 Dan Mooney's comments about building on OpenVMS.
3927
3928 19. Added support for partial matching via the PCRE_PARTIAL option for
3929 pcre_exec() and the \P data escape in pcretest.
3930
3931 20. Extended pcretest with 3 new pattern features:
3932
3933 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3934 write the compiled pattern to the file whose name is "rest-of-line".
3935 This is a straight binary dump of the data, with the saved pointer to
3936 the character tables forced to be NULL. The study data, if any, is
3937 written too. After writing, pcretest reads a new pattern.
3938
3939 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3940 compiled pattern from the given file. There must not be any
3941 occurrences of "<" in the file name (pretty unlikely); if there are,
3942 pcretest will instead treat the initial "<" as a pattern delimiter.
3943 After reading in the pattern, pcretest goes on to read data lines as
3944 usual.
3945
3946 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3947 and 16-bit fields in a compiled pattern, to simulate a pattern that
3948 was compiled on a host of opposite endianness.
3949
3950 21. The pcre-exec() function can now cope with patterns that were compiled on
3951 hosts of opposite endianness, with this restriction:
3952
3953 As for any compiled expression that is saved and used later, the tables
3954 pointer field cannot be preserved; the extra_data field in the arguments
3955 to pcre_exec() should be used to pass in a tables address if a value
3956 other than the default internal tables were used at compile time.
3957
3958 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3959 now diagnosed as an error. Previously, most of the time, a negative number
3960 would have been treated as zero, but if in addition "ovector" was passed as
3961 NULL, a crash could occur.
3962
3963 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3964 new versions from the libtool 1.5 distribution (the last one is a copy of
3965 a file called libtool.m4). This seems to have fixed the need to patch
3966 "configure" to support Darwin 1.3 (which I used to do). However, I still
3967 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3968 workstation).
3969
3970 24. Changed the PCRE licence to be the more standard "BSD" licence.
3971
3972
3973 Version 4.5 01-Dec-03
3974 ---------------------
3975
3976 1. There has been some re-arrangement of the code for the match() function so
3977 that it can be compiled in a version that does not call itself recursively.
3978 Instead, it keeps those local variables that need separate instances for
3979 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3980 needs to "recurse". Keeping track of where control must go is done by means
3981 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3982 hide most of the details from the main code, and operates only if
3983 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3984 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3985 operating.
3986
3987 To make it easier for callers to provide specially tailored get/free
3988 functions for this usage, two new functions, pcre_stack_malloc, and
3989 pcre_stack_free, are used. They are always called in strict stacking order,
3990 and the size of block requested is always the same.
3991
3992 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3993 PCRE has been compiled to use the stack or the heap for recursion. The
3994 -C option of pcretest uses this to show which version is compiled.
3995
3996 A new data escape \S, is added to pcretest; it causes the amounts of store
3997 obtained and freed by both kinds of malloc/free at match time to be added
3998 to the output.
3999
4000 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4001 what's available on my current Linux desktop machine.
4002
4003 3. When matching a UTF-8 string, the test for a valid string at the start has
4004 been extended. If start_offset is not zero, PCRE now checks that it points
4005 to a byte that is the start of a UTF-8 character. If not, it returns
4006 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4007 this is necessary because there may be backward assertions in the pattern.
4008 When matching the same subject several times, it may save resources to use
4009 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4010
4011 4. The code for checking the validity of UTF-8 strings has been tightened so
4012 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4013 containing "overlong sequences".
4014
4015 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4016 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4017 should have been "&", but it just so happened that all the cases this let
4018 through by mistake were picked up later in the function.
4019
4020 6. I had used a variable called "isblank" - this is a C99 function, causing
4021 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4022
4023 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4024 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4025 so the version is shown; (c) stop "make test" echoing "./RunTest".
4026
4027 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4028
4029 9. The prototype for memmove() for systems that don't have it was using
4030 size_t, but the inclusion of the header that defines size_t was later. I've
4031 moved the #includes for the C headers earlier to avoid this.
4032
4033 10. Added some adjustments to the code to make it easier to compiler on certain
4034 special systems:
4035
4036 (a) Some "const" qualifiers were missing.
4037 (b) Added the macro EXPORT before all exported functions; by default this
4038 is defined to be empty.
4039 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4040 that it reads its output file name as an argument instead of writing
4041 to the standard output and assuming this can be redirected.
4042
4043 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4044 class containing characters with values greater than 255, PCRE compilation
4045 went into a loop.
4046
4047 12. A recursive reference to a subpattern that was within another subpattern
4048 that had a minimum quantifier of zero caused PCRE to crash. For example,
4049 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4050 recursion. If the recursively-called subpattern itself had a zero repeat,
4051 that was OK.
4052
4053 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4054 buffer into which it was copied (for escape processing) was still set at
4055 1024, so long lines caused crashes.
4056
4057 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4058 "internal error: code overflow...". This applied to any character class
4059 that was followed by a possessive quantifier.
4060
4061 15. Modified the Makefile to add libpcre.la as a prerequisite for
4062 libpcreposix.la because I was told this is needed for a parallel build to
4063 work.
4064
4065 16. If a pattern that contained .* following optional items at the start was
4066 studied, the wrong optimizing data was generated, leading to matching
4067 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4068 matching string must start with a or b or c. The correct conclusion for
4069 this pattern is that a match can start with any character.
4070
4071
4072 Version 4.4 13-Aug-03
4073 ---------------------
4074
4075 1. In UTF-8 mode, a character class containing characters with values between
4076 127 and 255 was not handled correctly if the compiled pattern was studied.
4077 In fixing this, I have also improved the studying algorithm for such
4078 classes (slightly).
4079
4080 2. Three internal functions had redundant arguments passed to them. Removal
4081 might give a very teeny performance improvement.
4082
4083 3. Documentation bug: the value of the capture_top field in a callout is *one
4084 more than* the number of the hightest numbered captured substring.
4085
4086 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4087 in incorrectly linking with a previously installed version. They now link
4088 explicitly with libpcre.la.
4089
4090 5. configure.in no longer needs to recognize Cygwin specially.
4091
4092 6. A problem in pcre.in for Windows platforms is fixed.
