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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1415 - (show annotations)
Sun Dec 22 20:47:08 2013 UTC (6 years, 1 month ago) by zherczeg
File size: 258387 byte(s)
The auto-possessification of character sets were improved. The JIT compiler also optimizes more character set checks.
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
16 Version 8.34 15-December-2013
17 -----------------------------
18
19 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
20 executable memory. Patch inspired by Carsten Klein.
21
22 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
23 this macro is never tested and has no effect, because the work to support
24 coverage involves only compiling and linking options and special targets in
25 the Makefile. The comment in config.h implied that defining the macro would
26 enable coverage support, which is totally false. There was also support for
27 setting this macro in the CMake files (my fault, I just copied it from
28 configure). SUPPORT_GCOV has now been removed.
29
30 3. Make a small performance improvement in strlen16() and strlen32() in
31 pcretest.
32
33 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
34 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
35 these cases). There was also one in pcretest.c.
36
37 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
38
39 6. In UTF mode, the code for checking whether a group could match an empty
40 string (which is used for indefinitely repeated groups to allow for
41 breaking an infinite loop) was broken when the group contained a repeated
42 negated single-character class with a character that occupied more than one
43 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
44 UTF-8 mode). The effect was undefined: the group might or might not be
45 deemed as matching an empty string, or the program might have crashed.
46
47 7. The code for checking whether a group could match an empty string was not
48 recognizing that \h, \H, \v, \V, and \R must match a character.
49
50 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
51 an empty string. If it can, pcretest shows this in its information output.
52
53 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
54 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
55 matched by pcre_exec() without using JIT:
56
57 (a) If the rest of the pattern did not match after a maximal run of
58 grapheme clusters, the code for backing up to try with fewer of them
59 did not always back up over a full grapheme when characters that do not
60 have the modifier quality were involved, e.g. Hangul syllables.
61
62 (b) If the match point in a subject started with modifier character, and
63 there was no match, the code could incorrectly back up beyond the match
64 point, and potentially beyond the first character in the subject,
65 leading to a segfault or an incorrect match result.
66
67 10. A conditional group with an assertion condition could lead to PCRE
68 recording an incorrect first data item for a match if no other first data
69 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
70 first data item, and therefore matched "ca" after "c" instead of at the
71 start.
72
73 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
74 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
75
76 12. The source of pcregrep now includes z/OS-specific code so that it can be
77 compiled for z/OS as part of the special z/OS distribution.
78
79 13. Added the -T and -TM options to pcretest.
80
81 14. The code in pcre_compile.c for creating the table of named capturing groups
82 has been refactored. Instead of creating the table dynamically during the
83 actual compiling pass, the information is remembered during the pre-compile
84 pass (on the stack unless there are more than 20 named groups, in which
85 case malloc() is used) and the whole table is created before the actual
86 compile happens. This has simplified the code (it is now nearly 150 lines
87 shorter) and prepared the way for better handling of references to groups
88 with duplicate names.
89
90 15. A back reference to a named subpattern when there is more than one of the
91 same name now checks them in the order in which they appear in the pattern.
92 The first one that is set is used for the reference. Previously only the
93 first one was inspected. This change makes PCRE more compatible with Perl.
94
95 16. Unicode character properties were updated from Unicode 6.3.0.
96
97 17. The compile-time code for auto-possessification has been refactored, based
98 on a patch by Zoltan Herczeg. It now happens after instead of during
99 compilation. The code is cleaner, and more cases are handled. The option
100 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
101 options in pcretest are provided to set it. It can also be set by
102 (*NO_AUTO_POSSESS) at the start of a pattern.
103
104 18. The character VT has been added to the default ("C" locale) set of
105 characters that match \s and are generally treated as white space,
106 following this same change in Perl 5.18. There is now no difference between
107 "Perl space" and "POSIX space". Whether VT is treated as white space in
108 other locales depends on the locale.
109
110 19. The code for checking named groups as conditions, either for being set or
111 for being recursed, has been refactored (this is related to 14 and 15
112 above). Processing unduplicated named groups should now be as fast at
113 numerical groups, and processing duplicated groups should be faster than
114 before.
115
116 20. Two patches to the CMake build system, by Alexander Barkov:
117
118 (1) Replace the "source" command by "." in CMakeLists.txt because
119 "source" is a bash-ism.
120
121 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
122 without these the CMake build does not work on Solaris.
123
124 21. Perl has changed its handling of \8 and \9. If there is no previously
125 encountered capturing group of those numbers, they are treated as the
126 literal characters 8 and 9 instead of a binary zero followed by the
127 literals. PCRE now does the same.
128
129 22. Following Perl, added \o{} to specify codepoints in octal, making it
130 possible to specify values greater than 0777 and also making them
131 unambiguous.
132
133 23. Perl now gives an error for missing closing braces after \x{... instead of
134 treating the string as literal. PCRE now does the same.
135
136 24. RunTest used to grumble if an inappropriate test was selected explicitly,
137 but just skip it when running all tests. This make it awkward to run ranges
138 of tests when one of them was inappropriate. Now it just skips any
139 inappropriate tests, as it always did when running all tests.
140
141 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
142 character types such as \d or \w, too many callouts were inserted, and the
143 data that they returned was rubbish.
144
145 26. In UCP mode, \s was not matching two of the characters that Perl matches,
146 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
147 were matched by \h. The code has now been refactored so that the lists of
148 the horizontal and vertical whitespace characters used for \h and \v (which
149 are defined only in one place) are now also used for \s.
150
151 27. Add JIT support for the 64 bit TileGX architecture.
152 Patch by Jiong Wang (Tilera Corporation).
153
154 28. Possessive quantifiers for classes (both explicit and automatically
155 generated) now use special opcodes instead of wrapping in ONCE brackets.
156
157 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
158 (because it's meaningless), this was not happening when PCRE_CASELESS was
159 set. Not wrong, but inefficient.
160
161 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
162 properties for \w, \d, etc) is present in a test regex. Otherwise if the
163 test contains no characters greater than 255, Perl doesn't realise it
164 should be using Unicode semantics.
165
166 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
167 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
168 does in Unicode mode.
169
170 32. Added the "forbid" facility to pcretest so that putting tests into the
171 wrong test files can sometimes be quickly detected.
172
173 33. There is now a limit (default 250) on the depth of nesting of parentheses.
174 This limit is imposed to control the amount of system stack used at compile
175 time. It can be changed at build time by --with-parens-nest-limit=xxx or
176 the equivalent in CMake.
177
178 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
179 errors. Perl warns for these when in warning mode, but PCRE has no facility
180 for giving warnings.
181
182 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
183 However, this was not working for (?!) because it is optimized to (*FAIL),
184 for which PCRE does not allow quantifiers. The optimization is now disabled
185 when a quantifier follows (?!). I can't see any use for this, but it makes
186 things uniform.
187
188 36. Perl no longer allows group names to start with digits, so I have made this
189 change also in PCRE. It simplifies the code a bit.
190
191 37. In extended mode, Perl ignores spaces before a + that indicates a
192 possessive quantifier. PCRE allowed a space before the quantifier, but not
193 before the possessive +. It now does.
194
195 38. The use of \K (reset reported match start) within a repeated possessive
196 group such as (a\Kb)*+ was not working.
197
198 40. Document that the same character tables must be used at compile time and
199 run time, and that the facility to pass tables to pcre_exec() and
200 pcre_dfa_exec() is for use only with saved/restored patterns.
201
202 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
203 features for Builds with MSVC:
204
205 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
206 builds). The .rc files can be used to set FileDescription and many other
207 attributes.
208
209 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
210 This allows higher-level build scripts which want .pdb files to avoid
211 hard-coding the exact files needed."
212
213 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
214 mean "start of word" and "end of word", respectively, as a transition aid.
215
216 43. A minimizing repeat of a class containing codepoints greater than 255 in
217 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
218 compiled to use the heap for recursion.
219
220 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
221 is configured.
222
223
224 Version 8.33 28-May-2013
225 ------------------------
226
227 1. Added 'U' to some constants that are compared to unsigned integers, to
228 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
229 variables that are added to signed variables, to ensure the result is
230 signed and can be negated.
231
232 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
233 CMake config files.
234
235 3. Revise the creation of config.h.generic so that all boolean macros are
236 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
237 overriding via -D on the command line possible.
238
239 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
240 to unsigned int is reported to make a quite noticeable speed difference in
241 a specific Windows environment. Testing on Linux did also appear to show
242 some benefit (and it is clearly not harmful). Also fixed the definition of
243 Xop which should be unsigned.
244
245 5. Related to (4), changing the definition of the intermediate variable cc
246 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
247 improvements.
248
249 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
250 unnecessary spaces.
251
252 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
253 and later.
254
255 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
256
257 9. Optimizing fast_forward_start_bits in JIT.
258
259 10. Adding support for callouts in JIT, and fixing some issues revealed
260 during this work. Namely:
261
262 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
263
264 (b) Minimum length was not checked before the matching is started.
265
266 11. The value of capture_last that is passed to callouts was incorrect in some
267 cases when there was a capture on one path that was subsequently abandoned
268 after a backtrack. Also, the capture_last value is now reset after a
269 recursion, since all captures are also reset in this case.
270
271 12. The interpreter no longer returns the "too many substrings" error in the
272 case when an overflowing capture is in a branch that is subsequently
273 abandoned after a backtrack.
274
275 13. In the pathological case when an offset vector of size 2 is used, pcretest
276 now prints out the matched string after a yield of 0 or 1.
277
278 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
279 Only supported by the JIT compiler at the moment.
280
281 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
282
283 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
284 offset of the starting point of the matching process, provided the offsets
285 vector is large enough.
286
287 17. The \A escape now records a lookbehind value of 1, though its execution
288 does not actually inspect the previous character. This is to ensure that,
289 in partial multi-segment matching, at least one character from the old
290 segment is retained when a new segment is processed. Otherwise, if there
291 are no lookbehinds in the pattern, \A might match incorrectly at the start
292 of a new segment.
293
294 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
295
296 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
297 gives some modest performance improvement in 8-bit mode.
298
299 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
300 be expressed in certain programming languages using Universal Character
301 Names.
302
303 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
304 which points out that "non characters" are not "characters that may not
305 appear in Unicode strings" but rather "characters that are reserved for
306 internal use and have only local meaning".
307
308 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
309 there was a conditional group that depended on an assertion, if the
310 assertion was false, the callout that immediately followed the alternation
311 in the condition was skipped when pcre_exec() was used for matching.
312
313 23. Allow an explicit callout to be inserted before an assertion that is the
314 condition for a conditional group, for compatibility with automatic
315 callouts, which always insert a callout at this point.
316
317 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
318 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
319
320 25. (*PRUNE) is now supported by the JIT compiler.
321
322 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
323
324 27. Fix the case where there are two or more SKIPs with arguments that may be
325 ignored.
326
327 28. (*SKIP) is now supported by the JIT compiler.
328
329 29. (*THEN) is now supported by the JIT compiler.
330
331 30. Update RunTest with additional test selector options.
332
333 31. The way PCRE handles backtracking verbs has been changed in two ways.
334
335 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
336 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
337 backtracking. In some cases this makes it more Perl-compatible, but Perl's
338 rather obscure rules do not always do the same thing.
339
340 (2) Previously, backtracking verbs were confined within assertions. This is
341 no longer the case for positive assertions, except for (*ACCEPT). Again,
342 this sometimes improves Perl compatibility, and sometimes does not.
343
344 32. A number of tests that were in test 2 because Perl did things differently
345 have been moved to test 1, because either Perl or PCRE has changed, and
346 these tests are now compatible.
347
348 32. Backtracking control verbs are now handled in the same way in JIT and
349 interpreter.
350
351 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
352 contained a forward subroutine reference caused a compile error.
353
354 34. Auto-detect and optimize limited repetitions in JIT.
355
356 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
357 blocking (*UTF) etc.
358
359 36. In the interpreter, maximizing pattern repetitions for characters and
360 character types now use tail recursion, which reduces stack usage.
361
362 37. The value of the max lookbehind was not correctly preserved if a compiled
363 and saved regex was reloaded on a host of different endianness.
364
365 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
366 of the compiled pattern block, expand the flags field from 16 to 32 bits
367 because it was almost full.
368
369 39. Try madvise first before posix_madvise.
370
371 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
372 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
373 original change locked out all matches of empty strings. This has been
374 changed so that one match of an empty string per line is recognized.
375 Subsequent searches on the same line (for colouring or for --only-matching,
376 for example) do not recognize empty strings.
377
378 41. Applied a user patch to fix a number of spelling mistakes in comments.
379
380 42. Data lines longer than 65536 caused pcretest to crash.
381
382 43. Clarified the data type for length and startoffset arguments for pcre_exec
383 and pcre_dfa_exec in the function-specific man pages, where they were
384 explicitly stated to be in bytes, never having been updated. I also added
385 some clarification to the pcreapi man page.
386
387 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
388 a segmentation fault.
389
390
391 Version 8.32 30-November-2012
392 -----------------------------
393
394 1. Improved JIT compiler optimizations for first character search and single
395 character iterators.
396
397 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
398 Patch by Daniel Richard G.
399
400 3. Single character iterator optimizations in the JIT compiler.
401
402 4. Improved JIT compiler optimizations for character ranges.
403
404 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
405 Reported by Giuseppe D'Angelo.
406
407 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
408 of a line, was being set incorrectly in cases where .* appeared inside
409 atomic brackets at the start of a pattern, or where there was a subsequent
410 *PRUNE or *SKIP.
411
412 7. Improved instruction cache flush for POWER/PowerPC.
413 Patch by Daniel Richard G.
414
415 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
416 grep:
417
418 (a) There is now no limit to the number of patterns to be matched.
419
420 (b) An error is given if a pattern is too long.
421
422 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
423 are now supported.
424
425 (d) --exclude-from and --include-from (multiple use) have been added.
426
427 (e) Exclusions and inclusions now apply to all files and directories, not
428 just to those obtained from scanning a directory recursively.
429
430 (f) Multiple uses of -f and --file-list are now supported.
431
432 (g) In a Windows environment, the default for -d has been changed from
433 "read" (the GNU grep default) to "skip", because otherwise the presence
434 of a directory in the file list provokes an error.
435
436 (h) The documentation has been revised and clarified in places.
437
438 9. Improve the matching speed of capturing brackets.
439
440 10. Changed the meaning of \X so that it now matches a Unicode extended
441 grapheme cluster.
442
443 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
444 out POSIX threads when JIT support is configured.
445
446 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
447
448 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
449 structure could go wrong in environments where size_t is not the same size
450 as int.
451
452 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
453 set.
454
455 15. The EBCDIC support had decayed; later updates to the code had included
456 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
457 general tidy up of EBCDIC-related issues, and the documentation was also
458 not quite right. There is now a test that can be run on ASCII systems to
459 check some of the EBCDIC-related things (but is it not a full test).
460
461 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
462 in a small tidy to the code.
463
464 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
465
466 18. If the --only-matching (-o) option in pcregrep is specified multiple
467 times, each one causes appropriate output. For example, -o1 -o2 outputs the
468 substrings matched by the 1st and 2nd capturing parentheses. A separating
469 string can be specified by --om-separator (default empty).
470
471 19. Improving the first n character searches.
472
473 20. Turn case lists for horizontal and vertical white space into macros so that
474 they are defined only once.
475
476 21. This set of changes together give more compatible Unicode case-folding
477 behaviour for characters that have more than one other case when UCP
478 support is available.
479
480 (a) The Unicode property table now has offsets into a new table of sets of
481 three or more characters that are case-equivalent. The MultiStage2.py
482 script that generates these tables (the pcre_ucd.c file) now scans
483 CaseFolding.txt instead of UnicodeData.txt for character case
484 information.
485
486 (b) The code for adding characters or ranges of characters to a character
487 class has been abstracted into a generalized function that also handles
488 case-independence. In UTF-mode with UCP support, this uses the new data
489 to handle characters with more than one other case.
490
491 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
492 whose other case is greater than 256 are now correctly matched
493 caselessly. Previously, the high codepoint matched the low one, but not
494 vice versa.
495
496 (d) The processing of \h, \H, \v, and \ in character classes now makes use
497 of the new class addition function, using character lists defined as
498 macros alongside the case definitions of 20 above.
499
500 (e) Caseless back references now work with characters that have more than
501 one other case.
502
503 (f) General caseless matching of characters with more than one other case
504 is supported.
505
506 22. Unicode character properties were updated from Unicode 6.2.0
507
508 23. Improved CMake support under Windows. Patch by Daniel Richard G.
509
510 24. Add support for 32-bit character strings, and UTF-32
511
512 25. Major JIT compiler update (code refactoring and bugfixing).
513 Experimental Sparc 32 support is added.
514
515 26. Applied a modified version of Daniel Richard G's patch to create
516 pcre.h.generic and config.h.generic by "make" instead of in the
517 PrepareRelease script.
518
519 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
520 pcre_compile.c when checking for a zero character.
521
522 28. Introducing a native interface for JIT. Through this interface, the compiled
523 machine code can be directly executed. The purpose of this interface is to
524 provide fast pattern matching, so several sanity checks are not performed.
525 However, feature tests are still performed. The new interface provides
526 1.4x speedup compared to the old one.
527
528 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
529 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
530 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
531
532 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
533 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
534 the "old" RFC 2279). Instead, it ended up passing a negative length to
535 pcre_exec().
536
537 31. Add support for GCC's visibility feature to hide internal functions.
538
539 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
540 "unknown -C option" after outputting 0 or 1.
541
542 33. There is now support for generating a code coverage report for the test
543 suite in environments where gcc is the compiler and lcov is installed. This
544 is mainly for the benefit of the developers.
545
546 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
547 unaddressable using valgrind annotations, allowing valgrind to detect
548 invalid memory accesses. This is mainly for the benefit of the developers.
549
550 25. (*UTF) can now be used to start a pattern in any of the three libraries.
551
552 26. Give configure error if --enable-cpp but no C++ compiler found.
553
554
555 Version 8.31 06-July-2012
556 -------------------------
557
558 1. Fixing a wrong JIT test case and some compiler warnings.
559
560 2. Removed a bashism from the RunTest script.
561
562 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
563 to unsigned type, result still unsigned" that was given by an MS compiler
564 on encountering the code "-sizeof(xxx)".
565
566 4. Partial matching support is added to the JIT compiler.
567
568 5. Fixed several bugs concerned with partial matching of items that consist
569 of more than one character:
570
571 (a) /^(..)\1/ did not partially match "aba" because checking references was
572 done on an "all or nothing" basis. This also applied to repeated
573 references.
574
575 (b) \R did not give a hard partial match if \r was found at the end of the
576 subject.
577
578 (c) \X did not give a hard partial match after matching one or more
579 characters at the end of the subject.
580
581 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
582 a partial match for the string "\r".
583
584 (e) When newline was set to CRLF, the metacharacter "." did not recognize
585 a partial match for a CR character at the end of the subject string.
586
587 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
588 running pcretest, the text "(JIT)" added to the output whenever JIT is
589 actually used to run the match.
590
591 7. Individual JIT compile options can be set in pcretest by following -s+[+]
592 or /S+[+] with a digit between 1 and 7.
593
594 8. OP_NOT now supports any UTF character not just single-byte ones.
595
596 9. (*MARK) control verb is now supported by the JIT compiler.
597
598 10. The command "./RunTest list" lists the available tests without actually
599 running any of them. (Because I keep forgetting what they all are.)
600
601 11. Add PCRE_INFO_MAXLOOKBEHIND.
602
603 12. Applied a (slightly modified) user-supplied patch that improves performance
604 when the heap is used for recursion (compiled with --disable-stack-for-
605 recursion). Instead of malloc and free for each heap frame each time a
606 logical recursion happens, frames are retained on a chain and re-used where
607 possible. This sometimes gives as much as 30% improvement.
608
609 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
610 call.
611
612 14. As documented, (*COMMIT) is now confined to within a positive assertion.
613
614 15. It is now possible to link pcretest with libedit as an alternative to
615 libreadline.
616
617 16. (*COMMIT) control verb is now supported by the JIT compiler.
618
619 17. The Unicode data tables have been updated to Unicode 6.1.0.
620
621 18. Added --file-list option to pcregrep.
622
623 19. Added binary file support to pcregrep, including the -a, --binary-files,
624 -I, and --text options.
625
626 20. The madvise function is renamed for posix_madvise for QNX compatibility
627 reasons. Fixed by Giuseppe D'Angelo.
628
629 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
630
631 22. Changed the option for creating symbolic links for 16-bit man pages from
632 -s to -sf so that re-installing does not cause issues.
633
634 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
635
636 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
637 matches in certain environments (the workspace was not being correctly
638 retained). Also added to pcre_dfa_exec() a simple plausibility check on
639 some of the workspace data at the beginning of a restart.
640
641 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
642 was not doing so when it should - probably a typo introduced by SVN 528
643 (change 8.10/14).
644
645 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
646 \w+ when the character tables indicated that \x{c4} was a word character.
647 There were several related cases, all because the tests for doing a table
648 lookup were testing for characters less than 127 instead of 255.
649
650 27. If a pattern contains capturing parentheses that are not used in a match,
651 their slots in the ovector are set to -1. For those that are higher than
652 any matched groups, this happens at the end of processing. In the case when
653 there were back references that the ovector was too small to contain
654 (causing temporary malloc'd memory to be used during matching), and the
655 highest capturing number was not used, memory off the end of the ovector
656 was incorrectly being set to -1. (It was using the size of the temporary
657 memory instead of the true size.)
658
659 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
660 ovector size, it uses memory at the end of the block that it has got.
661
662 29. Check for an overlong MARK name and give an error at compile time. The
663 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
664
665 30. JIT compiler update.
666
667 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
668 Rill for the patch.
669
670 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
671
672 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
673
674 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
675 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
676 was enabled.
677
678 35. Improve JIT code generation for greedy plus quantifier.
679
680 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
681 1 to "aa" instead of to an empty string. The bug affected repeated groups
682 that could potentially match an empty string.
683
684 37. Optimizing single character iterators in JIT.
685
686 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
687 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
688 codepoints that are too big for the mode are faulted, and in a UTF mode,
689 disallowed codepoints are also faulted.
690
691 39. If PCRE was compiled with UTF support, in three places in the DFA
692 matcher there was code that should only have been obeyed in UTF mode, but
693 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
694 processing when bytes with values greater than 127 were present. In 16-bit
695 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
696 both cases the values are those that cannot be the first data item in a UTF
697 character. The three items that might have provoked this were recursions,
698 possessively repeated groups, and atomic groups.
