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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1417 - (show annotations)
Tue Dec 24 18:03:06 2013 UTC (5 years, 10 months ago) by ph10
File size: 258518 byte(s)
Get rid of some unitialized variable compiler warnings.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.35-RC1 xx-xxxx-201x
5 -----------------------------
6
7 1. A new flag is set, when property checks are present in an XCLASS.
8 When this flag is not set, PCRE can perform certain optimizations
9 such as studying these XCLASS-es.
10
11 2. The auto-possessification of character sets were improved: a normal
12 and an extended character set can be compared now. Furthermore
13 the JIT compiler optimizes more character set checks.
14
15 3. Got rid of some compiler warnings for potentially uninitialized variables
16 that show up only when compiled with -O2.
17
18
19 Version 8.34 15-December-2013
20 -----------------------------
21
22 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
23 executable memory. Patch inspired by Carsten Klein.
24
25 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
26 this macro is never tested and has no effect, because the work to support
27 coverage involves only compiling and linking options and special targets in
28 the Makefile. The comment in config.h implied that defining the macro would
29 enable coverage support, which is totally false. There was also support for
30 setting this macro in the CMake files (my fault, I just copied it from
31 configure). SUPPORT_GCOV has now been removed.
32
33 3. Make a small performance improvement in strlen16() and strlen32() in
34 pcretest.
35
36 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
37 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
38 these cases). There was also one in pcretest.c.
39
40 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
41
42 6. In UTF mode, the code for checking whether a group could match an empty
43 string (which is used for indefinitely repeated groups to allow for
44 breaking an infinite loop) was broken when the group contained a repeated
45 negated single-character class with a character that occupied more than one
46 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
47 UTF-8 mode). The effect was undefined: the group might or might not be
48 deemed as matching an empty string, or the program might have crashed.
49
50 7. The code for checking whether a group could match an empty string was not
51 recognizing that \h, \H, \v, \V, and \R must match a character.
52
53 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
54 an empty string. If it can, pcretest shows this in its information output.
55
56 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
57 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
58 matched by pcre_exec() without using JIT:
59
60 (a) If the rest of the pattern did not match after a maximal run of
61 grapheme clusters, the code for backing up to try with fewer of them
62 did not always back up over a full grapheme when characters that do not
63 have the modifier quality were involved, e.g. Hangul syllables.
64
65 (b) If the match point in a subject started with modifier character, and
66 there was no match, the code could incorrectly back up beyond the match
67 point, and potentially beyond the first character in the subject,
68 leading to a segfault or an incorrect match result.
69
70 10. A conditional group with an assertion condition could lead to PCRE
71 recording an incorrect first data item for a match if no other first data
72 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
73 first data item, and therefore matched "ca" after "c" instead of at the
74 start.
75
76 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
77 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
78
79 12. The source of pcregrep now includes z/OS-specific code so that it can be
80 compiled for z/OS as part of the special z/OS distribution.
81
82 13. Added the -T and -TM options to pcretest.
83
84 14. The code in pcre_compile.c for creating the table of named capturing groups
85 has been refactored. Instead of creating the table dynamically during the
86 actual compiling pass, the information is remembered during the pre-compile
87 pass (on the stack unless there are more than 20 named groups, in which
88 case malloc() is used) and the whole table is created before the actual
89 compile happens. This has simplified the code (it is now nearly 150 lines
90 shorter) and prepared the way for better handling of references to groups
91 with duplicate names.
92
93 15. A back reference to a named subpattern when there is more than one of the
94 same name now checks them in the order in which they appear in the pattern.
95 The first one that is set is used for the reference. Previously only the
96 first one was inspected. This change makes PCRE more compatible with Perl.
97
98 16. Unicode character properties were updated from Unicode 6.3.0.
99
100 17. The compile-time code for auto-possessification has been refactored, based
101 on a patch by Zoltan Herczeg. It now happens after instead of during
102 compilation. The code is cleaner, and more cases are handled. The option
103 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
104 options in pcretest are provided to set it. It can also be set by
105 (*NO_AUTO_POSSESS) at the start of a pattern.
106
107 18. The character VT has been added to the default ("C" locale) set of
108 characters that match \s and are generally treated as white space,
109 following this same change in Perl 5.18. There is now no difference between
110 "Perl space" and "POSIX space". Whether VT is treated as white space in
111 other locales depends on the locale.
112
113 19. The code for checking named groups as conditions, either for being set or
114 for being recursed, has been refactored (this is related to 14 and 15
115 above). Processing unduplicated named groups should now be as fast at
116 numerical groups, and processing duplicated groups should be faster than
117 before.
118
119 20. Two patches to the CMake build system, by Alexander Barkov:
120
121 (1) Replace the "source" command by "." in CMakeLists.txt because
122 "source" is a bash-ism.
123
124 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
125 without these the CMake build does not work on Solaris.
126
127 21. Perl has changed its handling of \8 and \9. If there is no previously
128 encountered capturing group of those numbers, they are treated as the
129 literal characters 8 and 9 instead of a binary zero followed by the
130 literals. PCRE now does the same.
131
132 22. Following Perl, added \o{} to specify codepoints in octal, making it
133 possible to specify values greater than 0777 and also making them
134 unambiguous.