4093
4094 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4095 pcretest, it used to include the size of the study block as part of its
4096 output. Unfortunately, the structure contains a field that has a different
4097 size on different hardware architectures. This meant that the tests that
4098 showed this size failed. As the block is currently always of a fixed size,
4099 this information isn't actually particularly useful in pcretest output, so
4100 I have just removed it.
4101
4102 8. Three pre-processor statements accidentally did not start in column 1.
4103 Sadly, there are *still* compilers around that complain, even though
4104 standard C has not required this for well over a decade. Sigh.
4105
4106 9. In pcretest, the code for checking callouts passed small integers in the
4107 callout_data field, which is a void * field. However, some picky compilers
4108 complained about the casts involved for this on 64-bit systems. Now
4109 pcretest passes the address of the small integer instead, which should get
4110 rid of the warnings.
4111
4112 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4113 both compile and run time, and gives an error if an invalid UTF-8 sequence
4114 is found. There is a option for disabling this check in cases where the
4115 string is known to be correct and/or the maximum performance is wanted.
4116
4117 11. In response to a bug report, I changed one line in Makefile.in from
4118
4119 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4120 to
4121 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4122
4123 to look similar to other lines, but I have no way of telling whether this
4124 is the right thing to do, as I do not use Windows. No doubt I'll get told
4125 if it's wrong...
4126
4127
4128 Version 4.3 21-May-03
4129 ---------------------
4130
4131 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4132 Makefile.
4133
4134 2. Some refactoring to improve the quality of the code:
4135
4136 (i) The utf8_table... variables are now declared "const".
4137
4138 (ii) The code for \cx, which used the "case flipping" table to upper case
4139 lower case letters, now just substracts 32. This is ASCII-specific,
4140 but the whole concept of \cx is ASCII-specific, so it seems
4141 reasonable.
4142
4143 (iii) PCRE was using its character types table to recognize decimal and
4144 hexadecimal digits in the pattern. This is silly, because it handles
4145 only 0-9, a-f, and A-F, but the character types table is locale-
4146 specific, which means strange things might happen. A private
4147 table is now used for this - though it costs 256 bytes, a table is
4148 much faster than multiple explicit tests. Of course, the standard
4149 character types table is still used for matching digits in subject
4150 strings against \d.
4151
4152 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4153 ending in _t are). So I've renamed it as ESC_tee.
4154
4155 3. The first argument for regexec() in the POSIX wrapper should have been
4156 defined as "const".
4157
4158 4. Changed pcretest to use malloc() for its buffers so that they can be
4159 Electric Fenced for debugging.
4160
4161 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4162 to read one or more bytes before the start of the subject string. Often this
4163 had no effect on PCRE's behaviour, but in some circumstances it could
4164 provoke a segmentation fault.
4165
4166 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4167 to try to read one or more bytes before the start of the subject string.
4168
4169 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4170 UTF-8 support could misbehave in various ways if the subject string
4171 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4172 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4173 back over UTF-8 characters.)
4174
4175
4176 Version 4.2 14-Apr-03
4177 ---------------------
4178
4179 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4180
4181 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4182 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4183 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4184 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4185 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4186 and BUILD_EXEEXT
4187 Note: automatic setting of the BUILD variables is not yet working
4188 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4189 compile-time but not at link-time
4190 [LINK]: use for linking executables only
4191 make different versions for Windows and non-Windows
4192 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4193 libraries
4194 [LINK_FOR_BUILD]: new variable
4195 [OBJEXT]: use throughout
4196 [EXEEXT]: use throughout
4197 <winshared>: new target
4198 <wininstall>: new target
4199 <dftables.o>: use native compiler
4200 <dftables>: use native linker
4201 <install>: handle Windows platform correctly
4202 <clean>: ditto
4203 <check>: ditto
4204 copy DLL to top builddir before testing
4205
4206 As part of these changes, -no-undefined was removed again. This was reported
4207 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4208 in any case.
4209
4210 3. Some tidies to get rid of compiler warnings:
4211
4212 . In the match_data structure, match_limit was an unsigned long int, whereas
4213 match_call_count was an int. I've made them both unsigned long ints.
4214
4215 . In pcretest the fact that a const uschar * doesn't automatically cast to
4216 a void * provoked a warning.
4217
4218 . Turning on some more compiler warnings threw up some "shadow" variables
4219 and a few more missing casts.
4220
4221 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4222 option, a class that contained a single character with a value between 128
4223 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4224
4225 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4226 option, a class that contained several characters, but with at least one
4227 whose value was between 128 and 255 caused PCRE to crash.
4228
4229
4230 Version 4.1 12-Mar-03
4231 ---------------------
4232
4233 1. Compiling with gcc -pedantic found a couple of places where casts were
4234 needed, and a string in dftables.c that was longer than standard compilers are
4235 required to support.
4236
4237 2. Compiling with Sun's compiler found a few more places where the code could
4238 be tidied up in order to avoid warnings.
4239
4240 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4241 first of these names is deprecated in the latest Autoconf in favour of the name
4242 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4243 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4244 analogy I have changed it to CFLAGS_FOR_BUILD.
4245
4246 4. Added -no-undefined to the linking command in the Makefile, because this is
4247 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4248 linking step for the pcreposix library.
4249
4250 5. PCRE was failing to diagnose the case of two named groups with the same
4251 name.
4252
4253 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4254 literal character that is needed in the subject for a match, and scans along to
4255 ensure that it is present before embarking on the full matching process. This
4256 saves time in cases of nested unlimited repeats that are never going to match.
4257 Problem: the scan can take a lot of time if the subject is very long (e.g.
4258 megabytes), thus penalizing straightforward matches. It is now done only if the
4259 amount of subject to be scanned is less than 1000 bytes.
4260
4261 7. A lesser problem with the same optimization is that it was recording the
4262 first character of an anchored pattern as "needed", thus provoking a search
4263 right along the subject, even when the first match of the pattern was going to
4264 fail. The "needed" character is now not set for anchored patterns, unless it
4265 follows something in the pattern that is of non-fixed length. Thus, it still
4266 fulfils its original purpose of finding quick non-matches in cases of nested
4267 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4268
4269
4270 Version 4.0 17-Feb-03
4271 ---------------------
4272
4273 1. If a comment in an extended regex that started immediately after a meta-item
4274 extended to the end of string, PCRE compiled incorrect data. This could lead to
4275 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4276
4277 2. Moved to autoconf 2.53 and libtool 1.4.2.