699
700 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
701 and pcregrep, because some OS require shared objects to be explicitly
702 passed to ld, causing the link step to fail if they are not.
703
704 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
705 mode, patterns that started with \h* or \R* might be incorrectly matched.
706
707
708 Version 8.30 04-February-2012
709 -----------------------------
710
711 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
712 name is defined in ctype.h.
713
714 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
715 only in quite long subpatterns.
716
717 3. Removed the function pcre_info(), which has been obsolete and deprecated
718 since it was replaced by pcre_fullinfo() in February 2000.
719
720 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
721 match a (*MARK), and the match failed at the start of the subject, a
722 reference to memory before the start of the subject could occur. This bug
723 was introduced by fix 17 of release 8.21.
724
725 5. A reference to an unset group with zero minimum repetition was giving
726 totally wrong answers (in non-JavaScript-compatibility mode). For example,
727 /(another)?(\1?)test/ matched against "hello world test". This bug was
728 introduced in release 8.13.
729
730 6. Add support for 16-bit character strings (a large amount of work involving
731 many changes and refactorings).
732
733 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
734 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
735 from a file.
736
737 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
738 rounding is not applied in this particular case).
739
740 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
741 if they appear, or are escaped, in patterns.
742
743 10. Get rid of a number of -Wunused-but-set-variable warnings.
744
745 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
746 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
747 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
748 also returns the mark "x". This bug applied to capturing parentheses,
749 non-capturing parentheses, and atomic parentheses. It also applied to some
750 assertions.
751
752 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
753 information out of configure.ac instead of relying on pcre.h.generic, which
754 is not stored in the repository.
755
756 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
757 -lreadline.
758
759 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
760
761 15. Applied Graycode's patch to put the top-level frame on the stack rather
762 than the heap when not using the stack for recursion. This gives a
763 performance improvement in many cases when recursion is not deep.
764
765 16. Experimental code added to "pcretest -C" to output the stack frame size.
766
767
768 Version 8.21 12-Dec-2011
769 ------------------------
770
771 1. Updating the JIT compiler.
772
773 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
774 are added as well.
775
776 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
777 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
778 calling _pcre_jit_exec. Some extra comments are added.
779
780 4. (*MARK) settings inside atomic groups that do not contain any capturing
781 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
782 was introduced by change 18 for 8.20.
783
784 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
785 ECMA-262 standard.
786
787 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
788 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
789 This bug was probably introduced by change 9 of 8.13.
790
791 7. While fixing 6 above, I noticed that a number of other items were being
792 incorrectly rejected as "not fixed length". This arose partly because newer
793 opcodes had not been added to the fixed-length checking code. I have (a)
794 corrected the bug and added tests for these items, and (b) arranged for an
795 error to occur if an unknown opcode is encountered while checking for fixed
796 length instead of just assuming "not fixed length". The items that were
797 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
798 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
799 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
800
801 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
802 being incorrectly compiled and would have given unpredicatble results.
803
804 9. A possessively repeated subpattern with minimum repeat count greater than
805 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
806 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
807 the first (A) could occur when it should not.
808
809 10. Add a cast and remove a redundant test from the code.
810
811 11. JIT should use pcre_malloc/pcre_free for allocation.
812
813 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
814 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
815 is anything other than /usr, -L is still shown).
816
817 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
818
819 14. Perl does not support \N without a following name in a [] class; PCRE now
820 also gives an error.
821
822 15. If a forward reference was repeated with an upper limit of around 2000,
823 it caused the error "internal error: overran compiling workspace". The
824 maximum number of forward references (including repeats) was limited by the
825 internal workspace, and dependent on the LINK_SIZE. The code has been
826 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
827 the default depends on LINK_SIZE. There is a new upper limit (for safety)
828 of around 200,000 forward references. While doing this, I also speeded up
829 the filling in of repeated forward references.
830
831 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
832 incorrectly expecting the subject to contain another "a" after the start.
833
834 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
835 in the match, the SKIP should be ignored. This was not happening; instead
836 the SKIP was being treated as NOMATCH. For patterns such as
837 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
838 tested.
839
840 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
841 now much more compatible with Perl, in particular in cases where the result
842 is a non-match for a non-anchored pattern. For example, if
843 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
844 "m", where previously it did not return a name. A side effect of this
845 change is that for partial matches, the last encountered mark name is
846 returned, as for non matches. A number of tests that were previously not
847 Perl-compatible have been moved into the Perl-compatible test files. The
848 refactoring has had the pleasing side effect of removing one argument from
849 the match() function, thus reducing its stack requirements.
850
851 19. If the /S+ option was used in pcretest to study a pattern using JIT,
852 subsequent uses of /S (without +) incorrectly behaved like /S+.
853
854 21. Retrieve executable code size support for the JIT compiler and fixing
855 some warnings.
856
857 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
858 not work when the shorter character appeared right at the end of the
859 subject string.
860
861 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
862 systems.
863
864 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
865 output it when the /M option is used in pcretest.
866
867 25. The CheckMan script was not being included in the distribution. Also, added
868 an explicit "perl" to run Perl scripts from the PrepareRelease script
869 because this is reportedly needed in Windows.
870
871 26. If study data was being save in a file and studying had not found a set of
872 "starts with" bytes for the pattern, the data written to the file (though
873 never used) was taken from uninitialized memory and so caused valgrind to
874 complain.
875
876 27. Updated RunTest.bat as provided by Sheri Pierce.
877
878 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
879
880 29. Computation of memory usage for the table of capturing group names was
881 giving an unnecessarily large value.
882
883
884 Version 8.20 21-Oct-2011
885 ------------------------
886
887 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
888 a POSIX class. After further experiments with Perl, which convinced me that
889 Perl has bugs and confusions, a closing square bracket is no longer allowed
890 in a POSIX name. This bug also affected patterns with classes that started
891 with full stops.
892
893 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
894 captured substring, but while checking the failing first alternative,
895 substring 1 is temporarily captured. If the output vector supplied to
896 pcre_exec() was not big enough for this capture, the yield of the function
897 was still zero ("insufficient space for captured substrings"). This cannot
898 be totally fixed without adding another stack variable, which seems a lot
899 of expense for a edge case. However, I have improved the situation in cases
900 such as /(a)(b)x|abc/ matched against "abc", where the return code
901 indicates that fewer than the maximum number of slots in the ovector have
902 been set.
903
904 3. Related to (2) above: when there are more back references in a pattern than
905 slots in the output vector, pcre_exec() uses temporary memory during
906 matching, and copies in the captures as far as possible afterwards. It was
907 using the entire output vector, but this conflicts with the specification
908 that only 2/3 is used for passing back captured substrings. Now it uses
909 only the first 2/3, for compatibility. This is, of course, another edge
910 case.
911
912 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
913 main code base, and can be used by building with --enable-jit. When this is
914 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
915 runtime --no-jit option is given.
916
917 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
918 ovector, the return from the function was zero, implying that there were
919 other matches that did not fit. The correct "exactly full" value is now
920 returned.
921
922 6. If a subpattern that was called recursively or as a subroutine contained
923 (*PRUNE) or any other control that caused it to give a non-standard return,
924 invalid errors such as "Error -26 (nested recursion at the same subject
925 position)" or even infinite loops could occur.
926
927 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
928 computing the minimum length on reaching *ACCEPT, and so ended up with the
929 wrong value of 1 rather than 0. Further investigation indicates that
930 computing a minimum subject length in the presence of *ACCEPT is difficult
931 (think back references, subroutine calls), and so I have changed the code
932 so that no minimum is registered for a pattern that contains *ACCEPT.
933
934 8. If (*THEN) was present in the first (true) branch of a conditional group,
935 it was not handled as intended. [But see 16 below.]
936
937 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
938 Sheri Pierce.
939
940 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
941 the first byte in a match must be "a".
942
943 11. Change 17 for 8.13 increased the recursion depth for patterns like
944 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
945 pattern contains any instances of (*THEN). If it does not, the old
946 optimizations are restored. It would be nice to do this on a per-group
947 basis, but at the moment that is not feasible.
948
949 12. In some environments, the output of pcretest -C is CRLF terminated. This
950 broke RunTest's code that checks for the link size. A single white space
951 character after the value is now allowed for.
952
953 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
954 For "fr", it uses the Windows-specific input and output files.
955
956 14. If (*THEN) appeared in a group that was called recursively or as a
957 subroutine, it did not work as intended. [But see next item.]
958
959 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
960 pattern fragments (but not containing any | characters). If A and B are
961 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
962 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
963 D. In other words, Perl considers parentheses that do not contain any |
964 characters to be part of a surrounding alternative, whereas PCRE was
965 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
966 differently. PCRE now behaves in the same way as Perl, except in the case
967 of subroutine/recursion calls such as (?1) which have in any case always
968 been different (but PCRE had them first :-).
969
970 16. Related to 15 above: Perl does not treat the | in a conditional group as
971 creating alternatives. Such a group is treated in the same way as an
972 ordinary group without any | characters when processing (*THEN). PCRE has
973 been changed to match Perl's behaviour.
974
975 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
976 RunGrepTest script failed.
977
978 18. Change 22 for version 13 caused atomic groups to use more stack. This is
979 inevitable for groups that contain captures, but it can lead to a lot of
980 stack use in large patterns. The old behaviour has been restored for atomic
981 groups that do not contain any capturing parentheses.
982
983 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
984 suppress the check for a minimum subject length at run time. (If it was
985 given to pcre_exec() or pcre_dfa_exec() it did work.)
986
987 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
988 fail to work when decoding hex characters in data strings in EBCDIC
989 environments.
990
991 21. It appears that in at least one Mac OS environment, the isxdigit() function
992 is implemented as a macro that evaluates to its argument more than once,
993 contravening the C 90 Standard (I haven't checked a later standard). There
994 was an instance in pcretest which caused it to go wrong when processing
995 \x{...} escapes in subject strings. The has been rewritten to avoid using
996 things like p++ in the argument of isxdigit().
997
998
999 Version 8.13 16-Aug-2011
1000 ------------------------
1001
1002 1. The Unicode data tables have been updated to Unicode 6.0.0.
1003
1004 2. Two minor typos in pcre_internal.h have been fixed.
1005
1006 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1007 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1008 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1009
1010 4. There were a number of related bugs in the code for matching backrefences
1011 caselessly in UTF-8 mode when codes for the characters concerned were
1012 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1013 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1014 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1015 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1016 2-byte code at the end of the subject (it thought there wasn't enough data
1017 left).
1018
1019 5. Comprehensive information about what went wrong is now returned by
1020 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1021 as the output vector has at least 2 elements. The offset of the start of
1022 the failing character and a reason code are placed in the vector.
1023
1024 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1025 now returned is for the first byte of the failing character, instead of the
1026 last byte inspected. This is an incompatible change, but I hope it is small
1027 enough not to be a problem. It makes the returned offset consistent with
1028 pcre_exec() and pcre_dfa_exec().
1029
1030 7. pcretest now gives a text phrase as well as the error number when
1031 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1032 failure, the offset and reason code are output.
1033
1034 8. When \R was used with a maximizing quantifier it failed to skip backwards
1035 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1036 back over a single character (\n). This seems wrong (because it treated the
1037 two characters as a single entity when going forwards), conflicts with the
1038 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1039 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1040 has been changed.
1041
1042 9. Some internal refactoring has changed the processing so that the handling
1043 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1044 time (the PCRE_DOTALL option was changed this way some time ago: version
1045 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1046 which was always a bit of a fudge. It also means that there is one less
1047 argument for the match() function, which reduces its stack requirements
1048 slightly. This change also fixes an incompatibility with Perl: the pattern
1049 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1050
1051 10. More internal refactoring has drastically reduced the number of recursive
1052 calls to match() for possessively repeated groups such as (abc)++ when
1053 using pcre_exec().
1054
1055 11. While implementing 10, a number of bugs in the handling of groups were
1056 discovered and fixed:
1057
1058 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1059 (a|)*(?1) gave a compile-time internal error.
1060 ((a|)+)+ did not notice that the outer group could match an empty string.
1061 (^a|^)+ was not marked as anchored.
1062 (.*a|.*)+ was not marked as matching at start or after a newline.
1063
1064 12. Yet more internal refactoring has removed another argument from the match()
1065 function. Special calls to this function are now indicated by setting a
1066 value in a variable in the "match data" data block.
1067
1068 13. Be more explicit in pcre_study() instead of relying on "default" for
1069 opcodes that mean there is no starting character; this means that when new
1070 ones are added and accidentally left out of pcre_study(), testing should
1071 pick them up.
1072
1073 14. The -s option of pcretest has been documented for ages as being an old
1074 synonym of -m (show memory usage). I have changed it to mean "force study
1075 for every regex", that is, assume /S for every regex. This is similar to -i
1076 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1077 using it. It makes it easier to run collections of tests with and without
1078 study enabled, and thereby test pcre_study() more easily. All the standard
1079 tests are now run with and without -s (but some patterns can be marked as
1080 "never study" - see 20 below).
1081
1082 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1083 restoration of the capturing data to the outer values was not happening
1084 correctly.
1085
1086 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1087 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1088 pattern had matched an empty string, and so incorrectly returned a no
1089 match.
1090
1091 17. There was optimizing code for the last branch of non-capturing parentheses,
1092 and also for the obeyed branch of a conditional subexpression, which used
1093 tail recursion to cut down on stack usage. Unfortunately, now that there is
1094 the possibility of (*THEN) occurring in these branches, tail recursion is
1095 no longer possible because the return has to be checked for (*THEN). These
1096 two optimizations have therefore been removed. [But see 8.20/11 above.]
1097
1098 18. If a pattern containing \R was studied, it was assumed that \R always
1099 matched two bytes, thus causing the minimum subject length to be
1100 incorrectly computed because \R can also match just one byte.
1101
1102 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1103 was incorrectly computed.
1104
1105 20. If /S is present twice on a test pattern in pcretest input, it now
1106 *disables* studying, thereby overriding the use of -s on the command line
1107 (see 14 above). This is necessary for one or two tests to keep the output
1108 identical in both cases.
1109
1110 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1111 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1112
1113 22. When an atomic group that contained a capturing parenthesis was
1114 successfully matched, but the branch in which it appeared failed, the
1115 capturing was not being forgotten if a higher numbered group was later
1116 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1117 group 1 to "a", when in fact it should be unset. This applied to multi-
1118 branched capturing and non-capturing groups, repeated or not, and also to
1119 positive assertions (capturing in negative assertions does not happen
1120 in PCRE) and also to nested atomic groups.
1121
1122 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1123 subject after a captured substring, to make it easier to tell which of a
1124 number of identical substrings has been captured.
1125
1126 24. The way atomic groups are processed by pcre_exec() has been changed so that
1127 if they are repeated, backtracking one repetition now resets captured
1128 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1129 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1130 "aaa". Previously, it would have been "a". As part of this code
1131 refactoring, the way recursive calls are handled has also been changed.
1132
1133 25. If an assertion condition captured any substrings, they were not passed
1134 back unless some other capturing happened later. For example, if
1135 (?(?=(a))a) was matched against "a", no capturing was returned.
1136
1137 26. When studying a pattern that contained subroutine calls or assertions,
1138 the code for finding the minimum length of a possible match was handling
1139 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1140 group 1 called group 2 while simultaneously a separate group 2 called group
1141 1). A stack overflow occurred in this case. I have fixed this by limiting
1142 the recursion depth to 10.
1143
1144 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1145 Fortmann. This supports explicit test numbers on the command line, and has
1146 argument validation and error reporting.
1147
1148 28. An instance of \X with an unlimited repeat could fail if at any point the
1149 first character it looked at was a mark character.
1150
1151 29. Some minor code refactoring concerning Unicode properties and scripts
1152 should reduce the stack requirement of match() slightly.
1153
1154 30. Added the '=' option to pcretest to check the setting of unused capturing
1155 slots at the end of the pattern, which are documented as being -1, but are
1156 not included in the return count.
1157
1158 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1159 compiled something random. Now it gives a compile-time error (as does
1160 Perl).
1161
1162 32. A *MARK encountered during the processing of a positive assertion is now
1163 recorded and passed back (compatible with Perl).
1164
1165 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1166 had alternative anchored branches, the search for a second match in a line
1167 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1168 matched the line "0102" twice. The same bug affected patterns that started
1169 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1170 twice.
1171
1172 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1173 does, and because of capturing effects, quantifying parenthesized
1174 assertions may at times be useful. Quantifiers are now allowed for
1175 parenthesized assertions.
1176
1177 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1178
1179 36. \g was being checked for fancy things in a character class, when it should
1180 just be a literal "g".
1181
1182 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1183 appearance of a nested POSIX class supersedes an apparent external class.
1184 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1185 unescaped square brackets may also appear as part of class names. For
1186 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1187 more like Perl. (But see 8.20/1 above.)
1188
1189 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1190 was because it thought it was \N{name}, which is not supported).
1191
1192 39. Add minix to OS list not supporting the -S option in pcretest.
1193
1194 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1195 cannot analyze patterns in sufficient detail to catch mutual recursions
1196 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1197 subgroup is called recursively as a subpattern for a second time at the
1198 same position in the subject string. In previous releases this might have
1199 been caught by the recursion limit, or it might have run out of stack.
1200
1201 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1202 happen only once. PCRE was, however incorrectly giving a compile time error
1203 "recursive call could loop indefinitely" because it cannot analyze the
1204 pattern in sufficient detail. The compile time test no longer happens when
1205 PCRE is compiling a conditional subpattern, but actual runaway loops are
1206 now caught at runtime (see 40 above).
1207
1208 42. It seems that Perl allows any characters other than a closing parenthesis
1209 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1210 has been changed to be the same.
1211
1212 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1213 as not to get warnings when autogen.sh is called. Also changed
1214 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1215
1216 44. To help people who use pcregrep to scan files containing exceedingly long
1217 lines, the following changes have been made:
1218
1219 (a) The default value of the buffer size parameter has been increased from
1220 8K to 20K. (The actual buffer used is three times this size.)
1221
1222 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1223 PCRE is built.
1224
1225 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1226 to be set at run time.
1227
1228 (d) Numerical values in pcregrep options can be followed by K or M, for
1229 example --buffer-size=50K.
1230
1231 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1232 given and the return code is set to 2.
1233
1234 45. Add a pointer to the latest mark to the callout data block.
1235
1236 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1237 partial match of an empty string instead of no match. This was specific to
1238 the use of ".".
1239
1240 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1241 complete match instead of a partial match. This bug was dependent on both
1242 the PCRE_UTF8 and PCRE_DOTALL options being set.
1243
1244 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1245 starting byte set, because \b was not being ignored.
1246
1247
1248 Version 8.12 15-Jan-2011
1249 ------------------------
1250
1251 1. Fixed some typos in the markup of the man pages, and wrote a script that
1252 checks for such things as part of the documentation building process.
1253
1254 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1255 --match-limit and --recursion-limit options (added for 8.11). In
1256 particular, this made one of the standard tests fail. (The integer value
1257 went into the wrong half of a long int.)
1258
1259 3. If the --colour option was given to pcregrep with -v (invert match), it
1260 did strange things, either producing crazy output, or crashing. It should,
1261 of course, ignore a request for colour when reporting lines that do not
1262 match.
1263
1264 4. Another pcregrep bug caused similar problems if --colour was specified with
1265 -M (multiline) and the pattern match finished with a line ending.
1266
1267 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1268 matched in multiline mode, the following line was shown as part of the
1269 match. This seems wrong, so I have changed it.
1270
1271 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1272 the check for further matches in the same line (so they could be coloured)
1273 to overrun the end of the current line. If another match was found, it was
1274 incorrectly shown (and then shown again when found in the next line).
1275
1276 7. If pcregrep was compiled under Windows, there was a reference to the
1277 function pcregrep_exit() before it was defined. I am assuming this was
1278 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1279 reported by a user. I've moved the definition above the reference.
1280
1281
1282 Version 8.11 10-Dec-2010
1283 ------------------------
1284
1285 1. (*THEN) was not working properly if there were untried alternatives prior
1286 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1287 backtracked to try for "b" instead of moving to the next alternative branch
1288 at the same level (in this case, to look for "c"). The Perl documentation
1289 is clear that when (*THEN) is backtracked onto, it goes to the "next
1290 alternative in the innermost enclosing group".
1291
1292 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1293 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1294 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1295 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1296 (*THEN).
1297
1298 3. If \s appeared in a character class, it removed the VT character from
1299 the class, even if it had been included by some previous item, for example
1300 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1301 of \s, but is part of the POSIX "space" class.)
1302
1303 4. A partial match never returns an empty string (because you can always
1304 match an empty string at the end of the subject); however the checking for
1305 an empty string was starting at the "start of match" point. This has been
1306 changed to the "earliest inspected character" point, because the returned
1307 data for a partial match starts at this character. This means that, for
1308 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1309 (previously it gave "no match").
1310
1311 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1312 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1313 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1314 has an implication that the given string is incomplete (because a partial
1315 match is preferred over a full match). For this reason, these items now
1316 give a partial match in this situation. [Aside: previously, the one case
1317 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1318 match rather than a full match, which was wrong by the old rules, but is
1319 now correct.]
1320
1321 6. There was a bug in the handling of #-introduced comments, recognized when
1322 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1323 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1324 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1325 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1326 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1327 places in pcre_compile().
1328
1329 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1330 comments when looking ahead for named forward references to subpatterns,
1331 the only newline sequence it recognized was NL. It now handles newlines
1332 according to the set newline convention.
1333
1334 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1335 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1336 cater for a lack of strerror(). These oversights have been fixed.
1337
1338 9. Added --match-limit and --recursion-limit to pcregrep.
1339
1340 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1341
1342 11. When the -o option was used, pcregrep was setting a return code of 1, even
1343 when matches were found, and --line-buffered was not being honoured.
1344
1345 12. Added an optional parentheses number to the -o and --only-matching options
1346 of pcregrep.
1347
1348 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1349 can match an empty string. The code to do it in pcretest and pcredemo
1350 needed fixing:
1351
1352 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1353 only one byte after an empty string match just before CRLF (this case
1354 just got forgotten; "any" and "anycrlf" were OK).
1355
1356 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1357 mode when an empty string match preceded an ASCII character followed by
1358 a non-ASCII character. (The code for advancing by one character rather
1359 than one byte was nonsense.)
1360
1361 (c) The pcredemo.c sample program did not have any code at all to handle
1362 the cases when CRLF is a valid newline sequence.
1363
1364 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1365 as a starting offset was within the subject string. There is now a new
1366 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1367 negative or greater than the length of the string. In order to test this,
1368 pcretest is extended to allow the setting of negative starting offsets.