135
136 23. Perl now gives an error for missing closing braces after \x{... instead of
137 treating the string as literal. PCRE now does the same.
138
139 24. RunTest used to grumble if an inappropriate test was selected explicitly,
140 but just skip it when running all tests. This make it awkward to run ranges
141 of tests when one of them was inappropriate. Now it just skips any
142 inappropriate tests, as it always did when running all tests.
143
144 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
145 character types such as \d or \w, too many callouts were inserted, and the
146 data that they returned was rubbish.
147
148 26. In UCP mode, \s was not matching two of the characters that Perl matches,
149 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
150 were matched by \h. The code has now been refactored so that the lists of
151 the horizontal and vertical whitespace characters used for \h and \v (which
152 are defined only in one place) are now also used for \s.
153
154 27. Add JIT support for the 64 bit TileGX architecture.
155 Patch by Jiong Wang (Tilera Corporation).
156
157 28. Possessive quantifiers for classes (both explicit and automatically
158 generated) now use special opcodes instead of wrapping in ONCE brackets.
159
160 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
161 (because it's meaningless), this was not happening when PCRE_CASELESS was
162 set. Not wrong, but inefficient.
163
164 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
165 properties for \w, \d, etc) is present in a test regex. Otherwise if the
166 test contains no characters greater than 255, Perl doesn't realise it
167 should be using Unicode semantics.
168
169 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
170 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
171 does in Unicode mode.
172
173 32. Added the "forbid" facility to pcretest so that putting tests into the
174 wrong test files can sometimes be quickly detected.
175
176 33. There is now a limit (default 250) on the depth of nesting of parentheses.
177 This limit is imposed to control the amount of system stack used at compile
178 time. It can be changed at build time by --with-parens-nest-limit=xxx or
179 the equivalent in CMake.
180
181 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
182 errors. Perl warns for these when in warning mode, but PCRE has no facility
183 for giving warnings.
184
185 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
186 However, this was not working for (?!) because it is optimized to (*FAIL),
187 for which PCRE does not allow quantifiers. The optimization is now disabled
188 when a quantifier follows (?!). I can't see any use for this, but it makes
189 things uniform.
190
191 36. Perl no longer allows group names to start with digits, so I have made this
192 change also in PCRE. It simplifies the code a bit.
193
194 37. In extended mode, Perl ignores spaces before a + that indicates a
195 possessive quantifier. PCRE allowed a space before the quantifier, but not
196 before the possessive +. It now does.
197
198 38. The use of \K (reset reported match start) within a repeated possessive
199 group such as (a\Kb)*+ was not working.
200
201 40. Document that the same character tables must be used at compile time and
202 run time, and that the facility to pass tables to pcre_exec() and
203 pcre_dfa_exec() is for use only with saved/restored patterns.
204
205 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
206 features for Builds with MSVC:
207
208 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
209 builds). The .rc files can be used to set FileDescription and many other
210 attributes.
211
212 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
213 This allows higher-level build scripts which want .pdb files to avoid
214 hard-coding the exact files needed."
215
216 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
217 mean "start of word" and "end of word", respectively, as a transition aid.
218
219 43. A minimizing repeat of a class containing codepoints greater than 255 in
220 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
221 compiled to use the heap for recursion.
222
223 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
224 is configured.
225
226
227 Version 8.33 28-May-2013
228 ------------------------
229
230 1. Added 'U' to some constants that are compared to unsigned integers, to
231 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
232 variables that are added to signed variables, to ensure the result is
233 signed and can be negated.
234
235 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
236 CMake config files.
237
238 3. Revise the creation of config.h.generic so that all boolean macros are
239 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
240 overriding via -D on the command line possible.
241
242 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
243 to unsigned int is reported to make a quite noticeable speed difference in
244 a specific Windows environment. Testing on Linux did also appear to show
245 some benefit (and it is clearly not harmful). Also fixed the definition of
246 Xop which should be unsigned.
247
248 5. Related to (4), changing the definition of the intermediate variable cc
249 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
250 improvements.
251
252 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
253 unnecessary spaces.
254
255 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
256 and later.
257
258 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
259
260 9. Optimizing fast_forward_start_bits in JIT.
261
262 10. Adding support for callouts in JIT, and fixing some issues revealed
263 during this work. Namely:
264
265 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
266
267 (b) Minimum length was not checked before the matching is started.
268
269 11. The value of capture_last that is passed to callouts was incorrect in some
270 cases when there was a capture on one path that was subsequently abandoned
271 after a backtrack. Also, the capture_last value is now reset after a
272 recursion, since all captures are also reset in this case.
273
274 12. The interpreter no longer returns the "too many substrings" error in the
275 case when an overflowing capture is in a branch that is subsequently
276 abandoned after a backtrack.
277
278 13. In the pathological case when an offset vector of size 2 is used, pcretest
279 now prints out the matched string after a yield of 0 or 1.
280
281 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
282 Only supported by the JIT compiler at the moment.
283
284 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
285
286 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
287 offset of the starting point of the matching process, provided the offsets
288 vector is large enough.