4278
4279 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4280 the special perltest8 script is no longer needed - all the tests can be run
4281 from a single perltest script.
4282
4283 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4284 by \s. It has now been removed in PCRE. This means it isn't recognized as
4285 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4286 class [:space:] *does* include VT, thereby creating a mess.
4287
4288 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4289 space and tab.
4290
4291 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4292 its new features into the main test script, reducing the number of scripts.
4293
4294 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4295 were backward compatible, and made the (?i) apply to the whole pattern, as if
4296 /i were given. Now it behaves more logically, and applies the option setting
4297 only to what follows. PCRE has been changed to follow suit. However, if it
4298 finds options settings right at the start of the pattern, it extracts them into
4299 the global options, as before. Thus, they show up in the info data.
4300
4301 8. Added support for the \Q...\E escape sequence. Characters in between are
4302 treated as literals. This is slightly different from Perl in that $ and @ are
4303 also handled as literals inside the quotes. In Perl, they will cause variable
4304 interpolation. Note the following examples:
4305
4306 Pattern PCRE matches Perl matches
4307
4308 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4309 \Qabc\$xyz\E abc\$xyz abc\$xyz
4310 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4311
4312 For compatibility with Perl, \Q...\E sequences are recognized inside character
4313 classes as well as outside them.
4314
4315 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4316 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4317 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4318 signed/unsigned warnings.
4319
4320 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4321 option for pcretest, so I've replaced it by a simple function that does just
4322 that job.
4323
4324 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4325 "pcregrep -".
4326
4327 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4328 Java package. This provides some syntactic sugar for simple cases of what my
4329 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4330 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4331 item, you can use this simplified notation. Note that only makes sense with
4332 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4333 greediness, whatever the setting of the PCRE_UNGREEDY option.
4334
4335 13. A change of greediness default within a pattern was not taking effect at
4336 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4337 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4338 was abstracted outside.
4339
4340 14. PCRE now supports the \G assertion. It is true when the current matching
4341 position is at the start point of the match. This differs from \A when the
4342 starting offset is non-zero. Used with the /g option of pcretest (or similar
4343 code), it works in the same way as it does for Perl's /g option. If all
4344 alternatives of a regex begin with \G, the expression is anchored to the start
4345 match position, and the "anchored" flag is set in the compiled expression.
4346
4347 15. Some bugs concerning the handling of certain option changes within patterns
4348 have been fixed. These applied to options other than (?ims). For example,
4349 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4350 been the other way round. Some of this was related to change 7 above.
4351
4352 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4353 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4354 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4355 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4356
4357 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4358 mode. Unlike ".", it always matches newline, whatever the setting of
4359 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4360 assertions. Perl allows it, but it doesn't (in general) work because it can't
4361 calculate the length of the lookbehind. At least, that's the case for Perl
4362 5.8.0 - I've been told they are going to document that it doesn't work in
4363 future.
4364
4365 18. Added an error diagnosis for escapes that PCRE does not support: these are
4366 \L, \l, \N, \P, \p, \U, \u, and \X.
4367
4368 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4369 reading past the end of the pattern in cases such as /[abcd/.
4370
4371 20. PCRE was getting more memory than necessary for patterns with classes that
4372 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4373
4374 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4375 compiling PCRE for use with Virtual Pascal.
4376
4377 22. Small fix to the Makefile to make it work properly if the build is done
4378 outside the source tree.
4379
4380 23. Added a new extension: a condition to go with recursion. If a conditional
4381 subpattern starts with (?(R) the "true" branch is used if recursion has
4382 happened, whereas the "false" branch is used only at the top level.
4383
4384 24. When there was a very long string of literal characters (over 255 bytes
4385 without UTF support, over 250 bytes with UTF support), the computation of how
4386 much memory was required could be incorrect, leading to segfaults or other
4387 strange effects.
4388
4389 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4390 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4391 there was a subsequent back reference to those brackets. This meant that, for
4392 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4393 possible to check for precisely this case. All we can do is abandon the
4394 optimization if .* occurs inside capturing brackets when there are any back
4395 references whatsoever. (See below for a better fix that came later.)
4396
4397 26. The handling of the optimization for finding the first character of a
4398 non-anchored pattern, and for finding a character that is required later in the
4399 match were failing in some cases. This didn't break the matching; it just
4400 failed to optimize when it could. The way this is done has been re-implemented.
4401
4402 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4403
4404 28. Added a new feature that provides some of the functionality that Perl
4405 provides with (?{...}). The facility is termed a "callout". The way it is done
4406 in PCRE is for the caller to provide an optional function, by setting
4407 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4408 global variable. By default it is unset, which disables all calling out. To get
4409 the function called, the regex must include (?C) at appropriate points. This
4410 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4411 This provides a means of identifying different callout points. When PCRE
4412 reaches such a point in the regex, if pcre_callout has been set, the external
4413 function is called. It is provided with data in a structure called
4414 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4415 matching continues; if it returns a non-zero value, the match at the current
4416 point fails. However, backtracking will occur if possible. [This was changed
4417 later and other features added - see item 49 below.]
4418
4419 29. pcretest is upgraded to test the callout functionality. It provides a
4420 callout function that displays information. By default, it shows the start of
4421 the match and the current position in the text. There are some new data escapes
4422 to vary what happens:
4423
4424 \C+ in addition, show current contents of captured substrings
4425 \C- do not supply a callout function
4426 \C!n return 1 when callout number n is reached
4427 \C!n!m return 1 when callout number n is reached for the mth time
4428
4429 30. If pcregrep was called with the -l option and just a single file name, it
4430 output "<stdin>" if a match was found, instead of the file name.
4431
4432 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4433 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4434 pcre_exec(). This saves a malloc/free per call. The default value of
4435 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4436 when configuring.
4437
4438 32. The default maximum size of a compiled pattern is 64K. There have been a
4439 few cases of people hitting this limit. The code now uses macros to handle the
4440 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4441 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4442 configuring. Tests 2 and 5 work only with 2-byte links because they output
4443 debugging information about compiled patterns.
4444
4445 33. Internal code re-arrangements:
4446
4447 (a) Moved the debugging function for printing out a compiled regex into
4448 its own source file (printint.c) and used #include to pull it into
4449 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4450 separate copies.