1369
1370 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1371 starting offset points to the beginning of a UTF-8 character was
1372 unnecessarily clumsy. I tidied it up.
1373
1374 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1375 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1376
1377 17. Nobody had reported that the --include_dir option, which was added in
1378 release 7.7 should have been called --include-dir (hyphen, not underscore)
1379 for compatibility with GNU grep. I have changed it to --include-dir, but
1380 left --include_dir as an undocumented synonym, and the same for
1381 --exclude-dir, though that is not available in GNU grep, at least as of
1382 release 2.5.4.
1383
1384 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1385 characters from a string of bytes) have been redefined so as not to use
1386 loops, in order to improve performance in some environments. At the same
1387 time, I abstracted some of the common code into auxiliary macros to save
1388 repetition (this should not affect the compiled code).
1389
1390 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1391 compile-time error is now given if \c is not followed by an ASCII
1392 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1393 different, and any byte value is allowed.)
1394
1395 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1396 START_OPTIMIZE option, which is now allowed at compile time - but just
1397 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1398 to pcregrep and other applications that have no direct access to PCRE
1399 options. The new /Y option in pcretest sets this option when calling
1400 pcre_compile().
1401
1402 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1403 back references. Groups containing recursive back references were forced to
1404 be atomic by that change, but in the case of named groups, the amount of
1405 memory required was incorrectly computed, leading to "Failed: internal
1406 error: code overflow". This has been fixed.
1407
1408 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1409 pcretest.c, to avoid build problems in some Borland environments.
1410
1411
1412 Version 8.10 25-Jun-2010
1413 ------------------------
1414
1415 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1416 THEN.
1417
1418 2. (*ACCEPT) was not working when inside an atomic group.
1419
1420 3. Inside a character class, \B is treated as a literal by default, but
1421 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1422 causes the error). The code is unchanged, but I tidied the documentation.
1423
1424 4. Inside a character class, PCRE always treated \R and \X as literals,
1425 whereas Perl faults them if its -w option is set. I have changed PCRE so
1426 that it faults them when PCRE_EXTRA is set.
1427
1428 5. Added support for \N, which always matches any character other than
1429 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1430
1431 6. When compiling pcregrep with newer versions of gcc which may have
1432 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1433 declared with attribute warn_unused_result" were given. Just casting the
1434 result to (void) does not stop the warnings; a more elaborate fudge is
1435 needed. I've used a macro to implement this.
1436
1437 7. Minor change to pcretest.c to avoid a compiler warning.
1438
1439 8. Added four artifical Unicode properties to help with an option to make
1440 \s etc use properties (see next item). The new properties are: Xan
1441 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1442
1443 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1444 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1445 this option. Modified pcretest to add /W to test this facility. Added
1446 REG_UCP to make it available via the POSIX interface.
1447
1448 10. Added --line-buffered to pcregrep.
1449
1450 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1451 studied, and the match started with a letter with a code point greater than
1452 127 whose first byte was different to the first byte of the other case of
1453 the letter, the other case of this starting letter was not recognized
1454 (#976).
1455
1456 12. If a pattern that was studied started with a repeated Unicode property
1457 test, for example, \p{Nd}+, there was the theoretical possibility of
1458 setting up an incorrect bitmap of starting bytes, but fortunately it could
1459 not have actually happened in practice until change 8 above was made (it
1460 added property types that matched character-matching opcodes).
1461
1462 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1463 possible starting bytes for non-anchored patterns.
1464
1465 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1466 \R, and also a number of cases that involve Unicode properties, both
1467 explicit and implicit when PCRE_UCP is set.
1468
1469 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1470 input, it could crash or give wrong results if characters with values
1471 greater than 0xc0 were present in the subject string. (Detail: it assumed
1472 UTF-8 input when processing these items.)
1473
1474 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1475 size_t is 64-bit (#991).
1476
1477 17. Added a check for running out of memory when PCRE is compiled with
1478 --disable-stack-for-recursion (#990).
1479
1480 18. If the last data line in a file for pcretest does not have a newline on
1481 the end, a newline was missing in the output.
1482
1483 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1484 less than 128) in its various bitmaps. However, there is a facility for
1485 generating tables according to the current locale when PCRE is compiled. It
1486 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1487 characters, are recognized by isspace() and therefore were getting set in
1488 these tables, and indeed these tables seem to approximate to ISO 8859. This
1489 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1490 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1491 which of course cannot start UTF-8 characters. I have changed the code so
1492 that only real ASCII characters (less than 128) and the correct starting
1493 bytes for UTF-8 encodings are set for characters greater than 127 when in
1494 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1495 altogether.)
1496
1497 20. Added the /T option to pcretest so as to be able to run tests with non-
1498 standard character tables, thus making it possible to include the tests
1499 used for 19 above in the standard set of tests.
1500
1501 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1502 reference to a subpattern the other side of a comment that contains an
1503 opening parenthesis caused either an internal compiling error, or a
1504 reference to the wrong subpattern.
1505
1506
1507 Version 8.02 19-Mar-2010
1508 ------------------------
1509
1510 1. The Unicode data tables have been updated to Unicode 5.2.0.
1511
1512 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1513 configured.
1514
1515 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1516 original author of that file, following a query about its status.
1517
1518 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1519 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1520
1521 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1522 quantifier applied to a forward-referencing subroutine call, could compile
1523 incorrect code or give the error "internal error: previously-checked
1524 referenced subpattern not found".
1525
1526 6. Both MS Visual Studio and Symbian OS have problems with initializing
1527 variables to point to external functions. For these systems, therefore,
1528 pcre_malloc etc. are now initialized to local functions that call the
1529 relevant global functions.
1530
1531 7. There were two entries missing in the vectors called coptable and poptable
1532 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1533 I've fixed the data, and added a kludgy way of testing at compile time that
1534 the lengths are correct (equal to the number of opcodes).
1535
1536 8. Following on from 7, I added a similar kludge to check the length of the
1537 eint vector in pcreposix.c.
1538
1539 9. Error texts for pcre_compile() are held as one long string to avoid too
1540 much relocation at load time. To find a text, the string is searched,
1541 counting zeros. There was no check for running off the end of the string,
1542 which could happen if a new error number was added without updating the
1543 string.
1544
1545 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1546
1547 11. \K was not working if it appeared in an atomic group or in a group that
1548 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1549 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1550 the assertion is positive, but not if it is negative.
1551
1552 12. Change 11 fortuitously reduced the size of the stack frame used in the
1553 "match()" function of pcre_exec.c by one pointer. Forthcoming
1554 implementation of support for (*MARK) will need an extra pointer on the
1555 stack; I have reserved it now, so that the stack frame size does not
1556 decrease.
1557
1558 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1559 item in branch that calls a recursion is a subroutine call - as in the
1560 second branch in the above example - was incorrectly given the compile-
1561 time error "recursive call could loop indefinitely" because pcre_compile()
1562 was not correctly checking the subroutine for matching a non-empty string.
1563
1564 14. The checks for overrunning compiling workspace could trigger after an
1565 overrun had occurred. This is a "should never occur" error, but it can be
1566 triggered by pathological patterns such as hundreds of nested parentheses.
1567 The checks now trigger 100 bytes before the end of the workspace.
1568
1569 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1570
1571
1572 Version 8.01 19-Jan-2010
1573 ------------------------
1574
1575 1. If a pattern contained a conditional subpattern with only one branch (in
1576 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1577 computed the wrong minimum data length (which is of course zero for such
1578 subpatterns). This could cause incorrect "no match" results.
1579
1580 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1581 the pattern is reset in the first branch, pcre_compile() failed with
1582 "internal error: code overflow at offset...". This happened only when
1583 the reset was to the original external option setting. (An optimization
1584 abstracts leading options settings into an external setting, which was the
1585 cause of this.)
1586
1587 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1588 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1589 assertion pattern did not match (meaning that the assertion was true), it
1590 was incorrectly treated as false if the SKIP had been reached during the
1591 matching. This also applied to assertions used as conditions.
1592
1593 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1594 assertion subpattern, including such a pattern used as a condition,
1595 unpredictable results occurred, instead of the error return
1596 PCRE_ERROR_DFA_UITEM.
1597
1598 5. The C++ GlobalReplace function was not working like Perl for the special
1599 situation when an empty string is matched. It now does the fancy magic
1600 stuff that is necessary.
1601
1602 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1603 removed. (These were left over from very, very early versions of PCRE.)
1604
1605 7. Some cosmetic changes to the code to make life easier when compiling it
1606 as part of something else:
1607
1608 (a) Change DEBUG to PCRE_DEBUG.
1609
1610 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1611 called "current" as "current_branch", to prevent a collision with the
1612 Linux macro when compiled as a kernel module.
1613
1614 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1615 prevent a collision with the Linux macro when compiled as a kernel
1616 module.
1617
1618 8. In pcre_compile() there are some checks for integer overflows that used to
1619 cast potentially large values to (double). This has been changed to that
1620 when building, a check for int64_t is made, and if it is found, it is used
1621 instead, thus avoiding the use of floating point arithmetic. (There is no
1622 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1623 double.
1624
1625 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1626 2005 (difference between two addresses compared to an unsigned value).
1627
1628 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1629 custom one, because of the following reported problem in Windows:
1630
1631 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1632 under Win32.
1633 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1634 therefore missing the function definition.
1635 - The compiler thus generates a "C" signature for the test function.
1636 - The linker fails to find the "C" function.
1637 - PCRE fails to configure if asked to do so against libbz2.
1638
1639 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1640 messages were output:
1641
1642 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1643 rerunning libtoolize, to keep the correct libtool macros in-tree.
1644 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1645
1646 I have done both of these things.
1647
1648 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1649 most of the time, it *can* run out if it is given a pattern that contains a
1650 runaway infinite recursion. I updated the discussion in the pcrestack man
1651 page.
1652
1653 13. Now that we have gone to the x.xx style of version numbers, the minor
1654 version may start with zero. Using 08 or 09 is a bad idea because users
1655 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1656 interpreted as invalid octal numbers. I've updated the previous comment in
1657 configure.ac, and also added a check that gives an error if 08 or 09 are
1658 used.
1659
1660 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1661 causing partial matching to fail when the end of the subject matched \W
1662 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1663
1664 15. There were some discrepancies between the declarations in pcre_internal.h
1665 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1666 their definitions. The declarations used "const uschar *" and the
1667 definitions used USPTR. Even though USPTR is normally defined as "const
1668 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1669 reported that: "This difference in casting confuses some C++ compilers, for
1670 example, SunCC recognizes above declarations as different functions and
1671 generates broken code for hbpcre." I have changed the declarations to use
1672 USPTR.
1673
1674 16. GNU libtool is named differently on some systems. The autogen.sh script now
1675 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1676 (FreeBSD).
1677
1678 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1679 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1680 comment: "Figure out how to create a longlong from a string: strtoll and
1681 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1682 instance, but it only takes 2 args instead of 3!"
1683
1684 18. A subtle bug concerned with back references has been fixed by a change of
1685 specification, with a corresponding code fix. A pattern such as
1686 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1687 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1688 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1689 same bug. Such groups have to be quantified to be useful, or contained
1690 inside another quantified group. (If there's no repetition, the reference
1691 can never match.) The problem arises because, having left the group and
1692 moved on to the rest of the pattern, a later failure that backtracks into
1693 the group uses the captured value from the final iteration of the group
1694 rather than the correct earlier one. I have fixed this in PCRE by forcing
1695 any group that contains a reference to itself to be an atomic group; that
1696 is, there cannot be any backtracking into it once it has completed. This is
1697 similar to recursive and subroutine calls.
1698
1699
1700 Version 8.00 19-Oct-09
1701 ----------------------
1702
1703 1. The table for translating pcre_compile() error codes into POSIX error codes
1704 was out-of-date, and there was no check on the pcre_compile() error code
1705 being within the table. This could lead to an OK return being given in
1706 error.
1707
1708 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1709 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1710 in a Windows environment.
1711
1712 3. The pcregrep --count option prints the count for each file even when it is
1713 zero, as does GNU grep. However, pcregrep was also printing all files when
1714 --files-with-matches was added. Now, when both options are given, it prints
1715 counts only for those files that have at least one match. (GNU grep just
1716 prints the file name in this circumstance, but including the count seems
1717 more useful - otherwise, why use --count?) Also ensured that the
1718 combination -clh just lists non-zero counts, with no names.
1719
1720 4. The long form of the pcregrep -F option was incorrectly implemented as
1721 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1722 but it seems right to fix it, and I didn't think it was worth preserving
1723 the old behaviour.
1724
1725 5. The command line items --regex=pattern and --regexp=pattern were not
1726 recognized by pcregrep, which required --regex pattern or --regexp pattern
1727 (with a space rather than an '='). The man page documented the '=' forms,
1728 which are compatible with GNU grep; these now work.
1729
1730 6. No libpcreposix.pc file was created for pkg-config; there was just
1731 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1732
1733 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1734 when UCP support is not needed, by modifying the Python script that
1735 generates it from Unicode data files. This should not matter if the module
1736 is correctly used as a library, but I received one complaint about 50K of
1737 unwanted data. My guess is that the person linked everything into his
1738 program rather than using a library. Anyway, it does no harm.
1739
1740 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1741 was a minimum greater than 1 for a wide character in a possessive
1742 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1743 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1744 character. Chaos in the form of incorrect output or a compiling loop could
1745 result.
1746
1747 9. The restrictions on what a pattern can contain when partial matching is
1748 requested for pcre_exec() have been removed. All patterns can now be
1749 partially matched by this function. In addition, if there are at least two
1750 slots in the offset vector, the offset of the earliest inspected character
1751 for the match and the offset of the end of the subject are set in them when
1752 PCRE_ERROR_PARTIAL is returned.
1753
1754 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1755 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1756 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1757 and may be more useful for multi-segment matching.
1758
1759 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1760 used to be given if ever the end of the subject was reached; now it is
1761 given only if matching could not proceed because another character was
1762 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1763 string "Z", which now yields "no match" instead of "partial match". In the
1764 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1765 final character ended with (*FAIL).
1766
1767 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1768 if the pattern had a "must contain" character that was already found in the
1769 earlier partial match, unless partial matching was again requested. For
1770 example, with the pattern /dog.(body)?/, the "must contain" character is
1771 "g". If the first part-match was for the string "dog", restarting with
1772 "sbody" failed. This bug has been fixed.
1773
1774 13. The string returned by pcre_dfa_exec() after a partial match has been
1775 changed so that it starts at the first inspected character rather than the
1776 first character of the match. This makes a difference only if the pattern
1777 starts with a lookbehind assertion or \b or \B (\K is not supported by
1778 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1779 matching functions compatible, and I think it's the right thing to do.
1780
1781 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1782 so that the demonstration program is easily available in environments where
1783 PCRE has not been installed from source.
1784
1785 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1786 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1787 library.
1788
1789 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1790 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1791 is not the first non-POSIX option to be added. Clearly some people find
1792 these options useful.
1793
1794 17. If a caller to the POSIX matching function regexec() passes a non-zero
1795 value for nmatch with a NULL value for pmatch, the value of
1796 nmatch is forced to zero.
1797
1798 18. RunGrepTest did not have a test for the availability of the -u option of
1799 the diff command, as RunTest does. It now checks in the same way as
1800 RunTest, and also checks for the -b option.
1801
1802 19. If an odd number of negated classes containing just a single character
1803 interposed, within parentheses, between a forward reference to a named
1804 subpattern and the definition of the subpattern, compilation crashed with
1805 an internal error, complaining that it could not find the referenced
1806 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1807 [The bug was that it was starting one character too far in when skipping
1808 over the character class, thus treating the ] as data rather than
1809 terminating the class. This meant it could skip too much.]
1810
1811 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1812 /g option in pcretest when the pattern contains \K, which makes it possible
1813 to have an empty string match not at the start, even when the pattern is
1814 anchored. Updated pcretest and pcredemo to use this option.
1815
1816 21. If the maximum number of capturing subpatterns in a recursion was greater
1817 than the maximum at the outer level, the higher number was returned, but
1818 with unset values at the outer level. The correct (outer level) value is
1819 now given.
1820
1821 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1822 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1823 make it do so. The string so far is captured, making this feature
1824 compatible with Perl.
1825
1826 23. The tests have been re-organized, adding tests 11 and 12, to make it
1827 possible to check the Perl 5.10 features against Perl 5.10.
1828
1829 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1830 pattern matches a fixed length string. PCRE did not allow this; now it
1831 does. Neither allows recursion.
1832
1833 25. I finally figured out how to implement a request to provide the minimum
1834 length of subject string that was needed in order to match a given pattern.
1835 (It was back references and recursion that I had previously got hung up
1836 on.) This code has now been added to pcre_study(); it finds a lower bound
1837 to the length of subject needed. It is not necessarily the greatest lower
1838 bound, but using it to avoid searching strings that are too short does give
1839 some useful speed-ups. The value is available to calling programs via
1840 pcre_fullinfo().
1841
1842 26. While implementing 25, I discovered to my embarrassment that pcretest had
1843 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1844 study optimizations had never been tested with that matching function.
1845 Oops. What is worse, even when it was passed study data, there was a bug in
1846 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1847 were also very few tests of studied patterns with pcre_dfa_exec().
1848
1849 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1850 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1851 on the other side of the coin, they are no longer allowed to have different
1852 names, because these cannot be distinguished in PCRE, and this has caused
1853 confusion. (This is a difference from Perl.)
1854
1855 28. When duplicate subpattern names are present (necessarily with different
1856 numbers, as required by 27 above), and a test is made by name in a
1857 conditional pattern, either for a subpattern having been matched, or for
1858 recursion in such a pattern, all the associated numbered subpatterns are
1859 tested, and the overall condition is true if the condition is true for any
1860 one of them. This is the way Perl works, and is also more like the way
1861 testing by number works.
1862
1863
1864 Version 7.9 11-Apr-09
1865 ---------------------
1866
1867 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1868 (pcretest), all targets were linked against these libraries. This included
1869 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1870 libraries. This caused unwanted dependencies to be created. This problem
1871 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1872 pcretest is linked with readline.
1873
1874 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1875 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1876 moved outside it again, because FALSE and TRUE are already defined in AIX,
1877 but BOOL is not.
1878
1879 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1880 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1881
1882 4. The pcregrep documentation said spaces were inserted as well as colons (or
1883 hyphens) following file names and line numbers when outputting matching
1884 lines. This is not true; no spaces are inserted. I have also clarified the
1885 wording for the --colour (or --color) option.
1886
1887 5. In pcregrep, when --colour was used with -o, the list of matching strings
1888 was not coloured; this is different to GNU grep, so I have changed it to be
1889 the same.
1890
1891 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1892 each matching line was coloured. Now it goes on to look for further matches
1893 of any of the test patterns, which is the same behaviour as GNU grep.
1894
1895 7. A pattern that could match an empty string could cause pcregrep to loop; it
1896 doesn't make sense to accept an empty string match in pcregrep, so I have
1897 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1898 seems to be how GNU grep behaves. [But see later change 40 for release
1899 8.33.]
1900
1901 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1902 start or after a newline", because the conditional assertion was not being
1903 correctly handled. The rule now is that both the assertion and what follows
1904 in the first alternative must satisfy the test.
1905
1906 9. If auto-callout was enabled in a pattern with a conditional group whose
1907 condition was an assertion, PCRE could crash during matching, both with
1908 pcre_exec() and pcre_dfa_exec().
1909
1910 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1911 used for matching.
1912
1913 11. Unicode property support in character classes was not working for
1914 characters (bytes) greater than 127 when not in UTF-8 mode.
1915
1916 12. Added the -M command line option to pcretest.
1917
1918 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1919
1920 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1921
1922 16. Added comments and documentation about mis-use of no_arg in the C++
1923 wrapper.
1924
1925 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1926 from Martin Jerabek that uses macro names for all relevant character and
1927 string constants.
1928
1929 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1930 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1931 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1932 these, but not everybody uses configure.
1933
1934 19. A conditional group that had only one branch was not being correctly
1935 recognized as an item that could match an empty string. This meant that an
1936 enclosing group might also not be so recognized, causing infinite looping
1937 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1938 with the subject "ab", where knowledge that the repeated group can match
1939 nothing is needed in order to break the loop.
1940
1941 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1942 exec(), but without supplying a callout function, matching went wrong.
1943
1944 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1945 leak if the size of the offset vector was greater than 30. When the vector
1946 is smaller, the saved offsets during recursion go onto a local stack
1947 vector, but for larger vectors malloc() is used. It was failing to free
1948 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1949 error, in fact).
1950
1951 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1952 heapframe that is used only when UTF-8 support is enabled. This caused no
1953 problem, but was untidy.
1954
1955 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1956 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1957 included within another project.
1958
1959 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1960 slightly modified by me:
1961
1962 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1963 not building pcregrep.
1964
1965 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1966 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1967
1968 25. Forward references, both numeric and by name, in patterns that made use of
1969 duplicate group numbers, could behave incorrectly or give incorrect errors,
1970 because when scanning forward to find the reference group, PCRE was not
1971 taking into account the duplicate group numbers. A pattern such as
1972 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1973
1974 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1975 the feature of a custom pointer more persuasive (as requested by a user).
1976
1977 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1978 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1979 pre-defined.
1980
1981 28. Added support for (*UTF8) at the start of a pattern.
1982
1983 29. Arrange for flags added by the "release type" setting in CMake to be shown
1984 in the configuration summary.
1985
1986
1987 Version 7.8 05-Sep-08
1988 ---------------------
1989
1990 1. Replaced UCP searching code with optimized version as implemented for Ad
1991 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1992 stage table and inline lookup instead of a function, giving speed ups of 2
1993 to 5 times on some simple patterns that I tested. Permission was given to
1994 distribute the MultiStage2.py script that generates the tables (it's not in
1995 the tarball, but is in the Subversion repository).
1996
1997 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1998 scripts.
1999
2000 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2001 a group with a zero qualifier. The result of the study could be incorrect,
2002 or the function might crash, depending on the pattern.
2003
2004 4. Caseless matching was not working for non-ASCII characters in back
2005 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2006 It now works when Unicode Property Support is available.
2007
2008 5. In pcretest, an escape such as \x{de} in the data was always generating
2009 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2010 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2011 truncation.
2012
2013 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2014
2015 7. Added two (int) casts to pcregrep when printing the difference of two
2016 pointers, in case they are 64-bit values.
2017
2018 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2019 test 2 if it fails.
2020
2021 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2022 and a #define of that name to empty if it is not externally set. This is to
2023 allow users of MSVC to set it if necessary.
2024
2025 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2026 the convenience functions in the pcre_get.c source file.