289
290 17. The \A escape now records a lookbehind value of 1, though its execution
291 does not actually inspect the previous character. This is to ensure that,
292 in partial multi-segment matching, at least one character from the old
293 segment is retained when a new segment is processed. Otherwise, if there
294 are no lookbehinds in the pattern, \A might match incorrectly at the start
295 of a new segment.
296
297 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
298
299 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
300 gives some modest performance improvement in 8-bit mode.
301
302 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
303 be expressed in certain programming languages using Universal Character
304 Names.
305
306 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
307 which points out that "non characters" are not "characters that may not
308 appear in Unicode strings" but rather "characters that are reserved for
309 internal use and have only local meaning".
310
311 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
312 there was a conditional group that depended on an assertion, if the
313 assertion was false, the callout that immediately followed the alternation
314 in the condition was skipped when pcre_exec() was used for matching.
315
316 23. Allow an explicit callout to be inserted before an assertion that is the
317 condition for a conditional group, for compatibility with automatic
318 callouts, which always insert a callout at this point.
319
320 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
321 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
322
323 25. (*PRUNE) is now supported by the JIT compiler.
324
325 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
326
327 27. Fix the case where there are two or more SKIPs with arguments that may be
328 ignored.
329
330 28. (*SKIP) is now supported by the JIT compiler.
331
332 29. (*THEN) is now supported by the JIT compiler.
333
334 30. Update RunTest with additional test selector options.
335
336 31. The way PCRE handles backtracking verbs has been changed in two ways.
337
338 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
339 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
340 backtracking. In some cases this makes it more Perl-compatible, but Perl's
341 rather obscure rules do not always do the same thing.
342
343 (2) Previously, backtracking verbs were confined within assertions. This is
344 no longer the case for positive assertions, except for (*ACCEPT). Again,
345 this sometimes improves Perl compatibility, and sometimes does not.
346
347 32. A number of tests that were in test 2 because Perl did things differently
348 have been moved to test 1, because either Perl or PCRE has changed, and
349 these tests are now compatible.
350
351 32. Backtracking control verbs are now handled in the same way in JIT and
352 interpreter.
353
354 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
355 contained a forward subroutine reference caused a compile error.
356
357 34. Auto-detect and optimize limited repetitions in JIT.
358
359 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
360 blocking (*UTF) etc.
361
362 36. In the interpreter, maximizing pattern repetitions for characters and
363 character types now use tail recursion, which reduces stack usage.
364
365 37. The value of the max lookbehind was not correctly preserved if a compiled
366 and saved regex was reloaded on a host of different endianness.
367
368 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
369 of the compiled pattern block, expand the flags field from 16 to 32 bits
370 because it was almost full.
371
372 39. Try madvise first before posix_madvise.
373
374 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
375 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
376 original change locked out all matches of empty strings. This has been
377 changed so that one match of an empty string per line is recognized.
378 Subsequent searches on the same line (for colouring or for --only-matching,
379 for example) do not recognize empty strings.
380
381 41. Applied a user patch to fix a number of spelling mistakes in comments.
382
383 42. Data lines longer than 65536 caused pcretest to crash.
384
385 43. Clarified the data type for length and startoffset arguments for pcre_exec
386 and pcre_dfa_exec in the function-specific man pages, where they were
387 explicitly stated to be in bytes, never having been updated. I also added
388 some clarification to the pcreapi man page.
389
390 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
391 a segmentation fault.
392
393
394 Version 8.32 30-November-2012
395 -----------------------------
396
397 1. Improved JIT compiler optimizations for first character search and single
398 character iterators.
399
400 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
401 Patch by Daniel Richard G.
402
403 3. Single character iterator optimizations in the JIT compiler.
404
405 4. Improved JIT compiler optimizations for character ranges.
406
407 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
408 Reported by Giuseppe D'Angelo.
409
410 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
411 of a line, was being set incorrectly in cases where .* appeared inside
412 atomic brackets at the start of a pattern, or where there was a subsequent
413 *PRUNE or *SKIP.
414
415 7. Improved instruction cache flush for POWER/PowerPC.
416 Patch by Daniel Richard G.
417
418 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
419 grep:
420
421 (a) There is now no limit to the number of patterns to be matched.
422
423 (b) An error is given if a pattern is too long.
424
425 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
426 are now supported.
427
428 (d) --exclude-from and --include-from (multiple use) have been added.
429
430 (e) Exclusions and inclusions now apply to all files and directories, not
431 just to those obtained from scanning a directory recursively.
432
433 (f) Multiple uses of -f and --file-list are now supported.
434
435 (g) In a Windows environment, the default for -d has been changed from
436 "read" (the GNU grep default) to "skip", because otherwise the presence
437 of a directory in the file list provokes an error.
438
439 (h) The documentation has been revised and clarified in places.
440
441 9. Improve the matching speed of capturing brackets.
442
443 10. Changed the meaning of \X so that it now matches a Unicode extended
444 grapheme cluster.
445
446 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
447 out POSIX threads when JIT support is configured.
448
449 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
450
451 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
452 structure could go wrong in environments where size_t is not the same size
453 as int.
454
455 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
456 set.
457
458 15. The EBCDIC support had decayed; later updates to the code had included
459 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
460 general tidy up of EBCDIC-related issues, and the documentation was also
461 not quite right. There is now a test that can be run on ASCII systems to
462 check some of the EBCDIC-related things (but is it not a full test).