4451
4452 (b) Defined the list of op-code names for debugging as a macro in
4453 internal.h so that it is next to the definition of the opcodes.
4454
4455 (c) Defined a table of op-code lengths for simpler skipping along compiled
4456 code. This is again a macro in internal.h so that it is next to the
4457 definition of the opcodes.
4458
4459 34. Added support for recursive calls to individual subpatterns, along the
4460 lines of Robin Houston's patch (but implemented somewhat differently).
4461
4462 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4463 allow it to read and process whole directories in Win32. This code was
4464 contributed by Lionel Fourquaux; it has not been tested by me.
4465
4466 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4467 used to name a group. Names consist of alphanumerics and underscores, and must
4468 be unique. Back references use the syntax (?P=name) and recursive calls use
4469 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4470 numbers. The function pcre_fullinfo() can be used after compilation to extract
4471 a name/number map. There are three relevant calls:
4472
4473 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4474 PCRE_INFO_NAMECOUNT yields the number of entries
4475 PCRE_INFO_NAMETABLE yields a pointer to the map.
4476
4477 The map is a vector of fixed-size entries. The size of each entry depends on
4478 the length of the longest name used. The first two bytes of each entry are the
4479 group number, most significant byte first. There follows the corresponding
4480 name, zero terminated. The names are in alphabetical order.
4481
4482 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4483 case instead of 255. Making it the same both with and without UTF-8 support
4484 means that the same test output works with both.
4485
4486 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4487 calling malloc() with a zero argument.
4488
4489 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4490 optimization. I've improved things by keeping a bitmap of backreferences with
4491 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4492 fact referenced, the optimization can be applied. It is unlikely that a
4493 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4494 the match to follow \n) will appear inside brackets with a number greater than
4495 31, but if it does, any back reference > 31 suppresses the optimization.
4496
4497 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4498 of disabling numbered capturing parentheses. Any opening parenthesis that is
4499 not followed by ? behaves as if it were followed by ?: but named parentheses
4500 can still be used for capturing (and they will acquire numbers in the usual
4501 way).
4502
4503 41. Redesigned the return codes from the match() function into yes/no/error so
4504 that errors can be passed back from deep inside the nested calls. A malloc
4505 failure while inside a recursive subpattern call now causes the
4506 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4507
4508 42. It is now possible to set a limit on the number of times the match()
4509 function is called in a call to pcre_exec(). This facility makes it possible to
4510 limit the amount of recursion and backtracking, though not in a directly
4511 obvious way, because the match() function is used in a number of different
4512 circumstances. The count starts from zero for each position in the subject
4513 string (for non-anchored patterns). The default limit is, for compatibility, a
4514 large number, namely 10 000 000. You can change this in two ways:
4515
4516 (a) When configuring PCRE before making, you can use --with-match-limit=n
4517 to set a default value for the compiled library.
4518
4519 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4520 a different value is set. See 45 below.
4521
4522 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4523
4524 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4525 of things that can be changed at compile time. The first argument specifies
4526 what is wanted and the second points to where the information is to be placed.
4527 The current list of available information is:
4528
4529 PCRE_CONFIG_UTF8
4530
4531 The output is an integer that is set to one if UTF-8 support is available;
4532 otherwise it is set to zero.
4533
4534 PCRE_CONFIG_NEWLINE
4535
4536 The output is an integer that it set to the value of the code that is used for
4537 newline. It is either LF (10) or CR (13).
4538
4539 PCRE_CONFIG_LINK_SIZE
4540
4541 The output is an integer that contains the number of bytes used for internal
4542 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4543
4544 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4545
4546 The output is an integer that contains the threshold above which the POSIX
4547 interface uses malloc() for output vectors. See item 31 above.
4548
4549 PCRE_CONFIG_MATCH_LIMIT
4550
4551 The output is an unsigned integer that contains the default limit of the number
4552 of match() calls in a pcre_exec() execution. See 42 above.
4553
4554 44. pcretest has been upgraded by the addition of the -C option. This causes it
4555 to extract all the available output from the new pcre_config() function, and to
4556 output it. The program then exits immediately.
4557
4558 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4559 order to support additional features. One way would have been to define
4560 pcre_exec2() (for example) with extra arguments, but this would not have been
4561 extensible, and would also have required all calls to the original function to
4562 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4563 is used for passing in "extra" data from pcre_study().
4564
4565 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4566 contains the following fields:
4567
4568 flags a bitmap indicating which of the following fields are set
4569 study_data opaque data from pcre_study()
4570 match_limit a way of specifying a limit on match() calls for a specific
4571 call to pcre_exec()
4572 callout_data data for callouts (see 49 below)
4573
4574 The flag bits are also defined in pcre.h, and are
4575
4576 PCRE_EXTRA_STUDY_DATA
4577 PCRE_EXTRA_MATCH_LIMIT
4578 PCRE_EXTRA_CALLOUT_DATA
4579
4580 The pcre_study() function now returns one of these new pcre_extra blocks, with
4581 the actual study data pointed to by the study_data field, and the
4582 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4583 before. That is, this change is entirely upwards-compatible and requires no
4584 change to existing code.
4585
4586 If you want to pass in additional data to pcre_exec(), you can either place it
4587 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4588 block.
4589
4590 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4591 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4592 times with different match limits, until it finds the minimum value needed for
4593 pcre_exec() to complete. The value is then output. This can be instructive; for
4594 most simple matches the number is quite small, but for pathological cases it
4595 gets very large very quickly.
4596
4597 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4598 returns the size of the data block pointed to by the study_data field in a
4599 pcre_extra block, that is, the value that was passed as the argument to
4600 pcre_malloc() when PCRE was getting memory in which to place the information
4601 created by pcre_study(). The fourth argument should point to a size_t variable.
4602 pcretest has been extended so that this information is shown after a successful
4603 pcre_study() call when information about the compiled regex is being displayed.
4604
4605 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4606 because what follows is always an absolute path. (Later: it turns out that this
4607 is more than cosmetic for MinGW, because it doesn't like empty path
4608 components.)
4609
4610 49. Some changes have been made to the callout feature (see 28 above):
4611
4612 (i) A callout function now has three choices for what it returns:
4613
4614 0 => success, carry on matching
4615 > 0 => failure at this point, but backtrack if possible
4616 < 0 => serious error, return this value from pcre_exec()
4617
4618 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4619 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4620 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4621 use by callout functions. It will never be used by PCRE itself.