2027
2028 11. An option change at the start of a pattern that had top-level alternatives
2029 could cause overwriting and/or a crash. This command provoked a crash in
2030 some environments:
2031
2032 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2033
2034 This potential security problem was recorded as CVE-2008-2371.
2035
2036 12. For a pattern where the match had to start at the beginning or immediately
2037 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2038 pcre_dfa_exec() could read past the end of the passed subject if there was
2039 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2040 pcretest so that it places the subject at the end of its malloc-ed buffer.
2041
2042 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2043 exec() might read past the end of the data buffer in UTF-8 mode.
2044
2045 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2046 the data contained the byte 0x85 as part of a UTF-8 character within its
2047 first line. This applied both to normal and DFA matching.
2048
2049 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2050 /^[^d]*?$/8 failed to match "abc".
2051
2052 16. Added a missing copyright notice to pcrecpp_internal.h.
2053
2054 17. Make it more clear in the documentation that values returned from
2055 pcre_exec() in ovector are byte offsets, not character counts.
2056
2057 18. Tidied a few places to stop certain compilers from issuing warnings.
2058
2059 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2060 supplied by Stefan Weber. I made a further small update for 7.8 because
2061 there is a change of source arrangements: the pcre_searchfuncs.c module is
2062 replaced by pcre_ucd.c.
2063
2064
2065 Version 7.7 07-May-08
2066 ---------------------
2067
2068 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2069 a string to a long long, pretend we don't even have a long long." This is
2070 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2071
2072 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2073 pre-7.6 versions, which defined a global no_arg variable instead of putting
2074 it in the RE class. (See also #8 below.)
2075
2076 3. Remove a line of dead code, identified by coverity and reported by Nuno
2077 Lopes.
2078
2079 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2080
2081 (1) The include/exclude patterns were being applied to the whole pathnames
2082 of files, instead of just to the final components.
2083
2084 (2) If there was more than one level of directory, the subdirectories were
2085 skipped unless they satisfied the include/exclude conditions. This is
2086 inconsistent with GNU grep (and could even be seen as contrary to the
2087 pcregrep specification - which I improved to make it absolutely clear).
2088 The action now is always to scan all levels of directory, and just
2089 apply the include/exclude patterns to regular files.
2090
2091 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2092 --exclude_dir in the tests to avoid scanning .svn directories.
2093
2094 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2095 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2096 doesn't support NULs in patterns.
2097
2098 7. Added some missing "const"s to declarations of static tables in
2099 pcre_compile.c and pcre_dfa_exec.c.
2100
2101 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2102 caused by fix #2 above. (Subsequently also a second patch to fix the
2103 first patch. And a third patch - this was a messy problem.)
2104
2105 9. Applied Craig's patch to remove the use of push_back().
2106
2107 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2108 matching function regexec().
2109
2110 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2111 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2112 references. PCRE supports relative numbers with this syntax (I don't think
2113 Oniguruma does).
2114
2115 12. Previously, a group with a zero repeat such as (...){0} was completely
2116 omitted from the compiled regex. However, this means that if the group
2117 was called as a subroutine from elsewhere in the pattern, things went wrong
2118 (an internal error was given). Such groups are now left in the compiled
2119 pattern, with a new opcode that causes them to be skipped at execution
2120 time.
2121
2122 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2123 to the way PCRE behaves:
2124
2125 (a) A lone ] character is dis-allowed (Perl treats it as data).
2126
2127 (b) A back reference to an unmatched subpattern matches an empty string
2128 (Perl fails the current match path).
2129
2130 (c) A data ] in a character class must be notated as \] because if the
2131 first data character in a class is ], it defines an empty class. (In
2132 Perl it is not possible to have an empty class.) The empty class []
2133 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2134 The negative empty class [^] matches any one character, independently
2135 of the DOTALL setting.
2136
2137 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2138 non-existent subpattern following a character class starting with ']' and
2139 containing () gave an internal compiling error instead of "reference to
2140 non-existent subpattern". Fortunately, when the pattern did exist, the
2141 compiled code was correct. (When scanning forwards to check for the
2142 existence of the subpattern, it was treating the data ']' as terminating
2143 the class, so got the count wrong. When actually compiling, the reference
2144 was subsequently set up correctly.)
2145
2146 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2147 it was being rejected as not supported by pcre_dfa_exec(), even though
2148 other assertions are supported. I have made pcre_dfa_exec() support
2149 (*FAIL).
2150
2151 16. The implementation of 13c above involved the invention of a new opcode,
2152 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2153 cannot be changed at match time, I realized I could make a small
2154 improvement to matching performance by compiling OP_ALLANY instead of
2155 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2156 on the OP_ANY path.
2157
2158 17. Compiling pcretest on Windows with readline support failed without the
2159 following two fixes: (1) Make the unistd.h include conditional on
2160 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2161
2162 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2163 ncurses library to be included for pcretest when ReadLine support is
2164 requested, but also to allow for it to be overridden. This patch came from
2165 Daniel Bergström.
2166
2167 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2168 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2169 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2170 spotting this.
2171
2172
2173 Version 7.6 28-Jan-08
2174 ---------------------
2175
2176 1. A character class containing a very large number of characters with
2177 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2178 overflow.
2179
2180 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2181 HAVE_LONG_LONG is not defined.
2182
2183 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2184 bring them up to date and include new features. This patch includes:
2185
2186 - Fixed PH's badly added libz and libbz2 support.
2187 - Fixed a problem with static linking.
2188 - Added pcredemo. [But later removed - see 7 below.]
2189 - Fixed dftables problem and added an option.
2190 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2191 HAVE_LONG_LONG.
2192 - Added readline support for pcretest.
2193 - Added an listing of the option settings after cmake has run.
2194
2195 4. A user submitted a patch to Makefile that makes it easy to create
2196 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2197 Makefile.am that cause it to include this special target, without
2198 affecting anything else. Note that the same mingw target plus all
2199 the other distribution libraries and programs are now supported
2200 when configuring with CMake (see 6 below) instead of with
2201 Configure/Make.
2202
2203 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2204 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2205 exported in the Windows port". It has not yet been confirmed that the patch
2206 solves the problem, but it does no harm.
2207
2208 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2209 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2210 with CMake, and also correct the comment about stack recursion.
2211
2212 7. Remove the automatic building of pcredemo from the ./configure system and
2213 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2214 of a program that users should build themselves after PCRE is installed, so
2215 building it automatically is not really right. What is more, it gave
2216 trouble in some build environments.
2217
2218 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2219
2220
2221 Version 7.5 10-Jan-08
2222 ---------------------
2223
2224 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2225 values in parens when parsing an RE using the C++ wrapper."
2226
2227 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2228 Characters greater than 255 were excluded from the class instead of being
2229 included.
2230
2231 3. The same bug as (2) above applied to negated POSIX classes such as
2232 [:^space:].
2233
2234 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2235 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2236 I have changed it.
2237
2238 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2239 first named subpattern) and a construct such as (?&a) would reference the
2240 first named subpattern whose name started with "a" (in other words, the
2241 length check was missing). Both these problems are fixed. "Subpattern name
2242 expected" is now given for (?&) (a zero-length name), and this patch also
2243 makes it give the same error for \k'' (previously it complained that that
2244 was a reference to a non-existent subpattern).
2245
2246 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2247 this is right because (?- can be followed by option settings as well as by
2248 digits. I have, however, made the messages clearer.
2249
2250 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2251 than the number used in the conditional) now cause a compile-time error.
2252 This is actually not compatible with Perl, which accepts such patterns, but
2253 treats the conditional as always being FALSE (as PCRE used to), but it
2254 seems to me that giving a diagnostic is better.
2255
2256 8. Change "alphameric" to the more common word "alphanumeric" in comments
2257 and messages.
2258
2259 9. Fix two occurrences of "backslash" in comments that should have been
2260 "backspace".
2261
2262 10. Remove two redundant lines of code that can never be obeyed (their function
2263 was moved elsewhere).
2264
2265 11. The program that makes PCRE's Unicode character property table had a bug
2266 which caused it to generate incorrect table entries for sequences of
2267 characters that have the same character type, but are in different scripts.
2268 It amalgamated them into a single range, with the script of the first of
2269 them. In other words, some characters were in the wrong script. There were
2270 thirteen such cases, affecting characters in the following ranges:
2271
2272 U+002b0 - U+002c1
2273 U+0060c - U+0060d
2274 U+0061e - U+00612
2275 U+0064b - U+0065e
2276 U+0074d - U+0076d
2277 U+01800 - U+01805
2278 U+01d00 - U+01d77
2279 U+01d9b - U+01dbf
2280 U+0200b - U+0200f
2281 U+030fc - U+030fe
2282 U+03260 - U+0327f
2283 U+0fb46 - U+0fbb1
2284 U+10450 - U+1049d
2285
2286 12. The -o option (show only the matching part of a line) for pcregrep was not
2287 compatible with GNU grep in that, if there was more than one match in a
2288 line, it showed only the first of them. It now behaves in the same way as
2289 GNU grep.
2290
2291 13. If the -o and -v options were combined for pcregrep, it printed a blank
2292 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2293 does the same. The return code can be used to tell if there were any
2294 non-matching lines.
2295
2296 14. Added --file-offsets and --line-offsets to pcregrep.
2297
2298 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2299 infinitely looping recursion. The bug was that positive lookaheads were not
2300 being skipped when checking for a possible empty match (negative lookaheads
2301 and both kinds of lookbehind were skipped).
2302
2303 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2304 inclusion of <windows.h> to before rather than after the definition of
2305 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2306
2307 17. Specifying a possessive quantifier with a specific limit for a Unicode
2308 character property caused pcre_compile() to compile bad code, which led at
2309 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2310 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2311 caused the error; without that there was no problem.
2312
2313 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2314
2315 19. Added --enable-pcretest-libreadline.
2316
2317 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2318 RE::GlobalReplace(). As a result, the number of replacements returned was
2319 double what it should be. I removed one of the increments, but Craig sent a
2320 later patch that removed the other one (the right fix) and added unit tests
2321 that check the return values (which was not done before).
2322
2323 21. Several CMake things:
2324
2325 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2326 the names libpcre and libpcreposix, not just pcre and pcreposix.
2327
2328 (2) The above change means that pcretest and pcregrep are now correctly
2329 linked with the newly-built libraries, not previously installed ones.
2330
2331 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2332
2333 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2334 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2335 UTF-8 newline character). The key issue is that the pattern starts .*;
2336 this means that the match must be either at the beginning, or after a
2337 newline. The bug was in the code for advancing after a failed match and
2338 checking that the new position followed a newline. It was not taking
2339 account of UTF-8 characters correctly.
2340
2341 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2342 character classes. PCRE was not treating the sequence [:...:] as a
2343 character class unless the ... were all letters. Perl, however, seems to
2344 allow any characters between [: and :], though of course it rejects as
2345 unknown any "names" that contain non-letters, because all the known class
2346 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2347 for example, whereas PCRE did not - it did not recognize a POSIX character
2348 class. This seemed a bit dangerous, so the code has been changed to be
2349 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2350 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2351 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2352 Perl does, and where it didn't before.
2353
2354 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2355 Windows environments %n is disabled by default.
2356
2357
2358 Version 7.4 21-Sep-07
2359 ---------------------
2360
2361 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2362 means that a class such as [\s] counted as "explicit reference to CR or
2363 LF". That isn't really right - the whole point of the change was to try to
2364 help when there was an actual mention of one of the two characters. So now
2365 the change happens only if \r or \n (or a literal CR or LF) character is
2366 encountered.
2367
2368 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2369 of both had grown to the point where there were only 3 bits left.
2370 Fortunately, there was spare space in the data structure, and so I have
2371 moved the internal flags into a new 16-bit field to free up more option
2372 bits.
2373
2374 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2375 but did not set the internal JCHANGED flag - either of these is enough to
2376 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2377 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2378 start sets both bits.
2379
2380 4. Added options (at build time, compile time, exec time) to change \R from
2381 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2382
2383 5. doc/pcresyntax.html was missing from the distribution.
2384
2385 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2386 compatibility, even though it is no longer used.
2387
2388 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2389 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2390 windows.h file is present (where different names are used). [This was
2391 reversed later after testing - see 16 below.]
2392
2393 8. Changed all #include <config.h> to #include "config.h". There were also
2394 some further <pcre.h> cases that I changed to "pcre.h".
2395
2396 9. When pcregrep was used with the --colour option, it missed the line ending
2397 sequence off the lines that it output.
2398
2399 10. It was pointed out to me that arrays of string pointers cause lots of
2400 relocations when a shared library is dynamically loaded. A technique of
2401 using a single long string with a table of offsets can drastically reduce
2402 these. I have refactored PCRE in four places to do this. The result is
2403 dramatic:
2404
2405 Originally: 290
2406 After changing UCP table: 187
2407 After changing error message table: 43
2408 After changing table of "verbs" 36
2409 After changing table of Posix names 22
2410
2411 Thanks to the folks working on Gregex for glib for this insight.
2412
2413 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2414 unicode-properties was also set.
2415
2416 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2417
2418 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2419 checked only for CRLF.
2420
2421 14. Added casts to pcretest.c to avoid compiler warnings.
2422
2423 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2424
2425 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2426 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2427 entirely. This removes changes made in 7 above.
2428
2429 17. The CMake files have been updated, and there is now more information about
2430 building with CMake in the NON-UNIX-USE document.
2431
2432
2433 Version 7.3 28-Aug-07
2434 ---------------------
2435
2436 1. In the rejigging of the build system that eventually resulted in 7.1, the
2437 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2438 brackets there is not right, since it causes compilers to look for an
2439 installed pcre.h, not the version that is in the source that is being
2440 compiled (which of course may be different). I have changed it back to:
2441
2442 #include "pcre.h"
2443
2444 I have a vague recollection that the change was concerned with compiling in
2445 different directories, but in the new build system, that is taken care of
2446 by the VPATH setting the Makefile.
2447
2448 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2449 when the subject happened to end in the byte 0x85 (e.g. if the last
2450 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2451 characters but of course it shouldn't be taken as a newline when it is part
2452 of another character. The bug was that, for an unlimited repeat of . in
2453 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2454 characters when looking for a newline.
2455
2456 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2457
2458 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2459 in debug output.
2460
2461 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2462 long printing in the pcrecpp unittest when running under MinGW.
2463
2464 6. ESC_K was left out of the EBCDIC table.
2465
2466 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2467 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2468 limit also applies to "virtual nesting" when a pattern is recursive, and in
2469 this case 1000 isn't so big. I have been able to remove this limit at the
2470 expense of backing off one optimization in certain circumstances. Normally,
2471 when pcre_exec() would call its internal match() function recursively and
2472 immediately return the result unconditionally, it uses a "tail recursion"
2473 feature to save stack. However, when a subpattern that can match an empty
2474 string has an unlimited repetition quantifier, it no longer makes this
2475 optimization. That gives it a stack frame in which to save the data for
2476 checking that an empty string has been matched. Previously this was taken
2477 from the 1000-entry workspace that had been reserved. So now there is no
2478 explicit limit, but more stack is used.
2479
2480 8. Applied Daniel's patches to solve problems with the import/export magic
2481 syntax that is required for Windows, and which was going wrong for the
2482 pcreposix and pcrecpp parts of the library. These were overlooked when this
2483 problem was solved for the main library.
2484
2485 9. There were some crude static tests to avoid integer overflow when computing
2486 the size of patterns that contain repeated groups with explicit upper
2487 limits. As the maximum quantifier is 65535, the maximum group length was
2488 set at 30,000 so that the product of these two numbers did not overflow a
2489 32-bit integer. However, it turns out that people want to use groups that
2490 are longer than 30,000 bytes (though not repeat them that many times).
2491 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2492 made it possible to implement the integer overflow checks in a much more
2493 dynamic way, which I have now done. The artificial limitation on group
2494 length has been removed - we now have only the limit on the total length of
2495 the compiled pattern, which depends on the LINK_SIZE setting.
2496
2497 10. Fixed a bug in the documentation for get/copy named substring when
2498 duplicate names are permitted. If none of the named substrings are set, the
2499 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2500 empty string.
2501
2502 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2503 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2504 because the ] is interpreted as the first data character and the
2505 terminating ] is not found. PCRE has been made compatible with Perl in this
2506 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2507 cause memory overwriting.
2508
2509 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2510 string has been matched (to stop an infinite loop). It was not recognizing
2511 a conditional subpattern that could match an empty string if that
2512 subpattern was within another subpattern. For example, it looped when
2513 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2514 condition was not nested. This bug has been fixed.
2515
2516 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2517 past the start of the subject in the presence of bytes with the top bit
2518 set, for example "\x8aBCD".
2519
2520 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2521 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2522
2523 14. Optimized (?!) to (*FAIL).
2524
2525 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2526 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2527 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2528 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2529 does: it's just the validity check that is more restrictive.
2530
2531 16. Inserted checks for integer overflows during escape sequence (backslash)
2532 processing, and also fixed erroneous offset values for syntax errors during
2533 backslash processing.
2534
2535 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2536 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2537
2538 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2539 caused an overrun.
2540
2541 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2542 something other than just ASCII characters) inside a group that had an
2543 unlimited repeat caused a loop at compile time (while checking to see
2544 whether the group could match an empty string).
2545
2546 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2547 [\P{Any}] did so. (Error in the code for printing property names.)
2548
2549 21. An orphan \E inside a character class could cause a crash.
2550
2551 22. A repeated capturing bracket such as (A)? could cause a wild memory
2552 reference during compilation.
2553
2554 23. There are several functions in pcre_compile() that scan along a compiled
2555 expression for various reasons (e.g. to see if it's fixed length for look
2556 behind). There were bugs in these functions when a repeated \p or \P was
2557 present in the pattern. These operators have additional parameters compared
2558 with \d, etc, and these were not being taken into account when moving along
2559 the compiled data. Specifically:
2560
2561 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2562 length.
2563
2564 (b) An item such as \pL+ within a repeated group could cause crashes or
2565 loops.
2566
2567 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2568 "reference to non-existent subpattern" error.
2569
2570 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2571
2572 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2573 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2574
2575 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2576
2577 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2578 character were causing crashes (broken optimization).
2579
2580 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2581 \p or \P) caused a compile-time loop.
2582
2583 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2584 break. For example, the unstudied pattern [\r\n]A does not match the string
2585 "\r\nA" because change 7.0/46 below moves the current point on by two
2586 characters after failing to match at the start. However, the pattern \nA
2587 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2588 the same is true. There doesn't seem any very clean way out of this, but
2589 what I have chosen to do makes the common cases work: PCRE now takes note
2590 of whether there can be an explicit match for \r or \n anywhere in the
2591 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2592 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2593 pattern has explicit CR or LF references.
2594
2595 29. Added (*CR) etc for changing newline setting at start of pattern.
2596
2597
2598 Version 7.2 19-Jun-07
2599 ---------------------
2600
2601 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2602 which is apparently normally available under Windows.
2603
2604 2. Re-jig the pcregrep tests with different newline settings in an attempt
2605 to make them independent of the local environment's newline setting.
2606
2607 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2608
2609 4. Some of the "internals" tests were previously cut out when the link size
2610 was not 2, because the output contained actual offsets. The recent new
2611 "Z" feature of pcretest means that these can be cut out, making the tests
2612 usable with all link sizes.
2613
2614 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2615 stack recursion. This gives a massive performance boost under BSD, but just
2616 a small improvement under Linux. However, it saves one field in the frame
2617 in all cases.
2618
2619 6. Added more features from the forthcoming Perl 5.10:
2620
2621 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2622 recursion call. It refers to the nth most recently opened parentheses.
2623
2624 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2625 to be opened parentheses.
2626
2627 (c) Conditions that refer to capturing parentheses can be specified
2628 relatively, for example, (?(-2)... or (?(+3)...
2629
2630 (d) \K resets the start of the current match so that everything before
2631 is not part of it.
2632
2633 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2634
2635 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2636 reference syntax.
2637
2638 (g) (?| introduces a group in which the numbering of parentheses in each
2639 alternative starts with the same number.
2640
2641 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2642
2643 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2644 PCRE_INFO_JCHANGED.
2645
2646 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2647 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2648 for detecting groups that can match an empty string.
2649
2650 9. A pattern with a very large number of alternatives (more than several
2651 hundred) was running out of internal workspace during the pre-compile
2652 phase, where pcre_compile() figures out how much memory will be needed. A
2653 bit of new cunning has reduced the workspace needed for groups with
2654 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2655 workspace instead of running out of the 4096 that are available.
2656
2657 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2658
2659 11. Applied patch from Google to remove an optimization that didn't quite work.
2660 The report of the bug said:
2661
2662 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2663 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2664 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2665
2666 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2667 it matched the wrong number of bytes.
2668
2669
2670 Version 7.1 24-Apr-07
2671 ---------------------
2672
2673 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2674 that is more "standard", making use of automake and other Autotools. There
2675 is some re-arrangement of the files and adjustment of comments consequent
2676 on this.
2677
2678 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2679 for recursive directory scanning broke on some systems because the files
2680 are not scanned in any specific order and on different systems the order
2681 was different. A call to "sort" has been inserted into RunGrepTest for the
2682 approprate test as a short-term fix. In the longer term there may be an
2683 alternative.
2684
2685 3. I had an email from Eric Raymond about problems translating some of PCRE's
2686 man pages to HTML (despite the fact that I distribute HTML pages, some
2687 people do their own conversions for various reasons). The problems
2688 concerned the use of low-level troff macros .br and .in. I have therefore
2689 removed all such uses from the man pages (some were redundant, some could
2690 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2691 HTML has been updated to handle .nf/.fi and to complain if it encounters
2692 .br or .in.
2693
2694 4. Updated comments in configure.ac that get placed in config.h.in and also
2695 arranged for config.h to be included in the distribution, with the name
2696 config.h.generic, for the benefit of those who have to compile without
2697 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2698
2699 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2700 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2701 makevp.bat for the current PCRE, using the additional files
2702 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2703
2704 6. A Windows user reported a minor discrepancy with test 2, which turned out
2705 to be caused by a trailing space on an input line that had got lost in his
2706 copy. The trailing space was an accident, so I've just removed it.
2707
2708 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2709 that is needed.
2710
2711 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2712 as "const" (a) because they are and (b) because it helps the PHP
2713 maintainers who have recently made a script to detect big data structures
2714 in the php code that should be moved to the .rodata section. I remembered
2715 to update Builducptable as well, so it won't revert if ucptable.h is ever
2716 re-created.
2717
2718 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2719 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2720 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2721 support is not required. This saves 1.5-2K of code, which is important in
2722 some applications.