463
464 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
465 in a small tidy to the code.
466
467 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
468
469 18. If the --only-matching (-o) option in pcregrep is specified multiple
470 times, each one causes appropriate output. For example, -o1 -o2 outputs the
471 substrings matched by the 1st and 2nd capturing parentheses. A separating
472 string can be specified by --om-separator (default empty).
473
474 19. Improving the first n character searches.
475
476 20. Turn case lists for horizontal and vertical white space into macros so that
477 they are defined only once.
478
479 21. This set of changes together give more compatible Unicode case-folding
480 behaviour for characters that have more than one other case when UCP
481 support is available.
482
483 (a) The Unicode property table now has offsets into a new table of sets of
484 three or more characters that are case-equivalent. The MultiStage2.py
485 script that generates these tables (the pcre_ucd.c file) now scans
486 CaseFolding.txt instead of UnicodeData.txt for character case
487 information.
488
489 (b) The code for adding characters or ranges of characters to a character
490 class has been abstracted into a generalized function that also handles
491 case-independence. In UTF-mode with UCP support, this uses the new data
492 to handle characters with more than one other case.
493
494 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
495 whose other case is greater than 256 are now correctly matched
496 caselessly. Previously, the high codepoint matched the low one, but not
497 vice versa.
498
499 (d) The processing of \h, \H, \v, and \ in character classes now makes use
500 of the new class addition function, using character lists defined as
501 macros alongside the case definitions of 20 above.
502
503 (e) Caseless back references now work with characters that have more than
504 one other case.
505
506 (f) General caseless matching of characters with more than one other case
507 is supported.
508
509 22. Unicode character properties were updated from Unicode 6.2.0
510
511 23. Improved CMake support under Windows. Patch by Daniel Richard G.
512
513 24. Add support for 32-bit character strings, and UTF-32
514
515 25. Major JIT compiler update (code refactoring and bugfixing).
516 Experimental Sparc 32 support is added.
517
518 26. Applied a modified version of Daniel Richard G's patch to create
519 pcre.h.generic and config.h.generic by "make" instead of in the
520 PrepareRelease script.
521
522 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
523 pcre_compile.c when checking for a zero character.
524
525 28. Introducing a native interface for JIT. Through this interface, the compiled
526 machine code can be directly executed. The purpose of this interface is to
527 provide fast pattern matching, so several sanity checks are not performed.
528 However, feature tests are still performed. The new interface provides
529 1.4x speedup compared to the old one.
530
531 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
532 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
533 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
534
535 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
536 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
537 the "old" RFC 2279). Instead, it ended up passing a negative length to
538 pcre_exec().
539
540 31. Add support for GCC's visibility feature to hide internal functions.
541
542 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
543 "unknown -C option" after outputting 0 or 1.
544
545 33. There is now support for generating a code coverage report for the test
546 suite in environments where gcc is the compiler and lcov is installed. This
547 is mainly for the benefit of the developers.
548
549 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
550 unaddressable using valgrind annotations, allowing valgrind to detect
551 invalid memory accesses. This is mainly for the benefit of the developers.
552
553 25. (*UTF) can now be used to start a pattern in any of the three libraries.
554
555 26. Give configure error if --enable-cpp but no C++ compiler found.
556
557
558 Version 8.31 06-July-2012
559 -------------------------
560
561 1. Fixing a wrong JIT test case and some compiler warnings.
562
563 2. Removed a bashism from the RunTest script.
564
565 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
566 to unsigned type, result still unsigned" that was given by an MS compiler
567 on encountering the code "-sizeof(xxx)".
568
569 4. Partial matching support is added to the JIT compiler.
570
571 5. Fixed several bugs concerned with partial matching of items that consist
572 of more than one character:
573
574 (a) /^(..)\1/ did not partially match "aba" because checking references was
575 done on an "all or nothing" basis. This also applied to repeated
576 references.
577
578 (b) \R did not give a hard partial match if \r was found at the end of the
579 subject.
580
581 (c) \X did not give a hard partial match after matching one or more
582 characters at the end of the subject.
583
584 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
585 a partial match for the string "\r".
586
587 (e) When newline was set to CRLF, the metacharacter "." did not recognize
588 a partial match for a CR character at the end of the subject string.
589
590 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
591 running pcretest, the text "(JIT)" added to the output whenever JIT is
592 actually used to run the match.
593
594 7. Individual JIT compile options can be set in pcretest by following -s+[+]
595 or /S+[+] with a digit between 1 and 7.
596
597 8. OP_NOT now supports any UTF character not just single-byte ones.
598
599 9. (*MARK) control verb is now supported by the JIT compiler.
600
601 10. The command "./RunTest list" lists the available tests without actually
602 running any of them. (Because I keep forgetting what they all are.)
603
604 11. Add PCRE_INFO_MAXLOOKBEHIND.
605
606 12. Applied a (slightly modified) user-supplied patch that improves performance
607 when the heap is used for recursion (compiled with --disable-stack-for-
608 recursion). Instead of malloc and free for each heap frame each time a
609 logical recursion happens, frames are retained on a chain and re-used where
610 possible. This sometimes gives as much as 30% improvement.