4622
4623 (ii) The pcre_extra structure (see 45 above) has a void * field called
4624 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4625 pcre_callout_block structure has a field of the same name. The contents of
4626 the field passed in the pcre_extra structure are passed to the callout
4627 function in the corresponding field in the callout block. This makes it
4628 easier to use the same callout-containing regex from multiple threads. For
4629 testing, the pcretest program has a new data escape
4630
4631 \C*n pass the number n (may be negative) as callout_data
4632
4633 If the callout function in pcretest receives a non-zero value as
4634 callout_data, it returns that value.
4635
4636 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4637 there were some redundant $(CFLAGS) in commands that are now specified as
4638 $(LINK), which already includes $(CFLAGS).
4639
4640 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4641 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4642 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4643 one-byte characters throughout. Note that case-insensitive matching applies
4644 only to characters whose values are less than 256. PCRE doesn't support the
4645 notion of cases for higher-valued characters.
4646
4647 (i) A character class whose characters are all within 0-255 is handled as
4648 a bit map, and the map is inverted for negative classes. Previously, a
4649 character > 255 always failed to match such a class; however it should
4650 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4651
4652 (ii) A negated character class with a single character < 255 is coded as
4653 "not this character" (OP_NOT). This wasn't working properly when the test
4654 character was multibyte, either singly or repeated.
4655
4656 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4657 mode, for example: \x{100}{2,3}.
4658
4659 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4660 singly or repeated) now correctly test multibyte characters. However,
4661 PCRE doesn't recognize any characters with values greater than 255 as
4662 digits, spaces, or word characters. Such characters always match \D, \S,
4663 and \W, and never match \d, \s, or \w.
4664
4665 (v) Classes may now contain characters and character ranges with values
4666 greater than 255. For example: [ab\x{100}-\x{400}].
4667
4668 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4669 PCRE in UTF-8 mode.
4670
4671 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4672 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4673 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4674 value.)
4675
4676 53. The single man page has become too large. I have therefore split it up into
4677 a number of separate man pages. These also give rise to individual HTML pages;
4678 these are now put in a separate directory, and there is an index.html page that
4679 lists them all. Some hyperlinking between the pages has been installed.
4680
4681 54. Added convenience functions for handling named capturing parentheses.
4682
4683 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4684 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4685 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4686 are faulted.
4687
4688 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4689 calling configure. These values are used when compiling the dftables.c program
4690 which is run to generate the source of the default character tables. They
4691 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4692 you will need to set these values.
4693
4694 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4695
4696
4697 Version 3.9 02-Jan-02
4698 ---------------------
4699
4700 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4701
4702 2. If --disable-static was given, the building process failed when trying to
4703 build pcretest and pcregrep. (For some reason it was using libtool to compile
4704 them, which is not right, as they aren't part of the library.)
4705
4706
4707 Version 3.8 18-Dec-01
4708 ---------------------
4709
4710 1. The experimental UTF-8 code was completely screwed up. It was packing the
4711 bytes in the wrong order. How dumb can you get?
4712
4713
4714 Version 3.7 29-Oct-01
4715 ---------------------
4716
4717 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4718 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4719 this didn't happen under Solaris 8, where I normally test things.
4720
4721 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4722 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4723 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4724 this fix an hour or so after the initial 3.7 release.)
4725
4726
4727 Version 3.6 23-Oct-01
4728 ---------------------
4729
4730 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4731 offsets passed as NULL with zero offset count.
4732
4733 2. The config.guess and config.sub files had not been updated when I moved to
4734 the latest autoconf.
4735
4736
4737 Version 3.5 15-Aug-01
4738 ---------------------
4739
4740 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4741 had been forgotten.
4742
4743 2. By using declared but undefined structures, we can avoid using "void"
4744 definitions in pcre.h while keeping the internal definitions of the structures
4745 private.
4746
4747 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4748 user point of view, this means that both static and shared libraries are built
4749 by default, but this can be individually controlled. More of the work of
4750 handling this static/shared cases is now inside libtool instead of PCRE's make
4751 file.
4752
4753 4. The pcretest utility is now installed along with pcregrep because it is
4754 useful for users (to test regexs) and by doing this, it automatically gets
4755 relinked by libtool. The documentation has been turned into a man page, so
4756 there are now .1, .txt, and .html versions in /doc.
4757
4758 5. Upgrades to pcregrep:
4759 (i) Added long-form option names like gnu grep.
4760 (ii) Added --help to list all options with an explanatory phrase.
4761 (iii) Added -r, --recursive to recurse into sub-directories.
4762 (iv) Added -f, --file to read patterns from a file.
4763
4764 6. pcre_exec() was referring to its "code" argument before testing that
4765 argument for NULL (and giving an error if it was NULL).
4766
4767 7. Upgraded Makefile.in to allow for compiling in a different directory from
4768 the source directory.
4769
4770 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4771 options bits, the pointer it was passed was to an int instead of to an unsigned
4772 long int. This mattered only on 64-bit systems.
4773
4774 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4775 generated) instead of pcre.in, which it its source. Also made the same change
4776 in several of the .c files.
4777
4778 10. A new release of gcc defines printf() as a macro, which broke pcretest
4779 because it had an ifdef in the middle of a string argument for printf(). Fixed
4780 by using separate calls to printf().
4781
4782 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4783 script, to force use of CR or LF instead of \n in the source. On non-Unix
4784 systems, the value can be set in config.h.
4785
4786 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4787 absolute limit. Changed the text of the error message to make this clear, and
4788 likewise updated the man page.
4789
4790 13. The limit of 99 on the number of capturing subpatterns has been removed.
4791 The new limit is 65535, which I hope will not be a "real" limit.
4792
4793
4794 Version 3.4 22-Aug-00
4795 ---------------------
4796
4797 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4798
4799 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4800
4801
4802 Version 3.3 01-Aug-00
4803 ---------------------
4804
4805 1. If an octal character was given, but the value was greater than \377, it
4806 was not getting masked to the least significant bits, as documented. This could
4807 lead to crashes in some systems.