2723
2724 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2725 so as not to refer to the tables, even though these functions will never be
2726 called when UTF-8 support is disabled. Otherwise there are problems with a
2727 shared library.
2728
2729 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2730
2731 (a) It was defining its arguments as char * instead of void *.
2732
2733 (b) It was assuming that all moves were upwards in memory; this was true
2734 a long time ago when I wrote it, but is no longer the case.
2735
2736 The emulated memove() is provided for those environments that have neither
2737 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2738 is clearly not the case, as these two bugs were recently reported.
2739
2740 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2741 and Detrail to create the HTML documentation, the .txt form of the man
2742 pages, and it removes trailing spaces from listed files. It also creates
2743 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2744 case, it wraps all the #defines with #ifndefs. This script should be run
2745 before "make dist".
2746
2747 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2748 with Unicode property support.
2749
2750 (a) For a maximizing quantifier, if the two different cases of the
2751 character were of different lengths in their UTF-8 codings (there are
2752 some cases like this - I found 11), and the matching function had to
2753 back up over a mixture of the two cases, it incorrectly assumed they
2754 were both the same length.
2755
2756 (b) When PCRE was configured to use the heap rather than the stack for
2757 recursion during matching, it was not correctly preserving the data for
2758 the other case of a UTF-8 character when checking ahead for a match
2759 while processing a minimizing repeat. If the check also involved
2760 matching a wide character, but failed, corruption could cause an
2761 erroneous result when trying to check for a repeat of the original
2762 character.
2763
2764 13. Some tidying changes to the testing mechanism:
2765
2766 (a) The RunTest script now detects the internal link size and whether there
2767 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2768 values substituted by "configure". (The RunGrepTest script already did
2769 this for UTF-8.) The configure.ac script no longer substitutes the
2770 relevant variables.
2771
2772 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2773 with length and offset values. This means that the output is different
2774 for different internal link sizes. Test 2 is skipped for link sizes
2775 other than 2 because of this, bypassing the problem. Unfortunately,
2776 there was also a test in test 3 (the locale tests) that used /B and
2777 failed for link sizes other than 2. Rather than cut the whole test out,
2778 I have added a new /Z option to pcretest that replaces the length and
2779 offset values with spaces. This is now used to make test 3 independent
2780 of link size. (Test 2 will be tidied up later.)
2781
2782 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2783 segmentation fault instead of returning the appropriate error message.
2784
2785 15. In multiline mode when the newline sequence was set to "any", the pattern
2786 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2787 This doesn't seem right; it now treats the CRLF combination as the line
2788 ending, and so does not match in that case. It's only a pattern such as ^$
2789 that would hit this one: something like ^ABC$ would have failed after \r
2790 and then tried again after \r\n.
2791
2792 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2793 in an attempt to make files that differ only in their line terminators
2794 compare equal. This works on Linux.
2795
2796 17. Under certain error circumstances pcregrep might try to free random memory
2797 as it exited. This is now fixed, thanks to valgrind.
2798
2799 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2800 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2801 was because its rules for how to advance for /g after matching an empty
2802 string at the end of a line did not allow for this case. They now check for
2803 it specially.
2804
2805 20. pcretest is supposed to handle patterns and data of any length, by
2806 extending its buffers when necessary. It was getting this wrong when the
2807 buffer for a data line had to be extended.
2808
2809 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2810 CRLF as a newline sequence.
2811
2812 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2813 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2814 I have nevertheless tidied it up.
2815
2816 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2817
2818 24. Added a man page for pcre-config.
2819
2820
2821 Version 7.0 19-Dec-06
2822 ---------------------
2823
2824 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2825 moving to gcc 4.1.1.
2826
2827 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2828 sys/time.h, which is documented as needed for this function. It doesn't
2829 seem to matter on Linux, but it showed up on some releases of OS X.
2830
2831 3. It seems that there are systems where bytes whose values are greater than
2832 127 match isprint() in the "C" locale. The "C" locale should be the
2833 default when a C program starts up. In most systems, only ASCII printing
2834 characters match isprint(). This difference caused the output from pcretest
2835 to vary, making some of the tests fail. I have changed pcretest so that:
2836
2837 (a) When it is outputting text in the compiled version of a pattern, bytes
2838 other than 32-126 are always shown as hex escapes.
2839
2840 (b) When it is outputting text that is a matched part of a subject string,
2841 it does the same, unless a different locale has been set for the match
2842 (using the /L modifier). In this case, it uses isprint() to decide.
2843
2844 4. Fixed a major bug that caused incorrect computation of the amount of memory
2845 required for a compiled pattern when options that changed within the
2846 pattern affected the logic of the preliminary scan that determines the
2847 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2848 that the computed length was too small. The symptoms of this bug were
2849 either the PCRE error "internal error: code overflow" from pcre_compile(),
2850 or a glibc crash with a message such as "pcretest: free(): invalid next
2851 size (fast)". Examples of patterns that provoked this bug (shown in
2852 pcretest format) are:
2853
2854 /(?-x: )/x
2855 /(?x)(?-x: \s*#\s*)/
2856 /((?i)[\x{c0}])/8
2857 /(?i:[\x{c0}])/8
2858
2859 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2860 is now done differently.
2861
2862 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2863 wrapper classes; (b) implement a new function in the C++ scanner that is
2864 more efficient than the old way of doing things because it avoids levels of
2865 recursion in the regex matching; (c) add a paragraph to the documentation
2866 for the FullMatch() function.
2867
2868 6. The escape sequence \n was being treated as whatever was defined as
2869 "newline". Not only was this contrary to the documentation, which states
2870 that \n is character 10 (hex 0A), but it also went horribly wrong when
2871 "newline" was defined as CRLF. This has been fixed.
2872
2873 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2874 was being set to -1 for the "end of line" case (supposedly a value that no
2875 character can have). Though this value is never used (the check for end of
2876 line is "zero bytes in current character"), it caused compiler complaints.
2877 I've changed it to 0xffffffff.
2878
2879 8. In pcre_version.c, the version string was being built by a sequence of
2880 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2881 string (as it is for production releases) called a macro with an empty
2882 argument. The C standard says the result of this is undefined. The gcc
2883 compiler treats it as an empty string (which was what was wanted) but it is
2884 reported that Visual C gives an error. The source has been hacked around to
2885 avoid this problem.
2886
2887 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2888 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2889 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2890 of them did).
2891
2892 10. Originally, pcretest opened its input and output without "b"; then I was
2893 told that "b" was needed in some environments, so it was added for release
2894 5.0 to both the input and output. (It makes no difference on Unix-like
2895 systems.) Later I was told that it is wrong for the input on Windows. I've
2896 now abstracted the modes into two macros, to make it easier to fiddle with
2897 them, and removed "b" from the input mode under Windows.
2898
2899 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2900
2901 12. Added -help and --help to pcretest as an official way of being reminded
2902 of the options.
2903
2904 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2905 and pcrecpp.cc because they annoy compilers at high warning levels.
2906
2907 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2908
2909 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2910 scripts are not C programs :-) and which was not noticed because it works
2911 on Linux.
2912
2913 16. pcretest is supposed to handle any length of pattern and data line (as one
2914 line or as a continued sequence of lines) by extending its input buffer if
2915 necessary. This feature was broken for very long pattern lines, leading to
2916 a string of junk being passed to pcre_compile() if the pattern was longer
2917 than about 50K.
2918
2919 17. I have done a major re-factoring of the way pcre_compile() computes the
2920 amount of memory needed for a compiled pattern. Previously, there was code
2921 that made a preliminary scan of the pattern in order to do this. That was
2922 OK when PCRE was new, but as the facilities have expanded, it has become
2923 harder and harder to keep it in step with the real compile phase, and there
2924 have been a number of bugs (see for example, 4 above). I have now found a
2925 cunning way of running the real compile function in a "fake" mode that
2926 enables it to compute how much memory it would need, while actually only
2927 ever using a few hundred bytes of working memory and without too many
2928 tests of the mode. This should make future maintenance and development
2929 easier. A side effect of this work is that the limit of 200 on the nesting
2930 depth of parentheses has been removed (though this was never a serious
2931 limitation, I suspect). However, there is a downside: pcre_compile() now
2932 runs more slowly than before (30% or more, depending on the pattern). I
2933 hope this isn't a big issue. There is no effect on runtime performance.
2934
2935 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2936 newline (only possible for the last line of a file) and it was a
2937 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2938
2939 19. Added additional timing features to pcretest. (1) The -tm option now times
2940 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2941 separate command line item, by a number that specifies the number of
2942 repeats to use when timing. The default is 50000; this gives better
2943 precision, but takes uncomfortably long for very large patterns.
2944
2945 20. Extended pcre_study() to be more clever in cases where a branch of a
2946 subpattern has no definite first character. For example, (a*|b*)[cd] would
2947 previously give no result from pcre_study(). Now it recognizes that the
2948 first character must be a, b, c, or d.
2949
2950 21. There was an incorrect error "recursive call could loop indefinitely" if
2951 a subpattern (or the entire pattern) that was being tested for matching an
2952 empty string contained only one non-empty item after a nested subpattern.
2953 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2954 incorrectly, because the \d was being skipped in the check.
2955
2956 22. The pcretest program now has a new pattern option /B and a command line
2957 option -b, which is equivalent to adding /B to every pattern. This causes
2958 it to show the compiled bytecode, without the additional information that
2959 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2960 is the same as /B/I).
2961
2962 23. A new optimization is now able automatically to treat some sequences such
2963 as a*b as a*+b. More specifically, if something simple (such as a character
2964 or a simple class like \d) has an unlimited quantifier, and is followed by
2965 something that cannot possibly match the quantified thing, the quantifier
2966 is automatically "possessified".
2967
2968 24. A recursive reference to a subpattern whose number was greater than 39
2969 went wrong under certain circumstances in UTF-8 mode. This bug could also
2970 have affected the operation of pcre_study().
2971
2972 25. Realized that a little bit of performance could be had by replacing
2973 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2974
2975 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2976
2977 27. Possessive quantifiers such as a++ were previously implemented by turning
2978 them into atomic groups such as ($>a+). Now they have their own opcodes,
2979 which improves performance. This includes the automatically created ones
2980 from 23 above.
2981
2982 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2983 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2984 the first matched character to be a colon. This applied both to named and
2985 numbered groups.
2986
2987 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2988
2989 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2990 building PCRE on Windows easier, so I have included it in the distribution.
2991
2992 31. There is now a check in pcretest against a ridiculously large number being
2993 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2994 loop, the loop is abandoned.
2995
2996 32. Forward references to subpatterns in conditions such as (?(2)...) where
2997 subpattern 2 is defined later cause pcre_compile() to search forwards in
2998 the pattern for the relevant set of parentheses. This search went wrong
2999 when there were unescaped parentheses in a character class, parentheses
3000 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3001
3002 33. "Subroutine" calls and backreferences were previously restricted to
3003 referencing subpatterns earlier in the regex. This restriction has now
3004 been removed.
3005
3006 34. Added a number of extra features that are going to be in Perl 5.10. On the
3007 whole, these are just syntactic alternatives for features that PCRE had
3008 previously implemented using the Python syntax or my own invention. The
3009 other formats are all retained for compatibility.
3010
3011 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3012 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3013 also .NET compatible.
3014
3015 (b) A recursion or subroutine call to a named group can now be defined as
3016 (?&name) as well as (?P>name).
3017
3018 (c) A backreference to a named group can now be defined as \k<name> or
3019 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3020 5.10, are also .NET compatible.
3021
3022 (d) A conditional reference to a named group can now use the syntax
3023 (?(<name>) or (?('name') as well as (?(name).
3024
3025 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3026 groups (named and numbered) that are never evaluated inline, but can be
3027 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3028 is always false. There may be only one alternative in such a group.
3029
3030 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3031 as the simple (?(R). The condition is true only if the most recent
3032 recursion is that of the given number or name. It does not search out
3033 through the entire recursion stack.
3034
3035 (g) The escape \gN or \g{N} has been added, where N is a positive or
3036 negative number, specifying an absolute or relative reference.
3037
3038 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3039 some "unreachable code" warnings.
3040
3041 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3042 things, this adds five new scripts.
3043
3044 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3045 There were also incompatibilities regarding the handling of \Q..\E inside
3046 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3047 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3048
3049 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3050 matches an empty string, and forcibly breaks the loop. There were bugs in
3051 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3052 against aaaa the result was just "a" rather than "aaaa", for example. Two
3053 separate and independent bugs (that affected different cases) have been
3054 fixed.
3055
3056 39. Refactored the code to abolish the use of different opcodes for small
3057 capturing bracket numbers. This is a tidy that I avoided doing when I
3058 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3059 The new approach is not only tidier, it makes it possible to reduce the
3060 memory needed to fix the previous bug (38).
3061
3062 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3063 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3064 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3065 mode.
3066
3067 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3068 report.
3069
3070 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3071 copy construction and assignment in the C++ wrapper.
3072
3073 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3074 couple of bugs that could have given wrong results in the "--newline=crlf"
3075 case.
3076
3077 44. Added a number of casts and did some reorganization of signed/unsigned int
3078 variables following suggestions from Dair Grant. Also renamed the variable
3079 "this" as "item" because it is a C++ keyword.
3080
3081 45. Arranged for dftables to add
3082
3083 #include "pcre_internal.h"
3084
3085 to pcre_chartables.c because without it, gcc 4.x may remove the array
3086 definition from the final binary if PCRE is built into a static library and
3087 dead code stripping is activated.
3088
3089 46. For an unanchored pattern, if a match attempt fails at the start of a
3090 newline sequence, and the newline setting is CRLF or ANY, and the next two
3091 characters are CRLF, advance by two characters instead of one.
3092
3093
3094 Version 6.7 04-Jul-06
3095 ---------------------
3096
3097 1. In order to handle tests when input lines are enormously long, pcretest has
3098 been re-factored so that it automatically extends its buffers when
3099 necessary. The code is crude, but this _is_ just a test program. The
3100 default size has been increased from 32K to 50K.
3101
3102 2. The code in pcre_study() was using the value of the re argument before
3103 testing it for NULL. (Of course, in any sensible call of the function, it
3104 won't be NULL.)
3105
3106 3. The memmove() emulation function in pcre_internal.h, which is used on
3107 systems that lack both memmove() and bcopy() - that is, hardly ever -
3108 was missing a "static" storage class specifier.
3109
3110 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3111 containing an extended class (one that cannot be represented by a bitmap
3112 because it contains high-valued characters or Unicode property items, e.g.
3113 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3114 pattern, but PCRE should not loop if you do not (it no longer does).
3115 [Detail: two cases were found: (a) a repeated subpattern containing an
3116 extended class; (b) a recursive reference to a subpattern that followed a
3117 previous extended class. It wasn't skipping over the extended class
3118 correctly when UTF-8 mode was not set.]
3119
3120 5. A negated single-character class was not being recognized as fixed-length
3121 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3122 compile error "lookbehind assertion is not fixed length".
3123
3124 6. The RunPerlTest auxiliary script was showing an unexpected difference
3125 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3126 write a Perl script that can interpret lines of an input file either as
3127 byte characters or as UTF-8, which is what "perltest" was being required to
3128 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3129 can't do is switch easily at run time between having the "use utf8;" pragma
3130 or not. In the end, I fudged it by using the RunPerlTest script to insert
3131 "use utf8;" explicitly for the UTF-8 tests.
3132
3133 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3134 the end of the subject string, contrary to the documentation and to what
3135 Perl does. This was true of both matching functions. Now it matches only at
3136 the start of the subject and immediately after *internal* newlines.
3137
3138 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3139 a pointer to an int instead of a pointer to an unsigned long int. This
3140 caused problems on 64-bit systems.
3141
3142 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3143 instance of the 'standard' template library not being so standard".
3144
3145 10. There was no check on the number of named subpatterns nor the maximum
3146 length of a subpattern name. The product of these values is used to compute
3147 the size of the memory block for a compiled pattern. By supplying a very
3148 long subpattern name and a large number of named subpatterns, the size
3149 computation could be caused to overflow. This is now prevented by limiting
3150 the length of names to 32 characters, and the number of named subpatterns
3151 to 10,000.
3152
3153 11. Subpatterns that are repeated with specific counts have to be replicated in
3154 the compiled pattern. The size of memory for this was computed from the
3155 length of the subpattern and the repeat count. The latter is limited to
3156 65535, but there was no limit on the former, meaning that integer overflow
3157 could in principle occur. The compiled length of a repeated subpattern is
3158 now limited to 30,000 bytes in order to prevent this.
3159
3160 12. Added the optional facility to have named substrings with the same name.
3161
3162 13. Added the ability to use a named substring as a condition, using the
3163 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3164 are numbers (not recommended). Forward references are permitted.
3165
3166 14. Added forward references in named backreferences (if you see what I mean).
3167
3168 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3169 pattern could run off the end of the subject. For example, the pattern
3170 "(?s)(.{1,5})"8 did this with the subject "ab".
3171
3172 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3173 PCRE_CASELESS was set when matching characters that were quantified with ?
3174 or *.
3175
3176 17. A character class other than a single negated character that had a minimum
3177 but no maximum quantifier - for example [ab]{6,} - was not handled
3178 correctly by pce_dfa_exec(). It would match only one character.
3179
3180 18. A valid (though odd) pattern that looked like a POSIX character
3181 class but used an invalid character after [ (for example [[,abc,]]) caused
3182 pcre_compile() to give the error "Failed: internal error: code overflow" or
3183 in some cases to crash with a glibc free() error. This could even happen if
3184 the pattern terminated after [[ but there just happened to be a sequence of
3185 letters, a binary zero, and a closing ] in the memory that followed.
3186
3187 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3188 over the years. Originally (before any Unicode support), just the bottom 8
3189 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3190 output from "man perlunicode" includes this:
3191
3192 The regular expression compiler produces polymorphic opcodes. That
3193 is, the pattern adapts to the data and automatically switches to
3194 the Unicode character scheme when presented with Unicode data--or
3195 instead uses a traditional byte scheme when presented with byte
3196 data.
3197
3198 Sadly, a wide octal escape does not cause a switch, and in a string with
3199 no other multibyte characters, these octal escapes are treated as before.
3200 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3201 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3202 Unicode string.
3203
3204 I have not perpetrated such confusion in PCRE. Up till now, it took just
3205 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3206 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3207 translate to the appropriate multibyte character.
3208
3209 29. Applied some refactoring to reduce the number of warnings from Microsoft
3210 and Borland compilers. This has included removing the fudge introduced
3211 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3212 a warning about an unused variable.
3213
3214 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3215 characters since release 4.0, because Perl (from release 5.004) does not.
3216 [Or at least, is documented not to: some releases seem to be in conflict
3217 with the documentation.] However, when a pattern was studied with
3218 pcre_study() and all its branches started with \s, PCRE still included VT
3219 as a possible starting character. Of course, this did no harm; it just
3220 caused an unnecessary match attempt.
3221
3222 22. Removed a now-redundant internal flag bit that recorded the fact that case
3223 dependency changed within the pattern. This was once needed for "required
3224 byte" processing, but is no longer used. This recovers a now-scarce options
3225 bit. Also moved the least significant internal flag bit to the most-
3226 significant bit of the word, which was not previously used (hangover from
3227 the days when it was an int rather than a uint) to free up another bit for
3228 the future.
3229
3230 23. Added support for CRLF line endings as well as CR and LF. As well as the
3231 default being selectable at build time, it can now be changed at runtime
3232 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3233 specify that it is scanning data with non-default line endings.
3234
3235 24. Changed the definition of CXXLINK to make it agree with the definition of
3236 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3237
3238 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3239 recursions. This makes a big different to stack usage for some patterns.
3240
3241 26. If a subpattern containing a named recursion or subroutine reference such
3242 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3243 the space required for the compiled pattern went wrong and gave too small a
3244 value. Depending on the environment, this could lead to "Failed: internal
3245 error: code overflow at offset 49" or "glibc detected double free or
3246 corruption" errors.
3247
3248 27. Applied patches from Google (a) to support the new newline modes and (b) to
3249 advance over multibyte UTF-8 characters in GlobalReplace.
3250
3251 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3252 difference for some implementation of PCRE in some Windows version.
3253
3254 29. Added some extra testing facilities to pcretest:
3255
3256 \q<number> in a data line sets the "match limit" value
3257 \Q<number> in a data line sets the "match recursion limt" value
3258 -S <number> sets the stack size, where <number> is in megabytes
3259
3260 The -S option isn't available for Windows.
3261
3262
3263 Version 6.6 06-Feb-06
3264 ---------------------
3265
3266 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3267 in pcreposix.h. I have copied the definition from pcre.h.
3268
3269 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3270 because pcre.h is no longer a built file.
3271
3272 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3273 not normally included in the compiled code.
3274
3275
3276 Version 6.5 01-Feb-06
3277 ---------------------
3278
3279 1. When using the partial match feature with pcre_dfa_exec(), it was not
3280 anchoring the second and subsequent partial matches at the new starting
3281 point. This could lead to incorrect results. For example, with the pattern
3282 /1234/, partially matching against "123" and then "a4" gave a match.
3283
3284 2. Changes to pcregrep:
3285
3286 (a) All non-match returns from pcre_exec() were being treated as failures
3287 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3288 error message is output. Some extra information is given for the
3289 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3290 probably the only errors that are likely to be caused by users (by
3291 specifying a regex that has nested indefinite repeats, for instance).
3292 If there are more than 20 of these errors, pcregrep is abandoned.
3293
3294 (b) A binary zero was treated as data while matching, but terminated the
3295 output line if it was written out. This has been fixed: binary zeroes
3296 are now no different to any other data bytes.
3297
3298 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3299 used to set a locale for matching. The --locale=xxxx long option has
3300 been added (no short equivalent) to specify a locale explicitly on the
3301 pcregrep command, overriding the environment variables.
3302
3303 (d) When -B was used with -n, some line numbers in the output were one less
3304 than they should have been.
3305
3306 (e) Added the -o (--only-matching) option.
3307
3308 (f) If -A or -C was used with -c (count only), some lines of context were
3309 accidentally printed for the final match.
3310
3311 (g) Added the -H (--with-filename) option.
3312
3313 (h) The combination of options -rh failed to suppress file names for files
3314 that were found from directory arguments.
3315
3316 (i) Added the -D (--devices) and -d (--directories) options.
3317
3318 (j) Added the -F (--fixed-strings) option.
3319
3320 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3321
3322 (l) Added the --colo(u)r option.
3323
3324 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3325 is not present by default.
3326
3327 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3328 items such as (?R) or (?1), when the recursion could match a number of
3329 alternatives. If it matched one of the alternatives, but subsequently,
3330 outside the recursion, there was a failure, the code tried to back up into
3331 the recursion. However, because of the way PCRE is implemented, this is not
3332 possible, and the result was an incorrect result from the match.