611
612 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
613 call.
614
615 14. As documented, (*COMMIT) is now confined to within a positive assertion.
616
617 15. It is now possible to link pcretest with libedit as an alternative to
618 libreadline.
619
620 16. (*COMMIT) control verb is now supported by the JIT compiler.
621
622 17. The Unicode data tables have been updated to Unicode 6.1.0.
623
624 18. Added --file-list option to pcregrep.
625
626 19. Added binary file support to pcregrep, including the -a, --binary-files,
627 -I, and --text options.
628
629 20. The madvise function is renamed for posix_madvise for QNX compatibility
630 reasons. Fixed by Giuseppe D'Angelo.
631
632 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
633
634 22. Changed the option for creating symbolic links for 16-bit man pages from
635 -s to -sf so that re-installing does not cause issues.
636
637 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
638
639 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
640 matches in certain environments (the workspace was not being correctly
641 retained). Also added to pcre_dfa_exec() a simple plausibility check on
642 some of the workspace data at the beginning of a restart.
643
644 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
645 was not doing so when it should - probably a typo introduced by SVN 528
646 (change 8.10/14).
647
648 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
649 \w+ when the character tables indicated that \x{c4} was a word character.
650 There were several related cases, all because the tests for doing a table
651 lookup were testing for characters less than 127 instead of 255.
652
653 27. If a pattern contains capturing parentheses that are not used in a match,
654 their slots in the ovector are set to -1. For those that are higher than
655 any matched groups, this happens at the end of processing. In the case when
656 there were back references that the ovector was too small to contain
657 (causing temporary malloc'd memory to be used during matching), and the
658 highest capturing number was not used, memory off the end of the ovector
659 was incorrectly being set to -1. (It was using the size of the temporary
660 memory instead of the true size.)
661
662 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
663 ovector size, it uses memory at the end of the block that it has got.
664
665 29. Check for an overlong MARK name and give an error at compile time. The
666 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
667
668 30. JIT compiler update.
669
670 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
671 Rill for the patch.
672
673 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
674
675 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
676
677 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
678 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
679 was enabled.
680
681 35. Improve JIT code generation for greedy plus quantifier.
682
683 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
684 1 to "aa" instead of to an empty string. The bug affected repeated groups
685 that could potentially match an empty string.
686
687 37. Optimizing single character iterators in JIT.
688
689 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
690 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
691 codepoints that are too big for the mode are faulted, and in a UTF mode,
692 disallowed codepoints are also faulted.
693
694 39. If PCRE was compiled with UTF support, in three places in the DFA
695 matcher there was code that should only have been obeyed in UTF mode, but
696 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
697 processing when bytes with values greater than 127 were present. In 16-bit
698 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
699 both cases the values are those that cannot be the first data item in a UTF
700 character. The three items that might have provoked this were recursions,
701 possessively repeated groups, and atomic groups.
702
703 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
704 and pcregrep, because some OS require shared objects to be explicitly
705 passed to ld, causing the link step to fail if they are not.
706
707 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
708 mode, patterns that started with \h* or \R* might be incorrectly matched.
709
710
711 Version 8.30 04-February-2012
712 -----------------------------
713
714 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
715 name is defined in ctype.h.
716
717 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
718 only in quite long subpatterns.
719
720 3. Removed the function pcre_info(), which has been obsolete and deprecated
721 since it was replaced by pcre_fullinfo() in February 2000.
722
723 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
724 match a (*MARK), and the match failed at the start of the subject, a
725 reference to memory before the start of the subject could occur. This bug
726 was introduced by fix 17 of release 8.21.
727
728 5. A reference to an unset group with zero minimum repetition was giving
729 totally wrong answers (in non-JavaScript-compatibility mode). For example,
730 /(another)?(\1?)test/ matched against "hello world test". This bug was
731 introduced in release 8.13.
732
733 6. Add support for 16-bit character strings (a large amount of work involving
734 many changes and refactorings).
735
736 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
737 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
738 from a file.
739
740 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
741 rounding is not applied in this particular case).
742
743 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
744 if they appear, or are escaped, in patterns.
745
746 10. Get rid of a number of -Wunused-but-set-variable warnings.
747
748 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
749 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
750 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
751 also returns the mark "x". This bug applied to capturing parentheses,
752 non-capturing parentheses, and atomic parentheses. It also applied to some
753 assertions.
754
755 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
756 information out of configure.ac instead of relying on pcre.h.generic, which
757 is not stored in the repository.
758
759 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
760 -lreadline.
761
762 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
763
764 15. Applied Graycode's patch to put the top-level frame on the stack rather
765 than the heap when not using the stack for recursion. This gives a
766 performance improvement in many cases when recursion is not deep.
767
768 16. Experimental code added to "pcretest -C" to output the stack frame size.
769
770
771 Version 8.21 12-Dec-2011
772 ------------------------
773
774 1. Updating the JIT compiler.
775
776 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
777 are added as well.
778
779 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
780 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
781 calling _pcre_jit_exec. Some extra comments are added.
782
783 4. (*MARK) settings inside atomic groups that do not contain any capturing
784 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
785 was introduced by change 18 for 8.20.