4808
4809 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4810 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4811
4812 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4813 These just pass their arguments on to (pcre_free)(), but they are provided
4814 because some uses of PCRE bind it to non-C systems that can call its functions,
4815 but cannot call free() or pcre_free() directly.
4816
4817 4. Add "make test" as a synonym for "make check". Corrected some comments in
4818 the Makefile.
4819
4820 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4821 Makefile.
4822
4823 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4824 command called pgrep for grepping around the active processes.
4825
4826 7. Added the beginnings of support for UTF-8 character strings.
4827
4828 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4829 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4830 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4831 out for the ar command.)
4832
4833
4834 Version 3.2 12-May-00
4835 ---------------------
4836
4837 This is purely a bug fixing release.
4838
4839 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4840 of ZA. This was just one example of several cases that could provoke this bug,
4841 which was introduced by change 9 of version 2.00. The code for breaking
4842 infinite loops after an iteration that matches an empty string was't working
4843 correctly.
4844
4845 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4846 when matched against abbab (for example). After matching an empty string, it
4847 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4848 caused it to match further down the string than it should.
4849
4850 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4851 was there because it doesn't seem to be needed, and it causes trouble on some
4852 systems, as it is not a Standard C header. It has been removed.
4853
4854 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4855 were reported on the Macintosh. The changes were from
4856
4857 while ((c = *(++ptr)) != 0 && c != '\n');
4858 to
4859 while ((c = *(++ptr)) != 0 && c != '\n') ;
4860
4861 Totally extraordinary, but if that's what it takes...
4862
4863 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4864 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4865 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4866 assumes the way PCRE uses memmove() (always moving upwards).
4867
4868 6. PCRE is being used in one environment where strchr() is not available. There
4869 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4870 faster code anyway.
4871
4872
4873 Version 3.1 09-Feb-00
4874 ---------------------
4875
4876 The only change in this release is the fixing of some bugs in Makefile.in for
4877 the "install" target:
4878
4879 (1) It was failing to install pcreposix.h.
4880
4881 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4882
4883
4884 Version 3.0 01-Feb-00
4885 ---------------------
4886
4887 1. Add support for the /+ modifier to perltest (to output $` like it does in
4888 pcretest).
4889
4890 2. Add support for the /g modifier to perltest.
4891
4892 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4893 matches null strings.
4894
4895 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4896 pattern. Perl treats empty patterns specially - it reuses the most recent
4897 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4898 effect.
4899
4900 5. The POSIX interface was broken in that it was just handing over the POSIX
4901 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4902 required a bigger vector, with some working space on the end. This means that
4903 the POSIX wrapper now has to get and free some memory, and copy the results.
4904
4905 6. Added some simple autoconf support, placing the test data and the
4906 documentation in separate directories, re-organizing some of the
4907 information files, and making it build pcre-config (a GNU standard). Also added
4908 libtool support for building PCRE as a shared library, which is now the
4909 default.
4910
4911 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4912 09 are not valid octal constants. Single digits will be used for minor values
4913 less than 10.
4914
4915 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4916 existing programs that set these in the POSIX interface can use PCRE without
4917 modification.
4918
4919 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4920 return all that pcre_info() returns, plus additional data. The pcre_info()
4921 function is retained for compatibility, but is considered to be obsolete.
4922
4923 10. Added experimental recursion feature (?R) to handle one common case that
4924 Perl 5.6 will be able to do with (?p{...}).
4925
4926 11. Added support for POSIX character classes like [:alpha:], which Perl is
4927 adopting.
4928
4929
4930 Version 2.08 31-Aug-99
4931 ----------------------
4932
4933 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4934 trying to match at the startoffset position, but instead was moving forward to
4935 the next newline as if a previous match had failed.
4936
4937 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4938 and could get into a loop if a null string was matched other than at the start
4939 of the subject.
4940
4941 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4942 be distinguished at compile time, and for completeness also added PCRE_DATE.
4943
4944 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4945 in GnuWin32 environments.
4946
4947
4948 Version 2.07 29-Jul-99
4949 ----------------------
4950
4951 1. The documentation is now supplied in plain text form and HTML as well as in
4952 the form of man page sources.
4953
4954 2. C++ compilers don't like assigning (void *) values to other pointer types.
4955 In particular this affects malloc(). Although there is no problem in Standard
4956 C, I've put in casts to keep C++ compilers happy.
4957
4958 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4959 should be (const char *).
4960
4961 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4962 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4963 However, I haven't made this a standard facility. The documentation doesn't
4964 mention it, and the Makefile doesn't support it.
4965
4966 5. The Makefile now contains an "install" target, with editable destinations at
4967 the top of the file. The pcretest program is not installed.
4968
4969 6. pgrep -V now gives the PCRE version number and date.
4970
4971 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4972 causing the entire string to be ignored, instead of just the last character.
4973
4974 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4975 non-matching string, it can take a very, very long time, even for strings of
4976 quite modest length, because of the nested recursion. PCRE now does better in
4977 some of these cases. It does this by remembering the last required literal
4978 character in the pattern, and pre-searching the subject to ensure it is present
4979 before running the real match. In other words, it applies a heuristic to detect
4980 some types of certain failure quickly, and in the above example, if presented
4981 with a string that has no trailing " it gives "no match" very quickly.
4982
4983 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4984 other alternatives are tried instead.
4985
4986
4987 Version 2.06 09-Jun-99
4988 ----------------------
4989
4990 1. Change pcretest's output for amount of store used to show just the code
4991 space, because the remainder (the data block) varies in size between 32-bit and
4992 64-bit systems.
4993
4994 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4995 start matching at. This allows lookbehinds to work when searching for multiple
4996 occurrences in a string.
4997
4998 3. Added additional options to pcretest for testing multiple occurrences:
4999
5000 /+ outputs the rest of the string that follows a match
5001 /g loops for multiple occurrences, using the new startoffset argument
5002 /G loops for multiple occurrences by passing an incremented pointer
5003
5004 4. PCRE wasn't doing the "first character" optimization for patterns starting
5005 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5006 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5007 the letter 'x'. On long subject strings, this gives a significant speed-up.
5008
5009
5010 Version 2.05 21-Apr-99
5011 ----------------------
5012
5013 1. Changed the type of magic_number from int to long int so that it works
5014 properly on 16-bit systems.