3333
3334 In order to prevent this happening, the specification of recursion has
3335 been changed so that all such subpatterns are automatically treated as
3336 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3337
3338 4. I had overlooked the fact that, in some locales, there are characters for
3339 which isalpha() is true but neither isupper() nor islower() are true. In
3340 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3341 and ordfeminine) are like this. This affected the treatment of \w and \W
3342 when they appeared in character classes, but not when they appeared outside
3343 a character class. The bit map for "word" characters is now created
3344 separately from the results of isalnum() instead of just taking it from the
3345 upper, lower, and digit maps. (Plus the underscore character, of course.)
3346
3347 5. The above bug also affected the handling of POSIX character classes such as
3348 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3349 permanent tables. Instead, the bit maps for such a class were previously
3350 created as the appropriate unions of the upper, lower, and digit bitmaps.
3351 Now they are created by subtraction from the [[:word:]] class, which has
3352 its own bitmap.
3353
3354 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3355 It is created by subtracting the vertical space characters (\x09, \x0a,
3356 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3357 subtraction was done in the overall bitmap for a character class, meaning
3358 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3359 be recognized. This bug has been fixed.
3360
3361 7. Patches from the folks at Google:
3362
3363 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3364 real life, but is still worth protecting against".
3365
3366 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3367 regular expressions".
3368
3369 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3370 have it.
3371
3372 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3373 "configure" and the latter not, in order to fix a problem somebody had
3374 with compiling the Arg class on HP-UX.
3375
3376 (e) Improve the error-handling of the C++ wrapper a little bit.
3377
3378 (f) New tests for checking recursion limiting.
3379
3380 8. The pcre_memmove() function, which is used only if the environment does not
3381 have a standard memmove() function (and is therefore rarely compiled),
3382 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3383 returning a result (though PCRE never actually uses the result).
3384
3385 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3386 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3387 returned instead of calling malloc() with an overflowing number that would
3388 most likely cause subsequent chaos.
3389
3390 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3391
3392 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3393 with this option is matched, the nmatch and pmatch options of regexec() are
3394 ignored.
3395
3396 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3397 provided in case anyone wants to the the POSIX interface with UTF-8
3398 strings.
3399
3400 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3401 C++ linking (needed for some HP-UX environments).
3402
3403 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3404 (unused parameter) and in the pcre_printint() function (omitted "default"
3405 switch label when the default is to do nothing).
3406
3407 15. Added some code to make it possible, when PCRE is compiled as a C++
3408 library, to replace subject pointers for pcre_exec() with a smart pointer
3409 class, thus making it possible to process discontinuous strings.
3410
3411 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3412 much the same function. They were added by different people who were trying
3413 to make PCRE easy to compile on non-Unix systems. It has been suggested
3414 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3415 for compiling on Windows systems. I have therefore replaced it with
3416 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3417 defaults to "extern" for C or "extern C" for C++, which works fine on
3418 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3419 SCOPE with something explicit in config.h. In addition:
3420
3421 (a) pcreposix.h still had just "extern" instead of either of these macros;
3422 I have replaced it with PCRE_DATA_SCOPE.
3423
3424 (b) Functions such as _pcre_xclass(), which are internal to the library,
3425 but external in the C sense, all had PCRE_EXPORT in their definitions.
3426 This is apparently wrong for the Windows case, so I have removed it.
3427 (It makes no difference on Unix-like systems.)
3428
3429 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3430 of recursive calls to match(). This is different to MATCH_LIMIT because
3431 that limits the total number of calls to match(), not all of which increase
3432 the depth of recursion. Limiting the recursion depth limits the amount of
3433 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3434 when PCRE is compiled, and changed at run time. A patch from Google adds
3435 this functionality to the C++ interface.
3436
3437 18. Changes to the handling of Unicode character properties:
3438
3439 (a) Updated the table to Unicode 4.1.0.
3440
3441 (b) Recognize characters that are not in the table as "Cn" (undefined).
3442
3443 (c) I revised the way the table is implemented to a much improved format
3444 which includes recognition of ranges. It now supports the ranges that
3445 are defined in UnicodeData.txt, and it also amalgamates other
3446 characters into ranges. This has reduced the number of entries in the
3447 table from around 16,000 to around 3,000, thus reducing its size
3448 considerably. I realized I did not need to use a tree structure after
3449 all - a binary chop search is just as efficient. Having reduced the
3450 number of entries, I extended their size from 6 bytes to 8 bytes to
3451 allow for more data.
3452
3453 (d) Added support for Unicode script names via properties such as \p{Han}.
3454
3455 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3456 matching that character.
3457
3458 20. When matching a repeated Unicode property with a minimum greater than zero,
3459 (for example \pL{2,}), PCRE could look past the end of the subject if it
3460 reached it while seeking the minimum number of characters. This could
3461 happen only if some of the characters were more than one byte long, because
3462 there is a check for at least the minimum number of bytes.
3463
3464 21. Refactored the implementation of \p and \P so as to be more general, to
3465 allow for more different types of property in future. This has changed the
3466 compiled form incompatibly. Anybody with saved compiled patterns that use
3467 \p or \P will have to recompile them.
3468
3469 22. Added "Any" and "L&" to the supported property types.
3470
3471 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3472 but give a compile time error if the value is greater than 0xff.
3473
3474 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3475 accidentally not being installed or uninstalled.
3476
3477 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3478 made were to insert the current release number. This seemed silly, because
3479 it made things harder for people building PCRE on systems that don't run
3480 "configure". I have turned pcre.h into a distributed file, no longer built
3481 by "configure", with the version identification directly included. There is
3482 no longer a pcre.h.in file.
3483
3484 However, this change necessitated a change to the pcre-config script as
3485 well. It is built from pcre-config.in, and one of the substitutions was the
3486 release number. I have updated configure.ac so that ./configure now finds
3487 the release number by grepping pcre.h.
3488
3489 26. Added the ability to run the tests under valgrind.
3490
3491
3492 Version 6.4 05-Sep-05
3493 ---------------------
3494
3495 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3496 "--" to be printed when multiple files were scanned, even when none of the
3497 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3498 consider it to be a bug, and have restored the previous behaviour.
3499
3500 2. A couple of code tidies to get rid of compiler warnings.
3501
3502 3. The pcretest program used to cheat by referring to symbols in the library
3503 whose names begin with _pcre_. These are internal symbols that are not
3504 really supposed to be visible externally, and in some environments it is
3505 possible to suppress them. The cheating is now confined to including
3506 certain files from the library's source, which is a bit cleaner.
3507
3508 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3509 file's purpose clearer.
3510
3511 5. Reorganized pcre_ucp_findchar().
3512
3513
3514 Version 6.3 15-Aug-05
3515 ---------------------
3516
3517 1. The file libpcre.pc.in did not have general read permission in the tarball.
3518
3519 2. There were some problems when building without C++ support:
3520
3521 (a) If C++ support was not built, "make install" and "make test" still
3522 tried to test it.
3523
3524 (b) There were problems when the value of CXX was explicitly set. Some
3525 changes have been made to try to fix these, and ...
3526
3527 (c) --disable-cpp can now be used to explicitly disable C++ support.
3528
3529 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3530 backslash in a target when C++ was disabled. This confuses some
3531 versions of "make", apparently. Using an intermediate variable solves
3532 this. (Same for CPP_LOBJ.)
3533
3534 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3535 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3536 necessary on certain architectures.
3537
3538 4. Added a setting of -export-symbols-regex to the link command to remove
3539 those symbols that are exported in the C sense, but actually are local
3540 within the library, and not documented. Their names all begin with
3541 "_pcre_". This is not a perfect job, because (a) we have to except some
3542 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3543 available (and never for static libraries). I have made a note to try to
3544 find a way round (a) in the future.
3545
3546
3547 Version 6.2 01-Aug-05
3548 ---------------------
3549
3550 1. There was no test for integer overflow of quantifier values. A construction
3551 such as {1111111111111111} would give undefined results. What is worse, if
3552 a minimum quantifier for a parenthesized subpattern overflowed and became
3553 negative, the calculation of the memory size went wrong. This could have
3554 led to memory overwriting.
3555
3556 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3557
3558 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3559 operating environments where this matters.
3560
3561 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3562 PCRE options from within the C++ wrapper.
3563
3564 5. Named capturing subpatterns were not being correctly counted when a pattern
3565 was compiled. This caused two problems: (a) If there were more than 100
3566 such subpatterns, the calculation of the memory needed for the whole
3567 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3568 back references of the form \12, where the number was greater than 9, were
3569 not recognized as back references, even though there were sufficient
3570 previous subpatterns.
3571
3572 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3573 versions of gcc, e.g. 2.95.4.
3574
3575
3576 Version 6.1 21-Jun-05
3577 ---------------------
3578
3579 1. There was one reference to the variable "posix" in pcretest.c that was not
3580 surrounded by "#if !defined NOPOSIX".
3581
3582 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3583 the cross-check on the old pcre_info() function, for the benefit of the
3584 cut-down version of PCRE that is currently imported into Exim.
3585
3586 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3587 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3588 patterns that start (?i) but I don't think that matters. The use of (?i) is
3589 just an example; this all applies to the other options as well.
3590
3591 4. Since libtool seems to echo the compile commands it is issuing, the output
3592 from "make" can be reduced a bit by putting "@" in front of each libtool
3593 compile command.
3594
3595 5. Patch from the folks at Google for configure.in to be a bit more thorough
3596 in checking for a suitable C++ installation before trying to compile the
3597 C++ stuff. This should fix a reported problem when a compiler was present,
3598 but no suitable headers.
3599
3600 6. The man pages all had just "PCRE" as their title. I have changed them to
3601 be the relevant file name. I have also arranged that these names are
3602 retained in the file doc/pcre.txt, which is a concatenation in text format
3603 of all the man pages except the little individual ones for each function.
3604
3605 7. The NON-UNIX-USE file had not been updated for the different set of source
3606 files that come with release 6. I also added a few comments about the C++
3607 wrapper.
3608
3609
3610 Version 6.0 07-Jun-05
3611 ---------------------
3612
3613 1. Some minor internal re-organization to help with my DFA experiments.
3614
3615 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3616 didn't matter for the library itself when fully configured, but did matter
3617 when compiling without UCP support, or within Exim, where the ucp files are
3618 not imported.
3619
3620 3. Refactoring of the library code to split up the various functions into
3621 different source modules. The addition of the new DFA matching code (see
3622 below) to a single monolithic source would have made it really too
3623 unwieldy, quite apart from causing all the code to be include in a
3624 statically linked application, when only some functions are used. This is
3625 relevant even without the DFA addition now that patterns can be compiled in
3626 one application and matched in another.
3627
3628 The downside of splitting up is that there have to be some external
3629 functions and data tables that are used internally in different modules of
3630 the library but which are not part of the API. These have all had their
3631 names changed to start with "_pcre_" so that they are unlikely to clash
3632 with other external names.
3633
3634 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3635 a different (DFA) algorithm. Although it is slower than the original
3636 function, it does have some advantages for certain types of matching
3637 problem.
3638
3639 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3640 including restarting after a partial match.
3641
3642 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3643 defined when compiling for Windows was sent to me. I have put it into the
3644 code, though I have no means of testing or verifying it.
3645
3646 7. Added the pcre_refcount() auxiliary function.
3647
3648 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3649 match before or at the first newline in the subject string. In pcretest,
3650 the /f option on a pattern can be used to set this.
3651
3652 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3653 would behave wrongly. This has been present in PCRE since release 4.0.
3654
3655 10. A number of changes to the pcregrep command:
3656
3657 (a) Refactored how -x works; insert ^(...)$ instead of setting
3658 PCRE_ANCHORED and checking the length, in preparation for adding
3659 something similar for -w.
3660
3661 (b) Added the -w (match as a word) option.
3662
3663 (c) Refactored the way lines are read and buffered so as to have more
3664 than one at a time available.
3665
3666 (d) Implemented a pcregrep test script.
3667
3668 (e) Added the -M (multiline match) option. This allows patterns to match
3669 over several lines of the subject. The buffering ensures that at least
3670 8K, or the rest of the document (whichever is the shorter) is available
3671 for matching (and similarly the previous 8K for lookbehind assertions).
3672
3673 (f) Changed the --help output so that it now says
3674
3675 -w, --word-regex(p)
3676
3677 instead of two lines, one with "regex" and the other with "regexp"
3678 because that confused at least one person since the short forms are the
3679 same. (This required a bit of code, as the output is generated
3680 automatically from a table. It wasn't just a text change.)
3681
3682 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3683 option but starts with a hyphen. Could be a pattern or a path name
3684 starting with a hyphen, for instance.
3685
3686 (h) "-" can be given as a file name to represent stdin.
3687
3688 (i) When file names are being printed, "(standard input)" is used for
3689 the standard input, for compatibility with GNU grep. Previously
3690 "<stdin>" was used.
3691
3692 (j) The option --label=xxx can be used to supply a name to be used for
3693 stdin when file names are being printed. There is no short form.
3694
3695 (k) Re-factored the options decoding logic because we are going to add
3696 two more options that take data. Such options can now be given in four
3697 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3698
3699 (l) Added the -A, -B, and -C options for requesting that lines of context
3700 around matches be printed.
3701
3702 (m) Added the -L option to print the names of files that do not contain
3703 any matching lines, that is, the complement of -l.
3704
3705 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3706 continue to scan other files.
3707
3708 (o) The -s option was incorrectly implemented. For compatibility with other
3709 greps, it now suppresses the error message for a non-existent or non-
3710 accessible file (but not the return code). There is a new option called
3711 -q that suppresses the output of matching lines, which was what -s was
3712 previously doing.
3713
3714 (p) Added --include and --exclude options to specify files for inclusion
3715 and exclusion when recursing.
3716
3717 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3718 Hopefully, it now does.
3719
3720 12. Missing cast in pcre_study().
3721
3722 13. Added an "uninstall" target to the makefile.
3723
3724 14. Replaced "extern" in the function prototypes in Makefile.in with
3725 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3726 world, but is set differently for Windows.
3727
3728 15. Added a second compiling function called pcre_compile2(). The only
3729 difference is that it has an extra argument, which is a pointer to an
3730 integer error code. When there is a compile-time failure, this is set
3731 non-zero, in addition to the error test pointer being set to point to an
3732 error message. The new argument may be NULL if no error number is required
3733 (but then you may as well call pcre_compile(), which is now just a
3734 wrapper). This facility is provided because some applications need a
3735 numeric error indication, but it has also enabled me to tidy up the way
3736 compile-time errors are handled in the POSIX wrapper.
3737
3738 16. Added VPATH=.libs to the makefile; this should help when building with one
3739 prefix path and installing with another. (Or so I'm told by someone who
3740 knows more about this stuff than I do.)
3741
3742 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3743 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3744 match everything, including newlines. This is not POSIX-compatible, but
3745 somebody wanted the feature. From pcretest it can be activated by using
3746 both the P and the s flags.
3747
3748 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3749
3750 19. libpcre.pc was being incorrectly installed as executable.
3751
3752 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3753 it now also looks for '\r' so that it will work unmodified on Windows.
3754
3755 21. Added Google's contributed C++ wrapper to the distribution.
3756
3757 22. Added some untidy missing memory free() calls in pcretest, to keep
3758 Electric Fence happy when testing.
3759
3760
3761
3762 Version 5.0 13-Sep-04
3763 ---------------------
3764
3765 1. Internal change: literal characters are no longer packed up into items
3766 containing multiple characters in a single byte-string. Each character
3767 is now matched using a separate opcode. However, there may be more than one
3768 byte in the character in UTF-8 mode.
3769
3770 2. The pcre_callout_block structure has two new fields: pattern_position and
3771 next_item_length. These contain the offset in the pattern to the next match
3772 item, and its length, respectively.
3773
3774 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3775 insertion of callouts before each pattern item. Added the /C option to
3776 pcretest to make use of this.
3777
3778 4. On the advice of a Windows user, the lines
3779
3780 #if defined(_WIN32) || defined(WIN32)
3781 _setmode( _fileno( stdout ), 0x8000 );
3782 #endif /* defined(_WIN32) || defined(WIN32) */
3783
3784 have been added to the source of pcretest. This apparently does useful
3785 magic in relation to line terminators.
3786
3787 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3788 for the benefit of those environments where the "b" makes a difference.
3789
3790 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3791 to know about it. I have put a hack into configure.in that adds in code
3792 to set GCC=yes if CC=icc. This seems to end up at a point in the
3793 generated configure script that is early enough to affect the setting of
3794 compiler options, which is what is needed, but I have no means of testing
3795 whether it really works. (The user who reported this had patched the
3796 generated configure script, which of course I cannot do.)
3797
3798 LATER: After change 22 below (new libtool files), the configure script
3799 seems to know about icc (and also ecc). Therefore, I have commented out
3800 this hack in configure.in.
3801
3802 7. Added support for pkg-config (2 patches were sent in).
3803
3804 8. Negated POSIX character classes that used a combination of internal tables
3805 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3806 [[:^ascii]]. Typically, they would match almost any characters. The other
3807 POSIX classes were not broken in this way.
3808
3809 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3810 to find the match, as PCRE was deluded into thinking that the match had to
3811 start at the start point or following a newline. The same bug applied to
3812 patterns with negative forward assertions or any backward assertions
3813 preceding ".*" at the start, unless the pattern required a fixed first
3814 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3815
3816 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3817 starting at the last subject character, bytes beyond the end of the subject
3818 string were read.
3819
3820 11. Renamed the variable "class" as "classbits" to make life easier for C++
3821 users. (Previously there was a macro definition, but it apparently wasn't
3822 enough.)
3823
3824 12. Added the new field "tables" to the extra data so that tables can be passed
3825 in at exec time, or the internal tables can be re-selected. This allows
3826 a compiled regex to be saved and re-used at a later time by a different
3827 program that might have everything at different addresses.
3828
3829 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3830 -R library as well as a -L library.
3831
3832 14. The debugging options of pcretest (-d on the command line or D on a
3833 pattern) showed incorrect output for anything following an extended class
3834 that contained multibyte characters and which was followed by a quantifier.
3835
3836 15. Added optional support for general category Unicode character properties
3837 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3838 support. It adds about 90K to the size of the library. The meanings of the
3839 inbuilt class escapes such as \d and \s have NOT been changed.
3840
3841 16. Updated pcredemo.c to include calls to free() to release the memory for the
3842 compiled pattern.
3843
3844 17. The generated file chartables.c was being created in the source directory
3845 instead of in the building directory. This caused the build to fail if the
3846 source directory was different from the building directory, and was
3847 read-only.
3848
3849 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3850 file. No doubt somebody will tell me if they don't make sense... Also added
3851 Dan Mooney's comments about building on OpenVMS.
3852
3853 19. Added support for partial matching via the PCRE_PARTIAL option for
3854 pcre_exec() and the \P data escape in pcretest.
3855
3856 20. Extended pcretest with 3 new pattern features:
3857
3858 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3859 write the compiled pattern to the file whose name is "rest-of-line".
3860 This is a straight binary dump of the data, with the saved pointer to
3861 the character tables forced to be NULL. The study data, if any, is
3862 written too. After writing, pcretest reads a new pattern.
3863
3864 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3865 compiled pattern from the given file. There must not be any
3866 occurrences of "<" in the file name (pretty unlikely); if there are,
3867 pcretest will instead treat the initial "<" as a pattern delimiter.
3868 After reading in the pattern, pcretest goes on to read data lines as
3869 usual.
3870
3871 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3872 and 16-bit fields in a compiled pattern, to simulate a pattern that
3873 was compiled on a host of opposite endianness.
3874
3875 21. The pcre-exec() function can now cope with patterns that were compiled on
3876 hosts of opposite endianness, with this restriction:
3877
3878 As for any compiled expression that is saved and used later, the tables
3879 pointer field cannot be preserved; the extra_data field in the arguments
3880 to pcre_exec() should be used to pass in a tables address if a value
3881 other than the default internal tables were used at compile time.
3882
3883 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3884 now diagnosed as an error. Previously, most of the time, a negative number
3885 would have been treated as zero, but if in addition "ovector" was passed as
3886 NULL, a crash could occur.
3887
3888 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3889 new versions from the libtool 1.5 distribution (the last one is a copy of
3890 a file called libtool.m4). This seems to have fixed the need to patch
3891 "configure" to support Darwin 1.3 (which I used to do). However, I still
3892 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3893 workstation).
3894
3895 24. Changed the PCRE licence to be the more standard "BSD" licence.
3896
3897
3898 Version 4.5 01-Dec-03
3899 ---------------------
3900
3901 1. There has been some re-arrangement of the code for the match() function so
3902 that it can be compiled in a version that does not call itself recursively.
3903 Instead, it keeps those local variables that need separate instances for
3904 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3905 needs to "recurse". Keeping track of where control must go is done by means
3906 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3907 hide most of the details from the main code, and operates only if
3908 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3909 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3910 operating.
3911
3912 To make it easier for callers to provide specially tailored get/free
3913 functions for this usage, two new functions, pcre_stack_malloc, and
3914 pcre_stack_free, are used. They are always called in strict stacking order,
3915 and the size of block requested is always the same.
3916
3917 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3918 PCRE has been compiled to use the stack or the heap for recursion. The
3919 -C option of pcretest uses this to show which version is compiled.
3920
3921 A new data escape \S, is added to pcretest; it causes the amounts of store
3922 obtained and freed by both kinds of malloc/free at match time to be added
3923 to the output.
3924
3925 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3926 what's available on my current Linux desktop machine.
3927
3928 3. When matching a UTF-8 string, the test for a valid string at the start has
3929 been extended. If start_offset is not zero, PCRE now checks that it points
3930 to a byte that is the start of a UTF-8 character. If not, it returns
3931 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3932 this is necessary because there may be backward assertions in the pattern.
3933 When matching the same subject several times, it may save resources to use
3934 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3935
3936 4. The code for checking the validity of UTF-8 strings has been tightened so
3937 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3938 containing "overlong sequences".
3939
3940 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3941 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3942 should have been "&", but it just so happened that all the cases this let
3943 through by mistake were picked up later in the function.
3944
3945 6. I had used a variable called "isblank" - this is a C99 function, causing
3946 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3947
3948 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3949 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3950 so the version is shown; (c) stop "make test" echoing "./RunTest".
3951
3952 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3953
3954 9. The prototype for memmove() for systems that don't have it was using
3955 size_t, but the inclusion of the header that defines size_t was later. I've
3956 moved the #includes for the C headers earlier to avoid this.