786
787 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
788 ECMA-262 standard.
789
790 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
791 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
792 This bug was probably introduced by change 9 of 8.13.
793
794 7. While fixing 6 above, I noticed that a number of other items were being
795 incorrectly rejected as "not fixed length". This arose partly because newer
796 opcodes had not been added to the fixed-length checking code. I have (a)
797 corrected the bug and added tests for these items, and (b) arranged for an
798 error to occur if an unknown opcode is encountered while checking for fixed
799 length instead of just assuming "not fixed length". The items that were
800 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
801 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
802 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
803
804 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
805 being incorrectly compiled and would have given unpredicatble results.
806
807 9. A possessively repeated subpattern with minimum repeat count greater than
808 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
809 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
810 the first (A) could occur when it should not.
811
812 10. Add a cast and remove a redundant test from the code.
813
814 11. JIT should use pcre_malloc/pcre_free for allocation.
815
816 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
817 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
818 is anything other than /usr, -L is still shown).
819
820 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
821
822 14. Perl does not support \N without a following name in a [] class; PCRE now
823 also gives an error.
824
825 15. If a forward reference was repeated with an upper limit of around 2000,
826 it caused the error "internal error: overran compiling workspace". The
827 maximum number of forward references (including repeats) was limited by the
828 internal workspace, and dependent on the LINK_SIZE. The code has been
829 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
830 the default depends on LINK_SIZE. There is a new upper limit (for safety)
831 of around 200,000 forward references. While doing this, I also speeded up
832 the filling in of repeated forward references.
833
834 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
835 incorrectly expecting the subject to contain another "a" after the start.
836
837 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
838 in the match, the SKIP should be ignored. This was not happening; instead
839 the SKIP was being treated as NOMATCH. For patterns such as
840 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
841 tested.
842
843 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
844 now much more compatible with Perl, in particular in cases where the result
845 is a non-match for a non-anchored pattern. For example, if
846 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
847 "m", where previously it did not return a name. A side effect of this
848 change is that for partial matches, the last encountered mark name is
849 returned, as for non matches. A number of tests that were previously not
850 Perl-compatible have been moved into the Perl-compatible test files. The
851 refactoring has had the pleasing side effect of removing one argument from
852 the match() function, thus reducing its stack requirements.
853
854 19. If the /S+ option was used in pcretest to study a pattern using JIT,
855 subsequent uses of /S (without +) incorrectly behaved like /S+.
856
857 21. Retrieve executable code size support for the JIT compiler and fixing
858 some warnings.
859
860 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
861 not work when the shorter character appeared right at the end of the
862 subject string.
863
864 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
865 systems.
866
867 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
868 output it when the /M option is used in pcretest.
869
870 25. The CheckMan script was not being included in the distribution. Also, added
871 an explicit "perl" to run Perl scripts from the PrepareRelease script
872 because this is reportedly needed in Windows.
873
874 26. If study data was being save in a file and studying had not found a set of
875 "starts with" bytes for the pattern, the data written to the file (though
876 never used) was taken from uninitialized memory and so caused valgrind to
877 complain.
878
879 27. Updated RunTest.bat as provided by Sheri Pierce.
880
881 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
882
883 29. Computation of memory usage for the table of capturing group names was
884 giving an unnecessarily large value.
885
886
887 Version 8.20 21-Oct-2011
888 ------------------------
889
890 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
891 a POSIX class. After further experiments with Perl, which convinced me that
892 Perl has bugs and confusions, a closing square bracket is no longer allowed
893 in a POSIX name. This bug also affected patterns with classes that started
894 with full stops.
895
896 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
897 captured substring, but while checking the failing first alternative,
898 substring 1 is temporarily captured. If the output vector supplied to
899 pcre_exec() was not big enough for this capture, the yield of the function
900 was still zero ("insufficient space for captured substrings"). This cannot
901 be totally fixed without adding another stack variable, which seems a lot
902 of expense for a edge case. However, I have improved the situation in cases
903 such as /(a)(b)x|abc/ matched against "abc", where the return code
904 indicates that fewer than the maximum number of slots in the ovector have
905 been set.
906
907 3. Related to (2) above: when there are more back references in a pattern than
908 slots in the output vector, pcre_exec() uses temporary memory during
909 matching, and copies in the captures as far as possible afterwards. It was
910 using the entire output vector, but this conflicts with the specification
911 that only 2/3 is used for passing back captured substrings. Now it uses
912 only the first 2/3, for compatibility. This is, of course, another edge
913 case.
914
915 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
916 main code base, and can be used by building with --enable-jit. When this is
917 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
918 runtime --no-jit option is given.
919
920 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
921 ovector, the return from the function was zero, implying that there were
922 other matches that did not fit. The correct "exactly full" value is now
923 returned.
924
925 6. If a subpattern that was called recursively or as a subroutine contained
926 (*PRUNE) or any other control that caused it to give a non-standard return,
927 invalid errors such as "Error -26 (nested recursion at the same subject
928 position)" or even infinite loops could occur.
929
930 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
931 computing the minimum length on reaching *ACCEPT, and so ended up with the
932 wrong value of 1 rather than 0. Further investigation indicates that
933 computing a minimum subject length in the presence of *ACCEPT is difficult
934 (think back references, subroutine calls), and so I have changed the code
935 so that no minimum is registered for a pattern that contains *ACCEPT.