5015
5016 2. Fixed a bug which caused patterns starting with .* not to work correctly
5017 when the subject string contained newline characters. PCRE was assuming
5018 anchoring for such patterns in all cases, which is not correct because .* will
5019 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5020 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5021 must be retried after every newline in the subject.
5022
5023
5024 Version 2.04 18-Feb-99
5025 ----------------------
5026
5027 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5028 computation of the store needed to hold the pattern was incorrect (too large).
5029 If such patterns were nested a few deep, this could multiply and become a real
5030 problem.
5031
5032 2. Added /M option to pcretest to show the memory requirement of a specific
5033 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5034
5035 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5036 compiled in such a way that the backtracking after subsequent failure was
5037 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5038 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5039
5040
5041 Version 2.03 02-Feb-99
5042 ----------------------
5043
5044 1. Fixed typo and small mistake in man page.
5045
5046 2. Added 4th condition (GPL supersedes if conflict) and created separate
5047 LICENCE file containing the conditions.
5048
5049 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5050 Perl, that is the internal \ allows the delimiter to be included in the
5051 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5052 the final delimiter, add \ to the end of the pattern (to test the error).
5053
5054 4. Added the convenience functions for extracting substrings after a successful
5055 match. Updated pcretest to make it able to test these functions.
5056
5057
5058 Version 2.02 14-Jan-99
5059 ----------------------
5060
5061 1. Initialized the working variables associated with each extraction so that
5062 their saving and restoring doesn't refer to uninitialized store.
5063
5064 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5065 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5066 fix the problem.
5067
5068 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5069 calls, and wasn't printing the correct value for compiling calls. Increased the
5070 default value of LOOPREPEAT, and the number of significant figures in the
5071 times.
5072
5073 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5074
5075 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5076 a building problem on Windows NT with a FAT file system.
5077
5078
5079 Version 2.01 21-Oct-98
5080 ----------------------
5081
5082 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5083 to character tables built by pcre_maketables() in the current locale. If NULL
5084 is passed, the default tables are used.
5085
5086
5087 Version 2.00 24-Sep-98
5088 ----------------------
5089
5090 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5091 it any more.
5092
5093 2. Allow quantification of (?>) groups, and make it work correctly.
5094
5095 3. The first character computation wasn't working for (?>) groups.
5096
5097 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5098 end of the subject) and add 5.005's \z, which really does match only at the
5099 very end of the subject.
5100
5101 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5102
5103 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5104 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5105 localized options. All options to pcre_study() were also removed.
5106
5107 7. Add other new features from 5.005:
5108
5109 $(?<= positive lookbehind
5110 $(?<! negative lookbehind
5111 (?imsx-imsx) added the unsetting capability
5112 such a setting is global if at outer level; local otherwise
5113 (?imsx-imsx:) non-capturing groups with option setting
5114 (?(cond)re|re) conditional pattern matching
5115
5116 A backreference to itself in a repeated group matches the previous
5117 captured string.
5118
5119 8. General tidying up of studying (both automatic and via "study")
5120 consequential on the addition of new assertions.
5121
5122 9. As in 5.005, unlimited repeated groups that could match an empty substring
5123 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5124 runtime if any iteration does actually match an empty substring.
5125
5126 10. Include the RunTest script in the distribution.
5127
5128 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5129 discrepancies, some of which were old and were also with respect to 5.004. They
5130 have now been fixed.
5131
5132
5133 Version 1.09 28-Apr-98
5134 ----------------------
5135
5136 1. A negated single character class followed by a quantifier with a minimum
5137 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5138 program crashes, or just wrong answers. This did not apply to negated classes
5139 containing more than one character, or to minima other than one.
5140
5141
5142 Version 1.08 27-Mar-98
5143 ----------------------
5144
5145 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5146
5147 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5148 latter must appear before anything that relies on it in the pattern.
5149
5150
5151 Version 1.07 16-Feb-98
5152 ----------------------
5153
5154 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5155 repeat of a potentially empty string).
5156
5157
5158 Version 1.06 23-Jan-98
5159 ----------------------
5160
5161 1. Added Markus Oberhumer's little patches for C++.
5162
5163 2. Literal strings longer than 255 characters were broken.
5164
5165
5166 Version 1.05 23-Dec-97
5167 ----------------------
5168
5169 1. Negated character classes containing more than one character were failing if
5170 PCRE_CASELESS was set at run time.
5171
5172
5173 Version 1.04 19-Dec-97
5174 ----------------------
5175
5176 1. Corrected the man page, where some "const" qualifiers had been omitted.
5177
5178 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5179 input syntax.
5180
5181 3. Fixed memory leak which occurred when a regex with back references was
5182 matched with an offsets vector that wasn't big enough. The temporary memory
5183 that is used in this case wasn't being freed if the match failed.
5184
5185 4. Tidied pcretest to ensure it frees memory that it gets.
5186
5187 5. Temporary memory was being obtained in the case where the passed offsets
5188 vector was exactly big enough.
5189
5190 6. Corrected definition of offsetof() from change 5 below.
5191
5192 7. I had screwed up change 6 below and broken the rules for the use of
5193 setjmp(). Now fixed.
5194
5195
5196 Version 1.03 18-Dec-97
5197 ----------------------
5198
5199 1. A erroneous regex with a missing opening parenthesis was correctly
5200 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5201 on some systems.
5202
5203 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5204 it was reported that one broken compiler failed on the former because "code" is
5205 also an independent variable.
5206
5207 3. The erroneous regex a[]b caused an array overrun reference.
5208
5209 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5210 fail on data ending with that character. (It was going on too far, and checking
5211 the next character, typically a binary zero.) This was specific to the
5212 optimized code for single-character negative classes.
5213
5214 5. Added a contributed patch from the TIN world which does the following:
5215
5216 + Add an undef for memmove, in case the the system defines a macro for it.
5217
5218 + Add a definition of offsetof(), in case there isn't one. (I don't know
5219 the reason behind this - offsetof() is part of the ANSI standard - but
5220 it does no harm).
5221
5222 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5223 most of the places where whitespace preceded '#'. I have given up and
5224 allowed the remaining 2 cases to be at the margin.
5225
5226 + Rename some variables in pcre to eliminate shadowing. This seems very
5227 pedantic, but does no harm, of course.