3957
3958 10. Added some adjustments to the code to make it easier to compiler on certain
3959 special systems:
3960
3961 (a) Some "const" qualifiers were missing.
3962 (b) Added the macro EXPORT before all exported functions; by default this
3963 is defined to be empty.
3964 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3965 that it reads its output file name as an argument instead of writing
3966 to the standard output and assuming this can be redirected.
3967
3968 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3969 class containing characters with values greater than 255, PCRE compilation
3970 went into a loop.
3971
3972 12. A recursive reference to a subpattern that was within another subpattern
3973 that had a minimum quantifier of zero caused PCRE to crash. For example,
3974 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3975 recursion. If the recursively-called subpattern itself had a zero repeat,
3976 that was OK.
3977
3978 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3979 buffer into which it was copied (for escape processing) was still set at
3980 1024, so long lines caused crashes.
3981
3982 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3983 "internal error: code overflow...". This applied to any character class
3984 that was followed by a possessive quantifier.
3985
3986 15. Modified the Makefile to add libpcre.la as a prerequisite for
3987 libpcreposix.la because I was told this is needed for a parallel build to
3988 work.
3989
3990 16. If a pattern that contained .* following optional items at the start was
3991 studied, the wrong optimizing data was generated, leading to matching
3992 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3993 matching string must start with a or b or c. The correct conclusion for
3994 this pattern is that a match can start with any character.
3995
3996
3997 Version 4.4 13-Aug-03
3998 ---------------------
3999
4000 1. In UTF-8 mode, a character class containing characters with values between
4001 127 and 255 was not handled correctly if the compiled pattern was studied.
4002 In fixing this, I have also improved the studying algorithm for such
4003 classes (slightly).
4004
4005 2. Three internal functions had redundant arguments passed to them. Removal
4006 might give a very teeny performance improvement.
4007
4008 3. Documentation bug: the value of the capture_top field in a callout is *one
4009 more than* the number of the hightest numbered captured substring.
4010
4011 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4012 in incorrectly linking with a previously installed version. They now link
4013 explicitly with libpcre.la.
4014
4015 5. configure.in no longer needs to recognize Cygwin specially.
4016
4017 6. A problem in pcre.in for Windows platforms is fixed.
4018
4019 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4020 pcretest, it used to include the size of the study block as part of its
4021 output. Unfortunately, the structure contains a field that has a different
4022 size on different hardware architectures. This meant that the tests that
4023 showed this size failed. As the block is currently always of a fixed size,
4024 this information isn't actually particularly useful in pcretest output, so
4025 I have just removed it.
4026
4027 8. Three pre-processor statements accidentally did not start in column 1.
4028 Sadly, there are *still* compilers around that complain, even though
4029 standard C has not required this for well over a decade. Sigh.
4030
4031 9. In pcretest, the code for checking callouts passed small integers in the
4032 callout_data field, which is a void * field. However, some picky compilers
4033 complained about the casts involved for this on 64-bit systems. Now
4034 pcretest passes the address of the small integer instead, which should get
4035 rid of the warnings.
4036
4037 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4038 both compile and run time, and gives an error if an invalid UTF-8 sequence
4039 is found. There is a option for disabling this check in cases where the
4040 string is known to be correct and/or the maximum performance is wanted.
4041
4042 11. In response to a bug report, I changed one line in Makefile.in from
4043
4044 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4045 to
4046 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4047
4048 to look similar to other lines, but I have no way of telling whether this
4049 is the right thing to do, as I do not use Windows. No doubt I'll get told
4050 if it's wrong...
4051
4052
4053 Version 4.3 21-May-03
4054 ---------------------
4055
4056 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4057 Makefile.
4058
4059 2. Some refactoring to improve the quality of the code:
4060
4061 (i) The utf8_table... variables are now declared "const".
4062
4063 (ii) The code for \cx, which used the "case flipping" table to upper case
4064 lower case letters, now just substracts 32. This is ASCII-specific,
4065 but the whole concept of \cx is ASCII-specific, so it seems
4066 reasonable.
4067
4068 (iii) PCRE was using its character types table to recognize decimal and
4069 hexadecimal digits in the pattern. This is silly, because it handles
4070 only 0-9, a-f, and A-F, but the character types table is locale-
4071 specific, which means strange things might happen. A private
4072 table is now used for this - though it costs 256 bytes, a table is
4073 much faster than multiple explicit tests. Of course, the standard
4074 character types table is still used for matching digits in subject
4075 strings against \d.
4076
4077 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4078 ending in _t are). So I've renamed it as ESC_tee.
4079
4080 3. The first argument for regexec() in the POSIX wrapper should have been
4081 defined as "const".
4082
4083 4. Changed pcretest to use malloc() for its buffers so that they can be
4084 Electric Fenced for debugging.
4085
4086 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4087 to read one or more bytes before the start of the subject string. Often this
4088 had no effect on PCRE's behaviour, but in some circumstances it could
4089 provoke a segmentation fault.
4090
4091 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4092 to try to read one or more bytes before the start of the subject string.
4093
4094 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4095 UTF-8 support could misbehave in various ways if the subject string
4096 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4097 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4098 back over UTF-8 characters.)
4099
4100
4101 Version 4.2 14-Apr-03
4102 ---------------------
4103
4104 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4105
4106 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4107 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4108 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4109 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4110 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4111 and BUILD_EXEEXT
4112 Note: automatic setting of the BUILD variables is not yet working
4113 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4114 compile-time but not at link-time
4115 [LINK]: use for linking executables only
4116 make different versions for Windows and non-Windows
4117 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4118 libraries
4119 [LINK_FOR_BUILD]: new variable
4120 [OBJEXT]: use throughout
4121 [EXEEXT]: use throughout
4122 <winshared>: new target
4123 <wininstall>: new target
4124 <dftables.o>: use native compiler
4125 <dftables>: use native linker
4126 <install>: handle Windows platform correctly
4127 <clean>: ditto
4128 <check>: ditto
4129 copy DLL to top builddir before testing
4130
4131 As part of these changes, -no-undefined was removed again. This was reported
4132 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4133 in any case.
4134
4135 3. Some tidies to get rid of compiler warnings:
4136
4137 . In the match_data structure, match_limit was an unsigned long int, whereas
4138 match_call_count was an int. I've made them both unsigned long ints.
4139
4140 . In pcretest the fact that a const uschar * doesn't automatically cast to
4141 a void * provoked a warning.
4142
4143 . Turning on some more compiler warnings threw up some "shadow" variables
4144 and a few more missing casts.
4145
4146 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4147 option, a class that contained a single character with a value between 128
4148 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4149
4150 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4151 option, a class that contained several characters, but with at least one
4152 whose value was between 128 and 255 caused PCRE to crash.
4153
4154
4155 Version 4.1 12-Mar-03
4156 ---------------------
4157
4158 1. Compiling with gcc -pedantic found a couple of places where casts were
4159 needed, and a string in dftables.c that was longer than standard compilers are
4160 required to support.
4161
4162 2. Compiling with Sun's compiler found a few more places where the code could
4163 be tidied up in order to avoid warnings.
4164
4165 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4166 first of these names is deprecated in the latest Autoconf in favour of the name
4167 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4168 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4169 analogy I have changed it to CFLAGS_FOR_BUILD.
4170
4171 4. Added -no-undefined to the linking command in the Makefile, because this is
4172 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4173 linking step for the pcreposix library.
4174
4175 5. PCRE was failing to diagnose the case of two named groups with the same
4176 name.
4177
4178 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4179 literal character that is needed in the subject for a match, and scans along to
4180 ensure that it is present before embarking on the full matching process. This
4181 saves time in cases of nested unlimited repeats that are never going to match.
4182 Problem: the scan can take a lot of time if the subject is very long (e.g.
4183 megabytes), thus penalizing straightforward matches. It is now done only if the
4184 amount of subject to be scanned is less than 1000 bytes.
4185
4186 7. A lesser problem with the same optimization is that it was recording the
4187 first character of an anchored pattern as "needed", thus provoking a search
4188 right along the subject, even when the first match of the pattern was going to
4189 fail. The "needed" character is now not set for anchored patterns, unless it
4190 follows something in the pattern that is of non-fixed length. Thus, it still
4191 fulfils its original purpose of finding quick non-matches in cases of nested
4192 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4193
4194
4195 Version 4.0 17-Feb-03
4196 ---------------------
4197
4198 1. If a comment in an extended regex that started immediately after a meta-item
4199 extended to the end of string, PCRE compiled incorrect data. This could lead to
4200 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4201
4202 2. Moved to autoconf 2.53 and libtool 1.4.2.
4203
4204 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4205 the special perltest8 script is no longer needed - all the tests can be run
4206 from a single perltest script.
4207
4208 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4209 by \s. It has now been removed in PCRE. This means it isn't recognized as
4210 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4211 class [:space:] *does* include VT, thereby creating a mess.
4212
4213 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4214 space and tab.
4215
4216 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4217 its new features into the main test script, reducing the number of scripts.
4218
4219 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4220 were backward compatible, and made the (?i) apply to the whole pattern, as if
4221 /i were given. Now it behaves more logically, and applies the option setting
4222 only to what follows. PCRE has been changed to follow suit. However, if it
4223 finds options settings right at the start of the pattern, it extracts them into
4224 the global options, as before. Thus, they show up in the info data.
4225
4226 8. Added support for the \Q...\E escape sequence. Characters in between are
4227 treated as literals. This is slightly different from Perl in that $ and @ are
4228 also handled as literals inside the quotes. In Perl, they will cause variable
4229 interpolation. Note the following examples:
4230
4231 Pattern PCRE matches Perl matches
4232
4233 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4234 \Qabc\$xyz\E abc\$xyz abc\$xyz
4235 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4236
4237 For compatibility with Perl, \Q...\E sequences are recognized inside character
4238 classes as well as outside them.
4239
4240 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4241 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4242 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4243 signed/unsigned warnings.
4244
4245 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4246 option for pcretest, so I've replaced it by a simple function that does just
4247 that job.
4248
4249 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4250 "pcregrep -".
4251
4252 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4253 Java package. This provides some syntactic sugar for simple cases of what my
4254 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4255 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4256 item, you can use this simplified notation. Note that only makes sense with
4257 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4258 greediness, whatever the setting of the PCRE_UNGREEDY option.
4259
4260 13. A change of greediness default within a pattern was not taking effect at
4261 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4262 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4263 was abstracted outside.
4264
4265 14. PCRE now supports the \G assertion. It is true when the current matching
4266 position is at the start point of the match. This differs from \A when the
4267 starting offset is non-zero. Used with the /g option of pcretest (or similar
4268 code), it works in the same way as it does for Perl's /g option. If all
4269 alternatives of a regex begin with \G, the expression is anchored to the start
4270 match position, and the "anchored" flag is set in the compiled expression.
4271
4272 15. Some bugs concerning the handling of certain option changes within patterns
4273 have been fixed. These applied to options other than (?ims). For example,
4274 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4275 been the other way round. Some of this was related to change 7 above.
4276
4277 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4278 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4279 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4280 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4281
4282 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4283 mode. Unlike ".", it always matches newline, whatever the setting of
4284 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4285 assertions. Perl allows it, but it doesn't (in general) work because it can't
4286 calculate the length of the lookbehind. At least, that's the case for Perl
4287 5.8.0 - I've been told they are going to document that it doesn't work in
4288 future.
4289
4290 18. Added an error diagnosis for escapes that PCRE does not support: these are
4291 \L, \l, \N, \P, \p, \U, \u, and \X.
4292
4293 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4294 reading past the end of the pattern in cases such as /[abcd/.
4295
4296 20. PCRE was getting more memory than necessary for patterns with classes that
4297 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4298
4299 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4300 compiling PCRE for use with Virtual Pascal.
4301
4302 22. Small fix to the Makefile to make it work properly if the build is done
4303 outside the source tree.
4304
4305 23. Added a new extension: a condition to go with recursion. If a conditional
4306 subpattern starts with (?(R) the "true" branch is used if recursion has
4307 happened, whereas the "false" branch is used only at the top level.
4308
4309 24. When there was a very long string of literal characters (over 255 bytes
4310 without UTF support, over 250 bytes with UTF support), the computation of how
4311 much memory was required could be incorrect, leading to segfaults or other
4312 strange effects.
4313
4314 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4315 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4316 there was a subsequent back reference to those brackets. This meant that, for
4317 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4318 possible to check for precisely this case. All we can do is abandon the
4319 optimization if .* occurs inside capturing brackets when there are any back
4320 references whatsoever. (See below for a better fix that came later.)
4321
4322 26. The handling of the optimization for finding the first character of a
4323 non-anchored pattern, and for finding a character that is required later in the
4324 match were failing in some cases. This didn't break the matching; it just
4325 failed to optimize when it could. The way this is done has been re-implemented.
4326
4327 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4328
4329 28. Added a new feature that provides some of the functionality that Perl
4330 provides with (?{...}). The facility is termed a "callout". The way it is done
4331 in PCRE is for the caller to provide an optional function, by setting
4332 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4333 global variable. By default it is unset, which disables all calling out. To get
4334 the function called, the regex must include (?C) at appropriate points. This
4335 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4336 This provides a means of identifying different callout points. When PCRE
4337 reaches such a point in the regex, if pcre_callout has been set, the external
4338 function is called. It is provided with data in a structure called
4339 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4340 matching continues; if it returns a non-zero value, the match at the current
4341 point fails. However, backtracking will occur if possible. [This was changed
4342 later and other features added - see item 49 below.]
4343
4344 29. pcretest is upgraded to test the callout functionality. It provides a
4345 callout function that displays information. By default, it shows the start of
4346 the match and the current position in the text. There are some new data escapes
4347 to vary what happens:
4348
4349 \C+ in addition, show current contents of captured substrings
4350 \C- do not supply a callout function
4351 \C!n return 1 when callout number n is reached
4352 \C!n!m return 1 when callout number n is reached for the mth time
4353
4354 30. If pcregrep was called with the -l option and just a single file name, it
4355 output "<stdin>" if a match was found, instead of the file name.
4356
4357 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4358 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4359 pcre_exec(). This saves a malloc/free per call. The default value of
4360 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4361 when configuring.
4362
4363 32. The default maximum size of a compiled pattern is 64K. There have been a
4364 few cases of people hitting this limit. The code now uses macros to handle the
4365 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4366 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4367 configuring. Tests 2 and 5 work only with 2-byte links because they output
4368 debugging information about compiled patterns.
4369
4370 33. Internal code re-arrangements:
4371
4372 (a) Moved the debugging function for printing out a compiled regex into
4373 its own source file (printint.c) and used #include to pull it into
4374 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4375 separate copies.
4376
4377 (b) Defined the list of op-code names for debugging as a macro in
4378 internal.h so that it is next to the definition of the opcodes.
4379
4380 (c) Defined a table of op-code lengths for simpler skipping along compiled
4381 code. This is again a macro in internal.h so that it is next to the
4382 definition of the opcodes.
4383
4384 34. Added support for recursive calls to individual subpatterns, along the
4385 lines of Robin Houston's patch (but implemented somewhat differently).
4386
4387 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4388 allow it to read and process whole directories in Win32. This code was
4389 contributed by Lionel Fourquaux; it has not been tested by me.
4390
4391 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4392 used to name a group. Names consist of alphanumerics and underscores, and must
4393 be unique. Back references use the syntax (?P=name) and recursive calls use
4394 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4395 numbers. The function pcre_fullinfo() can be used after compilation to extract
4396 a name/number map. There are three relevant calls:
4397
4398 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4399 PCRE_INFO_NAMECOUNT yields the number of entries
4400 PCRE_INFO_NAMETABLE yields a pointer to the map.
4401
4402 The map is a vector of fixed-size entries. The size of each entry depends on
4403 the length of the longest name used. The first two bytes of each entry are the
4404 group number, most significant byte first. There follows the corresponding
4405 name, zero terminated. The names are in alphabetical order.
4406
4407 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4408 case instead of 255. Making it the same both with and without UTF-8 support
4409 means that the same test output works with both.
4410
4411 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4412 calling malloc() with a zero argument.
4413
4414 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4415 optimization. I've improved things by keeping a bitmap of backreferences with
4416 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4417 fact referenced, the optimization can be applied. It is unlikely that a
4418 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4419 the match to follow \n) will appear inside brackets with a number greater than
4420 31, but if it does, any back reference > 31 suppresses the optimization.
4421
4422 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4423 of disabling numbered capturing parentheses. Any opening parenthesis that is
4424 not followed by ? behaves as if it were followed by ?: but named parentheses
4425 can still be used for capturing (and they will acquire numbers in the usual
4426 way).
4427
4428 41. Redesigned the return codes from the match() function into yes/no/error so
4429 that errors can be passed back from deep inside the nested calls. A malloc
4430 failure while inside a recursive subpattern call now causes the
4431 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4432
4433 42. It is now possible to set a limit on the number of times the match()
4434 function is called in a call to pcre_exec(). This facility makes it possible to
4435 limit the amount of recursion and backtracking, though not in a directly
4436 obvious way, because the match() function is used in a number of different
4437 circumstances. The count starts from zero for each position in the subject
4438 string (for non-anchored patterns). The default limit is, for compatibility, a
4439 large number, namely 10 000 000. You can change this in two ways:
4440
4441 (a) When configuring PCRE before making, you can use --with-match-limit=n
4442 to set a default value for the compiled library.
4443
4444 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4445 a different value is set. See 45 below.
4446
4447 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4448
4449 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4450 of things that can be changed at compile time. The first argument specifies
4451 what is wanted and the second points to where the information is to be placed.
4452 The current list of available information is:
4453
4454 PCRE_CONFIG_UTF8
4455
4456 The output is an integer that is set to one if UTF-8 support is available;
4457 otherwise it is set to zero.
4458
4459 PCRE_CONFIG_NEWLINE
4460
4461 The output is an integer that it set to the value of the code that is used for
4462 newline. It is either LF (10) or CR (13).
4463
4464 PCRE_CONFIG_LINK_SIZE
4465
4466 The output is an integer that contains the number of bytes used for internal
4467 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4468
4469 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4470
4471 The output is an integer that contains the threshold above which the POSIX
4472 interface uses malloc() for output vectors. See item 31 above.
4473
4474 PCRE_CONFIG_MATCH_LIMIT
4475
4476 The output is an unsigned integer that contains the default limit of the number
4477 of match() calls in a pcre_exec() execution. See 42 above.
4478
4479 44. pcretest has been upgraded by the addition of the -C option. This causes it
4480 to extract all the available output from the new pcre_config() function, and to
4481 output it. The program then exits immediately.
4482
4483 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4484 order to support additional features. One way would have been to define
4485 pcre_exec2() (for example) with extra arguments, but this would not have been
4486 extensible, and would also have required all calls to the original function to
4487 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4488 is used for passing in "extra" data from pcre_study().
4489
4490 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4491 contains the following fields:
4492
4493 flags a bitmap indicating which of the following fields are set
4494 study_data opaque data from pcre_study()
4495 match_limit a way of specifying a limit on match() calls for a specific
4496 call to pcre_exec()
4497 callout_data data for callouts (see 49 below)
4498
4499 The flag bits are also defined in pcre.h, and are
4500
4501 PCRE_EXTRA_STUDY_DATA
4502 PCRE_EXTRA_MATCH_LIMIT
4503 PCRE_EXTRA_CALLOUT_DATA
4504
4505 The pcre_study() function now returns one of these new pcre_extra blocks, with
4506 the actual study data pointed to by the study_data field, and the
4507 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4508 before. That is, this change is entirely upwards-compatible and requires no
4509 change to existing code.
4510
4511 If you want to pass in additional data to pcre_exec(), you can either place it
4512 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4513 block.
4514
4515 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4516 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4517 times with different match limits, until it finds the minimum value needed for
4518 pcre_exec() to complete. The value is then output. This can be instructive; for
4519 most simple matches the number is quite small, but for pathological cases it
4520 gets very large very quickly.
4521
4522 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4523 returns the size of the data block pointed to by the study_data field in a
4524 pcre_extra block, that is, the value that was passed as the argument to
4525 pcre_malloc() when PCRE was getting memory in which to place the information
4526 created by pcre_study(). The fourth argument should point to a size_t variable.
4527 pcretest has been extended so that this information is shown after a successful
4528 pcre_study() call when information about the compiled regex is being displayed.
4529
4530 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4531 because what follows is always an absolute path. (Later: it turns out that this
4532 is more than cosmetic for MinGW, because it doesn't like empty path
4533 components.)
4534
4535 49. Some changes have been made to the callout feature (see 28 above):
4536
4537 (i) A callout function now has three choices for what it returns:
4538
4539 0 => success, carry on matching
4540 > 0 => failure at this point, but backtrack if possible
4541 < 0 => serious error, return this value from pcre_exec()
4542
4543 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4544 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4545 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4546 use by callout functions. It will never be used by PCRE itself.
4547
4548 (ii) The pcre_extra structure (see 45 above) has a void * field called
4549 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4550 pcre_callout_block structure has a field of the same name. The contents of
4551 the field passed in the pcre_extra structure are passed to the callout
4552 function in the corresponding field in the callout block. This makes it
4553 easier to use the same callout-containing regex from multiple threads. For
4554 testing, the pcretest program has a new data escape
4555
4556 \C*n pass the number n (may be negative) as callout_data
4557
4558 If the callout function in pcretest receives a non-zero value as
4559 callout_data, it returns that value.
4560
4561 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4562 there were some redundant $(CFLAGS) in commands that are now specified as
4563 $(LINK), which already includes $(CFLAGS).
4564
4565 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4566 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4567 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4568 one-byte characters throughout. Note that case-insensitive matching applies
4569 only to characters whose values are less than 256. PCRE doesn't support the
4570 notion of cases for higher-valued characters.
4571
4572 (i) A character class whose characters are all within 0-255 is handled as
4573 a bit map, and the map is inverted for negative classes. Previously, a
4574 character > 255 always failed to match such a class; however it should
4575 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4576
4577 (ii) A negated character class with a single character < 255 is coded as
4578 "not this character" (OP_NOT). This wasn't working properly when the test
4579 character was multibyte, either singly or repeated.
4580
4581 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4582 mode, for example: \x{100}{2,3}.
4583
4584 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4585 singly or repeated) now correctly test multibyte characters. However,
4586 PCRE doesn't recognize any characters with values greater than 255 as
4587 digits, spaces, or word characters. Such characters always match \D, \S,
4588 and \W, and never match \d, \s, or \w.
4589
4590 (v) Classes may now contain characters and character ranges with values
4591 greater than 255. For example: [ab\x{100}-\x{400}].
4592
4593 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4594 PCRE in UTF-8 mode.