936
937 8. If (*THEN) was present in the first (true) branch of a conditional group,
938 it was not handled as intended. [But see 16 below.]
939
940 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
941 Sheri Pierce.
942
943 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
944 the first byte in a match must be "a".
945
946 11. Change 17 for 8.13 increased the recursion depth for patterns like
947 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
948 pattern contains any instances of (*THEN). If it does not, the old
949 optimizations are restored. It would be nice to do this on a per-group
950 basis, but at the moment that is not feasible.
951
952 12. In some environments, the output of pcretest -C is CRLF terminated. This
953 broke RunTest's code that checks for the link size. A single white space
954 character after the value is now allowed for.
955
956 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
957 For "fr", it uses the Windows-specific input and output files.
958
959 14. If (*THEN) appeared in a group that was called recursively or as a
960 subroutine, it did not work as intended. [But see next item.]
961
962 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
963 pattern fragments (but not containing any | characters). If A and B are
964 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
965 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
966 D. In other words, Perl considers parentheses that do not contain any |
967 characters to be part of a surrounding alternative, whereas PCRE was
968 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
969 differently. PCRE now behaves in the same way as Perl, except in the case
970 of subroutine/recursion calls such as (?1) which have in any case always
971 been different (but PCRE had them first :-).
972
973 16. Related to 15 above: Perl does not treat the | in a conditional group as
974 creating alternatives. Such a group is treated in the same way as an
975 ordinary group without any | characters when processing (*THEN). PCRE has
976 been changed to match Perl's behaviour.
977
978 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
979 RunGrepTest script failed.
980
981 18. Change 22 for version 13 caused atomic groups to use more stack. This is
982 inevitable for groups that contain captures, but it can lead to a lot of
983 stack use in large patterns. The old behaviour has been restored for atomic
984 groups that do not contain any capturing parentheses.
985
986 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
987 suppress the check for a minimum subject length at run time. (If it was
988 given to pcre_exec() or pcre_dfa_exec() it did work.)
989
990 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
991 fail to work when decoding hex characters in data strings in EBCDIC
992 environments.
993
994 21. It appears that in at least one Mac OS environment, the isxdigit() function
995 is implemented as a macro that evaluates to its argument more than once,
996 contravening the C 90 Standard (I haven't checked a later standard). There
997 was an instance in pcretest which caused it to go wrong when processing
998 \x{...} escapes in subject strings. The has been rewritten to avoid using
999 things like p++ in the argument of isxdigit().
1000
1001
1002 Version 8.13 16-Aug-2011
1003 ------------------------
1004
1005 1. The Unicode data tables have been updated to Unicode 6.0.0.
1006
1007 2. Two minor typos in pcre_internal.h have been fixed.
1008
1009 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1010 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1011 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1012
1013 4. There were a number of related bugs in the code for matching backrefences
1014 caselessly in UTF-8 mode when codes for the characters concerned were
1015 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1016 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1017 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1018 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1019 2-byte code at the end of the subject (it thought there wasn't enough data
1020 left).
1021
1022 5. Comprehensive information about what went wrong is now returned by
1023 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1024 as the output vector has at least 2 elements. The offset of the start of
1025 the failing character and a reason code are placed in the vector.
1026
1027 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1028 now returned is for the first byte of the failing character, instead of the
1029 last byte inspected. This is an incompatible change, but I hope it is small
1030 enough not to be a problem. It makes the returned offset consistent with
1031 pcre_exec() and pcre_dfa_exec().
1032
1033 7. pcretest now gives a text phrase as well as the error number when
1034 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1035 failure, the offset and reason code are output.
1036
1037 8. When \R was used with a maximizing quantifier it failed to skip backwards
1038 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1039 back over a single character (\n). This seems wrong (because it treated the
1040 two characters as a single entity when going forwards), conflicts with the
1041 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1042 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1043 has been changed.
1044
1045 9. Some internal refactoring has changed the processing so that the handling
1046 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1047 time (the PCRE_DOTALL option was changed this way some time ago: version
1048 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1049 which was always a bit of a fudge. It also means that there is one less
1050 argument for the match() function, which reduces its stack requirements
1051 slightly. This change also fixes an incompatibility with Perl: the pattern
1052 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1053
1054 10. More internal refactoring has drastically reduced the number of recursive
1055 calls to match() for possessively repeated groups such as (abc)++ when
1056 using pcre_exec().
1057
1058 11. While implementing 10, a number of bugs in the handling of groups were
1059 discovered and fixed:
1060
1061 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1062 (a|)*(?1) gave a compile-time internal error.
1063 ((a|)+)+ did not notice that the outer group could match an empty string.
1064 (^a|^)+ was not marked as anchored.
1065 (.*a|.*)+ was not marked as matching at start or after a newline.
1066
1067 12. Yet more internal refactoring has removed another argument from the match()
1068 function. Special calls to this function are now indicated by setting a
1069 value in a variable in the "match data" data block.
1070
1071 13. Be more explicit in pcre_study() instead of relying on "default" for
1072 opcodes that mean there is no starting character; this means that when new
1073 ones are added and accidentally left out of pcre_study(), testing should
1074 pick them up.