5228
5229 6. Moved the call to setjmp() into its own function, to get rid of warnings
5230 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5231
5232 7. Constructs such as \d{8,} were compiling into the equivalent of
5233 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5234 outcome, but in this particular case used more store than had been allocated,
5235 which caused the bug to be discovered because it threw up an internal error.
5236
5237 8. The debugging code in both pcre and pcretest for outputting the compiled
5238 form of a regex was going wrong in the case of back references followed by
5239 curly-bracketed repeats.
5240
5241
5242 Version 1.02 12-Dec-97
5243 ----------------------
5244
5245 1. Typos in pcre.3 and comments in the source fixed.
5246
5247 2. Applied a contributed patch to get rid of places where it used to remove
5248 'const' from variables, and fixed some signed/unsigned and uninitialized
5249 variable warnings.
5250
5251 3. Added the "runtest" target to Makefile.
5252
5253 4. Set default compiler flag to -O2 rather than just -O.
5254
5255
5256 Version 1.01 19-Nov-97
5257 ----------------------
5258
5259 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5260 like /([ab]*)*/, that is, for classes with more than one character in them.
5261
5262 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5263 as /((?>a*))*/ (a PCRE_EXTRA facility).
5264
5265
5266 Version 1.00 18-Nov-97
5267 ----------------------
5268
5269 1. Added compile-time macros to support systems such as SunOS4 which don't have
5270 memmove() or strerror() but have other things that can be used instead.
5271
5272 2. Arranged that "make clean" removes the executables.
5273
5274
5275 Version 0.99 27-Oct-97
5276 ----------------------
5277
5278 1. Fixed bug in code for optimizing classes with only one character. It was
5279 initializing a 32-byte map regardless, which could cause it to run off the end
5280 of the memory it had got.
5281
5282 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5283
5284
5285 Version 0.98 22-Oct-97
5286 ----------------------
5287
5288 1. Fixed bug in code for handling temporary memory usage when there are more
5289 back references than supplied space in the ovector. This could cause segfaults.
5290
5291
5292 Version 0.97 21-Oct-97
5293 ----------------------
5294
5295 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5296
5297 2. Optimized negated single characters not to use a bit map.
5298
5299 3. Brought error texts together as macro definitions; clarified some of them;
5300 fixed one that was wrong - it said "range out of order" when it meant "invalid
5301 escape sequence".
5302
5303 4. Changed some char * arguments to const char *.
5304
5305 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5306
5307 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5308 pcretest.
5309
5310
5311 Version 0.96 16-Oct-97
5312 ----------------------
5313
5314 1. Added a simple "pgrep" utility to the distribution.
5315
5316 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5317 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5318 where "ddd" means "one or more decimal digits".
5319
5320 3. Fixed serious bug. If a pattern had a back reference, but the call to
5321 pcre_exec() didn't supply a large enough ovector to record the related
5322 identifying subpattern, the match always failed. PCRE now remembers the number
5323 of the largest back reference, and gets some temporary memory in which to save
5324 the offsets during matching if necessary, in order to ensure that
5325 backreferences always work.
5326
5327 4. Increased the compatibility with Perl in a number of ways:
5328
5329 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5330 to request this handling. The option can be set at compile or exec time.
5331
5332 (b) $ matches before a terminating newline by default; an option
5333 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5334 mode). The option can be set at compile or exec time.
5335
5336 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5337 the same as Perl's. If the decimal number it represents is less than 10
5338 or there aren't that many previous left capturing parentheses, an octal
5339 escape is read. Inside a character class, it's always an octal escape,
5340 even if it is a single digit.
5341
5342 (d) An escaped but undefined alphabetic character is taken as a literal,
5343 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5344 escapes.
5345
5346 (e) {0} is now permitted. (The previous item is removed from the compiled
5347 pattern).
5348
5349 5. Changed all the names of code files so that the basic parts are no longer
5350 than 10 characters, and abolished the teeny "globals.c" file.
5351
5352 6. Changed the handling of character classes; they are now done with a 32-byte
5353 bit map always.
5354
5355 7. Added the -d and /D options to pcretest to make it possible to look at the
5356 internals of compilation without having to recompile pcre.
5357
5358
5359 Version 0.95 23-Sep-97
5360 ----------------------
5361
5362 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5363 \x20 at the start of a run of normal characters. These were being treated as
5364 real characters, instead of the source characters being re-checked.
5365
5366
5367 Version 0.94 18-Sep-97
5368 ----------------------
5369
5370 1. The functions are now thread-safe, with the caveat that the global variables
5371 containing pointers to malloc() and free() or alternative functions are the
5372 same for all threads.
5373
5374 2. Get pcre_study() to generate a bitmap of initial characters for non-
5375 anchored patterns when this is possible, and use it if passed to pcre_exec().
5376
5377
5378 Version 0.93 15-Sep-97
5379 ----------------------
5380
5381 1. /(b)|(:+)/ was computing an incorrect first character.
5382
5383 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5384 but not actually doing anything yet.
5385
5386 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5387 as Perl does (e.g. [-az] or [az-]).
5388
5389 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5390 all possible positions.
5391
5392 5. Split up into different modules to avoid including unneeded functions in a
5393 compiled binary. However, compile and exec are still in one module. The "study"
5394 function is split off.
5395
5396 6. The character tables are now in a separate module whose source is generated
5397 by an auxiliary program - but can then be edited by hand if required. There are
5398 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5399 toupper() in the code.
5400
5401 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5402 make them global. Abolish the function for setting them, as the caller can now
5403 set them directly.
5404
5405
5406 Version 0.92 11-Sep-97
5407 ----------------------
5408
5409 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5410 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5411
5412 2. Caseless matching was not working in character classes if the characters in
5413 the pattern were in upper case.
5414
5415 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5416
5417 4. Make PCRE_ANCHORED public and accept as a compile option.
5418
5419 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5420 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5421 pass them.
5422
5423 6. Give an error if bad option bits passed at compile or run time.
5424
5425 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5426 pcretest to cause it to pass that flag.
5427
5428 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5429 options, and the first character, if set.
5430
5431 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5432
5433
5434 Version 0.91 10-Sep-97
5435 ----------------------
5436
5437 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5438 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5439
5440 2. PCRE was looping on encountering an indefinitely repeated back reference to
5441 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5442 Perl does - treats the match as successful.
5443
5444 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5