4595
4596 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4597 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4598 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4599 value.)
4600
4601 53. The single man page has become too large. I have therefore split it up into
4602 a number of separate man pages. These also give rise to individual HTML pages;
4603 these are now put in a separate directory, and there is an index.html page that
4604 lists them all. Some hyperlinking between the pages has been installed.
4605
4606 54. Added convenience functions for handling named capturing parentheses.
4607
4608 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4609 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4610 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4611 are faulted.
4612
4613 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4614 calling configure. These values are used when compiling the dftables.c program
4615 which is run to generate the source of the default character tables. They
4616 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4617 you will need to set these values.
4618
4619 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4620
4621
4622 Version 3.9 02-Jan-02
4623 ---------------------
4624
4625 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4626
4627 2. If --disable-static was given, the building process failed when trying to
4628 build pcretest and pcregrep. (For some reason it was using libtool to compile
4629 them, which is not right, as they aren't part of the library.)
4630
4631
4632 Version 3.8 18-Dec-01
4633 ---------------------
4634
4635 1. The experimental UTF-8 code was completely screwed up. It was packing the
4636 bytes in the wrong order. How dumb can you get?
4637
4638
4639 Version 3.7 29-Oct-01
4640 ---------------------
4641
4642 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4643 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4644 this didn't happen under Solaris 8, where I normally test things.
4645
4646 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4647 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4648 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4649 this fix an hour or so after the initial 3.7 release.)
4650
4651
4652 Version 3.6 23-Oct-01
4653 ---------------------
4654
4655 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4656 offsets passed as NULL with zero offset count.
4657
4658 2. The config.guess and config.sub files had not been updated when I moved to
4659 the latest autoconf.
4660
4661
4662 Version 3.5 15-Aug-01
4663 ---------------------
4664
4665 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4666 had been forgotten.
4667
4668 2. By using declared but undefined structures, we can avoid using "void"
4669 definitions in pcre.h while keeping the internal definitions of the structures
4670 private.
4671
4672 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4673 user point of view, this means that both static and shared libraries are built
4674 by default, but this can be individually controlled. More of the work of
4675 handling this static/shared cases is now inside libtool instead of PCRE's make
4676 file.
4677
4678 4. The pcretest utility is now installed along with pcregrep because it is
4679 useful for users (to test regexs) and by doing this, it automatically gets
4680 relinked by libtool. The documentation has been turned into a man page, so
4681 there are now .1, .txt, and .html versions in /doc.
4682
4683 5. Upgrades to pcregrep:
4684 (i) Added long-form option names like gnu grep.
4685 (ii) Added --help to list all options with an explanatory phrase.
4686 (iii) Added -r, --recursive to recurse into sub-directories.
4687 (iv) Added -f, --file to read patterns from a file.
4688
4689 6. pcre_exec() was referring to its "code" argument before testing that
4690 argument for NULL (and giving an error if it was NULL).
4691
4692 7. Upgraded Makefile.in to allow for compiling in a different directory from
4693 the source directory.
4694
4695 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4696 options bits, the pointer it was passed was to an int instead of to an unsigned
4697 long int. This mattered only on 64-bit systems.
4698
4699 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4700 generated) instead of pcre.in, which it its source. Also made the same change
4701 in several of the .c files.
4702
4703 10. A new release of gcc defines printf() as a macro, which broke pcretest
4704 because it had an ifdef in the middle of a string argument for printf(). Fixed
4705 by using separate calls to printf().
4706
4707 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4708 script, to force use of CR or LF instead of \n in the source. On non-Unix
4709 systems, the value can be set in config.h.
4710
4711 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4712 absolute limit. Changed the text of the error message to make this clear, and
4713 likewise updated the man page.
4714
4715 13. The limit of 99 on the number of capturing subpatterns has been removed.
4716 The new limit is 65535, which I hope will not be a "real" limit.
4717
4718
4719 Version 3.4 22-Aug-00
4720 ---------------------
4721
4722 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4723
4724 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4725
4726
4727 Version 3.3 01-Aug-00
4728 ---------------------
4729
4730 1. If an octal character was given, but the value was greater than \377, it
4731 was not getting masked to the least significant bits, as documented. This could
4732 lead to crashes in some systems.
4733
4734 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4735 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4736
4737 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4738 These just pass their arguments on to (pcre_free)(), but they are provided
4739 because some uses of PCRE bind it to non-C systems that can call its functions,
4740 but cannot call free() or pcre_free() directly.
4741
4742 4. Add "make test" as a synonym for "make check". Corrected some comments in
4743 the Makefile.
4744
4745 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4746 Makefile.
4747
4748 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4749 command called pgrep for grepping around the active processes.
4750
4751 7. Added the beginnings of support for UTF-8 character strings.
4752
4753 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4754 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4755 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4756 out for the ar command.)
4757
4758
4759 Version 3.2 12-May-00
4760 ---------------------
4761
4762 This is purely a bug fixing release.
4763
4764 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4765 of ZA. This was just one example of several cases that could provoke this bug,
4766 which was introduced by change 9 of version 2.00. The code for breaking
4767 infinite loops after an iteration that matches an empty string was't working
4768 correctly.
4769
4770 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4771 when matched against abbab (for example). After matching an empty string, it
4772 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4773 caused it to match further down the string than it should.
4774
4775 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4776 was there because it doesn't seem to be needed, and it causes trouble on some
4777 systems, as it is not a Standard C header. It has been removed.
4778
4779 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4780 were reported on the Macintosh. The changes were from
4781
4782 while ((c = *(++ptr)) != 0 && c != '\n');
4783 to
4784 while ((c = *(++ptr)) != 0 && c != '\n') ;
4785
4786 Totally extraordinary, but if that's what it takes...
4787
4788 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4789 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4790 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4791 assumes the way PCRE uses memmove() (always moving upwards).
4792
4793 6. PCRE is being used in one environment where strchr() is not available. There
4794 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4795 faster code anyway.
4796
4797
4798 Version 3.1 09-Feb-00
4799 ---------------------
4800
4801 The only change in this release is the fixing of some bugs in Makefile.in for
4802 the "install" target:
4803
4804 (1) It was failing to install pcreposix.h.
4805
4806 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4807
4808
4809 Version 3.0 01-Feb-00
4810 ---------------------
4811
4812 1. Add support for the /+ modifier to perltest (to output $` like it does in
4813 pcretest).
4814
4815 2. Add support for the /g modifier to perltest.
4816
4817 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4818 matches null strings.
4819
4820 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4821 pattern. Perl treats empty patterns specially - it reuses the most recent
4822 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4823 effect.
4824
4825 5. The POSIX interface was broken in that it was just handing over the POSIX
4826 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4827 required a bigger vector, with some working space on the end. This means that
4828 the POSIX wrapper now has to get and free some memory, and copy the results.
4829
4830 6. Added some simple autoconf support, placing the test data and the
4831 documentation in separate directories, re-organizing some of the
4832 information files, and making it build pcre-config (a GNU standard). Also added
4833 libtool support for building PCRE as a shared library, which is now the
4834 default.
4835
4836 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4837 09 are not valid octal constants. Single digits will be used for minor values
4838 less than 10.
4839
4840 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4841 existing programs that set these in the POSIX interface can use PCRE without
4842 modification.
4843
4844 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4845 return all that pcre_info() returns, plus additional data. The pcre_info()
4846 function is retained for compatibility, but is considered to be obsolete.
4847
4848 10. Added experimental recursion feature (?R) to handle one common case that
4849 Perl 5.6 will be able to do with (?p{...}).
4850
4851 11. Added support for POSIX character classes like [:alpha:], which Perl is
4852 adopting.
4853
4854
4855 Version 2.08 31-Aug-99
4856 ----------------------
4857
4858 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4859 trying to match at the startoffset position, but instead was moving forward to
4860 the next newline as if a previous match had failed.
4861
4862 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4863 and could get into a loop if a null string was matched other than at the start
4864 of the subject.
4865
4866 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4867 be distinguished at compile time, and for completeness also added PCRE_DATE.
4868
4869 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4870 in GnuWin32 environments.
4871
4872
4873 Version 2.07 29-Jul-99
4874 ----------------------
4875
4876 1. The documentation is now supplied in plain text form and HTML as well as in
4877 the form of man page sources.
4878
4879 2. C++ compilers don't like assigning (void *) values to other pointer types.
4880 In particular this affects malloc(). Although there is no problem in Standard
4881 C, I've put in casts to keep C++ compilers happy.
4882
4883 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4884 should be (const char *).
4885
4886 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4887 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4888 However, I haven't made this a standard facility. The documentation doesn't
4889 mention it, and the Makefile doesn't support it.
4890
4891 5. The Makefile now contains an "install" target, with editable destinations at
4892 the top of the file. The pcretest program is not installed.
4893
4894 6. pgrep -V now gives the PCRE version number and date.
4895
4896 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4897 causing the entire string to be ignored, instead of just the last character.
4898
4899 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4900 non-matching string, it can take a very, very long time, even for strings of
4901 quite modest length, because of the nested recursion. PCRE now does better in
4902 some of these cases. It does this by remembering the last required literal
4903 character in the pattern, and pre-searching the subject to ensure it is present
4904 before running the real match. In other words, it applies a heuristic to detect
4905 some types of certain failure quickly, and in the above example, if presented
4906 with a string that has no trailing " it gives "no match" very quickly.
4907
4908 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4909 other alternatives are tried instead.
4910
4911
4912 Version 2.06 09-Jun-99
4913 ----------------------
4914
4915 1. Change pcretest's output for amount of store used to show just the code
4916 space, because the remainder (the data block) varies in size between 32-bit and
4917 64-bit systems.
4918
4919 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4920 start matching at. This allows lookbehinds to work when searching for multiple
4921 occurrences in a string.
4922
4923 3. Added additional options to pcretest for testing multiple occurrences:
4924
4925 /+ outputs the rest of the string that follows a match
4926 /g loops for multiple occurrences, using the new startoffset argument
4927 /G loops for multiple occurrences by passing an incremented pointer
4928
4929 4. PCRE wasn't doing the "first character" optimization for patterns starting
4930 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4931 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4932 the letter 'x'. On long subject strings, this gives a significant speed-up.
4933
4934
4935 Version 2.05 21-Apr-99
4936 ----------------------
4937
4938 1. Changed the type of magic_number from int to long int so that it works
4939 properly on 16-bit systems.
4940
4941 2. Fixed a bug which caused patterns starting with .* not to work correctly
4942 when the subject string contained newline characters. PCRE was assuming
4943 anchoring for such patterns in all cases, which is not correct because .* will
4944 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4945 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4946 must be retried after every newline in the subject.
4947
4948
4949 Version 2.04 18-Feb-99
4950 ----------------------
4951
4952 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4953 computation of the store needed to hold the pattern was incorrect (too large).
4954 If such patterns were nested a few deep, this could multiply and become a real
4955 problem.
4956
4957 2. Added /M option to pcretest to show the memory requirement of a specific
4958 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4959
4960 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4961 compiled in such a way that the backtracking after subsequent failure was
4962 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4963 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4964
4965
4966 Version 2.03 02-Feb-99
4967 ----------------------
4968
4969 1. Fixed typo and small mistake in man page.
4970
4971 2. Added 4th condition (GPL supersedes if conflict) and created separate
4972 LICENCE file containing the conditions.
4973
4974 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4975 Perl, that is the internal \ allows the delimiter to be included in the
4976 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4977 the final delimiter, add \ to the end of the pattern (to test the error).
4978
4979 4. Added the convenience functions for extracting substrings after a successful
4980 match. Updated pcretest to make it able to test these functions.
4981
4982
4983 Version 2.02 14-Jan-99
4984 ----------------------
4985
4986 1. Initialized the working variables associated with each extraction so that
4987 their saving and restoring doesn't refer to uninitialized store.
4988
4989 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4990 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4991 fix the problem.
4992
4993 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4994 calls, and wasn't printing the correct value for compiling calls. Increased the
4995 default value of LOOPREPEAT, and the number of significant figures in the
4996 times.
4997
4998 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4999
5000 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5001 a building problem on Windows NT with a FAT file system.
5002
5003
5004 Version 2.01 21-Oct-98
5005 ----------------------
5006
5007 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5008 to character tables built by pcre_maketables() in the current locale. If NULL
5009 is passed, the default tables are used.
5010
5011
5012 Version 2.00 24-Sep-98
5013 ----------------------
5014
5015 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5016 it any more.
5017
5018 2. Allow quantification of (?>) groups, and make it work correctly.
5019
5020 3. The first character computation wasn't working for (?>) groups.
5021
5022 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5023 end of the subject) and add 5.005's \z, which really does match only at the
5024 very end of the subject.
5025
5026 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5027
5028 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5029 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5030 localized options. All options to pcre_study() were also removed.
5031
5032 7. Add other new features from 5.005:
5033
5034 $(?<= positive lookbehind
5035 $(?<! negative lookbehind
5036 (?imsx-imsx) added the unsetting capability
5037 such a setting is global if at outer level; local otherwise
5038 (?imsx-imsx:) non-capturing groups with option setting
5039 (?(cond)re|re) conditional pattern matching
5040
5041 A backreference to itself in a repeated group matches the previous
5042 captured string.
5043
5044 8. General tidying up of studying (both automatic and via "study")
5045 consequential on the addition of new assertions.
5046
5047 9. As in 5.005, unlimited repeated groups that could match an empty substring
5048 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5049 runtime if any iteration does actually match an empty substring.
5050
5051 10. Include the RunTest script in the distribution.
5052
5053 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5054 discrepancies, some of which were old and were also with respect to 5.004. They
5055 have now been fixed.
5056
5057
5058 Version 1.09 28-Apr-98
5059 ----------------------
5060
5061 1. A negated single character class followed by a quantifier with a minimum
5062 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5063 program crashes, or just wrong answers. This did not apply to negated classes
5064 containing more than one character, or to minima other than one.
5065
5066
5067 Version 1.08 27-Mar-98
5068 ----------------------
5069
5070 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5071
5072 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5073 latter must appear before anything that relies on it in the pattern.
5074
5075
5076 Version 1.07 16-Feb-98
5077 ----------------------
5078
5079 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5080 repeat of a potentially empty string).
5081
5082
5083 Version 1.06 23-Jan-98
5084 ----------------------
5085
5086 1. Added Markus Oberhumer's little patches for C++.
5087
5088 2. Literal strings longer than 255 characters were broken.
5089
5090
5091 Version 1.05 23-Dec-97
5092 ----------------------
5093
5094 1. Negated character classes containing more than one character were failing if
5095 PCRE_CASELESS was set at run time.
5096
5097
5098 Version 1.04 19-Dec-97
5099 ----------------------
5100
5101 1. Corrected the man page, where some "const" qualifiers had been omitted.
5102
5103 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5104 input syntax.
5105
5106 3. Fixed memory leak which occurred when a regex with back references was
5107 matched with an offsets vector that wasn't big enough. The temporary memory
5108 that is used in this case wasn't being freed if the match failed.
5109
5110 4. Tidied pcretest to ensure it frees memory that it gets.
5111
5112 5. Temporary memory was being obtained in the case where the passed offsets
5113 vector was exactly big enough.
5114
5115 6. Corrected definition of offsetof() from change 5 below.
5116
5117 7. I had screwed up change 6 below and broken the rules for the use of
5118 setjmp(). Now fixed.
5119
5120
5121 Version 1.03 18-Dec-97
5122 ----------------------
5123
5124 1. A erroneous regex with a missing opening parenthesis was correctly
5125 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5126 on some systems.
5127
5128 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5129 it was reported that one broken compiler failed on the former because "code" is
5130 also an independent variable.
5131
5132 3. The erroneous regex a[]b caused an array overrun reference.
5133
5134 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5135 fail on data ending with that character. (It was going on too far, and checking
5136 the next character, typically a binary zero.) This was specific to the
5137 optimized code for single-character negative classes.
5138
5139 5. Added a contributed patch from the TIN world which does the following:
5140
5141 + Add an undef for memmove, in case the the system defines a macro for it.
5142
5143 + Add a definition of offsetof(), in case there isn't one. (I don't know
5144 the reason behind this - offsetof() is part of the ANSI standard - but
5145 it does no harm).
5146
5147 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5148 most of the places where whitespace preceded '#'. I have given up and
5149 allowed the remaining 2 cases to be at the margin.
5150
5151 + Rename some variables in pcre to eliminate shadowing. This seems very
5152 pedantic, but does no harm, of course.
5153
5154 6. Moved the call to setjmp() into its own function, to get rid of warnings
5155 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5156
5157 7. Constructs such as \d{8,} were compiling into the equivalent of
5158 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5159 outcome, but in this particular case used more store than had been allocated,
5160 which caused the bug to be discovered because it threw up an internal error.
5161
5162 8. The debugging code in both pcre and pcretest for outputting the compiled
5163 form of a regex was going wrong in the case of back references followed by
5164 curly-bracketed repeats.
5165
5166
5167 Version 1.02 12-Dec-97
5168 ----------------------
5169
5170 1. Typos in pcre.3 and comments in the source fixed.
5171
5172 2. Applied a contributed patch to get rid of places where it used to remove
5173 'const' from variables, and fixed some signed/unsigned and uninitialized
5174 variable warnings.
5175
5176 3. Added the "runtest" target to Makefile.
5177
5178 4. Set default compiler flag to -O2 rather than just -O.
5179
5180
5181 Version 1.01 19-Nov-97
5182 ----------------------
5183
5184 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5185 like /([ab]*)*/, that is, for classes with more than one character in them.
5186
5187 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5188 as /((?>a*))*/ (a PCRE_EXTRA facility).
5189
5190
5191 Version 1.00 18-Nov-97
5192 ----------------------
5193
5194 1. Added compile-time macros to support systems such as SunOS4 which don't have
5195 memmove() or strerror() but have other things that can be used instead.
5196
5197 2. Arranged that "make clean" removes the executables.
5198
5199
5200 Version 0.99 27-Oct-97
5201 ----------------------
5202
5203 1. Fixed bug in code for optimizing classes with only one character. It was
5204 initializing a 32-byte map regardless, which could cause it to run off the end
5205 of the memory it had got.
5206
5207 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5208
5209
5210 Version 0.98 22-Oct-97
5211 ----------------------
5212
5213 1. Fixed bug in code for handling temporary memory usage when there are more
5214 back references than supplied space in the ovector. This could cause segfaults.
5215
5216
5217 Version 0.97 21-Oct-97
5218 ----------------------
5219
5220 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5221
5222 2. Optimized negated single characters not to use a bit map.
5223
5224 3. Brought error texts together as macro definitions; clarified some of them;
5225 fixed one that was wrong - it said "range out of order" when it meant "invalid
5226 escape sequence".
5227
5228 4. Changed some char * arguments to const char *.
5229
5230 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5231
5232 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5233 pcretest.
5234
5235
5236 Version 0.96 16-Oct-97
5237 ----------------------
5238
5239 1. Added a simple "pgrep" utility to the distribution.
5240
5241 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5242 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5243 where "ddd" means "one or more decimal digits".
5244
5245 3. Fixed serious bug. If a pattern had a back reference, but the call to
5246 pcre_exec() didn't supply a large enough ovector to record the related
5247 identifying subpattern, the match always failed. PCRE now remembers the number
5248 of the largest back reference, and gets some temporary memory in which to save
5249 the offsets during matching if necessary, in order to ensure that
5250 backreferences always work.
5251
5252 4. Increased the compatibility with Perl in a number of ways:
5253
5254 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5255 to request this handling. The option can be set at compile or exec time.
5256
5257 (b) $ matches before a terminating newline by default; an option
5258 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5259 mode). The option can be set at compile or exec time.
5260
5261 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5262 the same as Perl's. If the decimal number it represents is less than 10
5263 or there aren't that many previous left capturing parentheses, an octal
5264 escape is read. Inside a character class, it's always an octal escape,
5265 even if it is a single digit.
5266
5267 (d) An escaped but undefined alphabetic character is taken as a literal,
5268 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5269 escapes.
5270
5271 (e) {0} is now permitted. (The previous item is removed from the compiled
5272 pattern).
5273
5274 5. Changed all the names of code files so that the basic parts are no longer
5275 than 10 characters, and abolished the teeny "globals.c" file.
5276
5277 6. Changed the handling of character classes; they are now done with a 32-byte
5278 bit map always.
5279
5280 7. Added the -d and /D options to pcretest to make it possible to look at the
5281 internals of compilation without having to recompile pcre.
5282
5283
5284 Version 0.95 23-Sep-97
5285 ----------------------
5286
5287 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5288 \x20 at the start of a run of normal characters. These were being treated as
5289 real characters, instead of the source characters being re-checked.
5290
5291
5292 Version 0.94 18-Sep-97
5293 ----------------------
5294
5295 1. The functions are now thread-safe, with the caveat that the global variables
5296 containing pointers to malloc() and free() or alternative functions are the
5297 same for all threads.
5298
5299 2. Get pcre_study() to generate a bitmap of initial characters for non-
5300 anchored patterns when this is possible, and use it if passed to pcre_exec().
5301
5302
5303 Version 0.93 15-Sep-97
5304 ----------------------
5305
5306 1. /(b)|(:+)/ was computing an incorrect first character.
5307
5308 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5309 but not actually doing anything yet.
5310
5311 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5312 as Perl does (e.g. [-az] or [az-]).
5313
5314 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5315 all possible positions.
5316
5317 5. Split up into different modules to avoid including unneeded functions in a
5318 compiled binary. However, compile and exec are still in one module. The "study"
5319 function is split off.
5320
5321 6. The character tables are now in a separate module whose source is generated
5322 by an auxiliary program - but can then be edited by hand if required. There are
5323 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5324 toupper() in the code.
5325
5326 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5327 make them global. Abolish the function for setting them, as the caller can now
5328 set them directly.
5329
5330
5331 Version 0.92 11-Sep-97
5332 ----------------------
5333
5334 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5335 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5336
5337 2. Caseless matching was not working in character classes if the characters in
5338 the pattern were in upper case.
5339
5340 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5341
5342 4. Make PCRE_ANCHORED public and accept as a compile option.
5343
5344 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5345 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5346 pass them.
5347
5348 6. Give an error if bad option bits passed at compile or run time.
5349
5350 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5351 pcretest to cause it to pass that flag.
5352
5353 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5354 options, and the first character, if set.
5355
5356 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5357
5358
5359 Version 0.91 10-Sep-97
5360 ----------------------
5361
5362 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5363 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5364
5365 2. PCRE was looping on encountering an indefinitely repeated back reference to
5366 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5367 Perl does - treats the match as successful.
5368
5369 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5