1075
1076 14. The -s option of pcretest has been documented for ages as being an old
1077 synonym of -m (show memory usage). I have changed it to mean "force study
1078 for every regex", that is, assume /S for every regex. This is similar to -i
1079 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1080 using it. It makes it easier to run collections of tests with and without
1081 study enabled, and thereby test pcre_study() more easily. All the standard
1082 tests are now run with and without -s (but some patterns can be marked as
1083 "never study" - see 20 below).
1084
1085 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1086 restoration of the capturing data to the outer values was not happening
1087 correctly.
1088
1089 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1090 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1091 pattern had matched an empty string, and so incorrectly returned a no
1092 match.
1093
1094 17. There was optimizing code for the last branch of non-capturing parentheses,
1095 and also for the obeyed branch of a conditional subexpression, which used
1096 tail recursion to cut down on stack usage. Unfortunately, now that there is
1097 the possibility of (*THEN) occurring in these branches, tail recursion is
1098 no longer possible because the return has to be checked for (*THEN). These
1099 two optimizations have therefore been removed. [But see 8.20/11 above.]
1100
1101 18. If a pattern containing \R was studied, it was assumed that \R always
1102 matched two bytes, thus causing the minimum subject length to be
1103 incorrectly computed because \R can also match just one byte.
1104
1105 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1106 was incorrectly computed.
1107
1108 20. If /S is present twice on a test pattern in pcretest input, it now
1109 *disables* studying, thereby overriding the use of -s on the command line
1110 (see 14 above). This is necessary for one or two tests to keep the output
1111 identical in both cases.
1112
1113 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1114 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1115
1116 22. When an atomic group that contained a capturing parenthesis was
1117 successfully matched, but the branch in which it appeared failed, the
1118 capturing was not being forgotten if a higher numbered group was later
1119 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1120 group 1 to "a", when in fact it should be unset. This applied to multi-
1121 branched capturing and non-capturing groups, repeated or not, and also to
1122 positive assertions (capturing in negative assertions does not happen
1123 in PCRE) and also to nested atomic groups.
1124
1125 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1126 subject after a captured substring, to make it easier to tell which of a
1127 number of identical substrings has been captured.
1128
1129 24. The way atomic groups are processed by pcre_exec() has been changed so that
1130 if they are repeated, backtracking one repetition now resets captured
1131 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1132 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1133 "aaa". Previously, it would have been "a". As part of this code
1134 refactoring, the way recursive calls are handled has also been changed.
1135
1136 25. If an assertion condition captured any substrings, they were not passed
1137 back unless some other capturing happened later. For example, if
1138 (?(?=(a))a) was matched against "a", no capturing was returned.
1139
1140 26. When studying a pattern that contained subroutine calls or assertions,
1141 the code for finding the minimum length of a possible match was handling
1142 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1143 group 1 called group 2 while simultaneously a separate group 2 called group
1144 1). A stack overflow occurred in this case. I have fixed this by limiting
1145 the recursion depth to 10.
1146
1147 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1148 Fortmann. This supports explicit test numbers on the command line, and has
1149 argument validation and error reporting.
1150
1151 28. An instance of \X with an unlimited repeat could fail if at any point the
1152 first character it looked at was a mark character.
1153
1154 29. Some minor code refactoring concerning Unicode properties and scripts
1155 should reduce the stack requirement of match() slightly.
1156
1157 30. Added the '=' option to pcretest to check the setting of unused capturing
1158 slots at the end of the pattern, which are documented as being -1, but are
1159 not included in the return count.
1160
1161 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1162 compiled something random. Now it gives a compile-time error (as does
1163 Perl).
1164
1165 32. A *MARK encountered during the processing of a positive assertion is now
1166 recorded and passed back (compatible with Perl).
1167
1168 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1169 had alternative anchored branches, the search for a second match in a line
1170 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1171 matched the line "0102" twice. The same bug affected patterns that started
1172 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1173 twice.
1174
1175 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1176 does, and because of capturing effects, quantifying parenthesized
1177 assertions may at times be useful. Quantifiers are now allowed for
1178 parenthesized assertions.
1179
1180 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1181
1182 36. \g was being checked for fancy things in a character class, when it should
1183 just be a literal "g".
1184
1185 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1186 appearance of a nested POSIX class supersedes an apparent external class.
1187 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1188 unescaped square brackets may also appear as part of class names. For
1189 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1190 more like Perl. (But see 8.20/1 above.)
1191
1192 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1193 was because it thought it was \N{name}, which is not supported).
1194
1195 39. Add minix to OS list not supporting the -S option in pcretest.
1196
1197 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1198 cannot analyze patterns in sufficient detail to catch mutual recursions
1199 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1200 subgroup is called recursively as a subpattern for a second time at the
1201 same position in the subject string. In previous releases this might have
1202 been caught by the recursion limit, or it might have run out of stack.
1203
1204 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1205 happen only once. PCRE was, however incorrectly giving a compile time error
1206 "recursive call could loop indefinitely" because it cannot analyze the
1207 pattern in sufficient detail. The compile time test no longer happens when
1208 PCRE is compiling a conditional subpattern, but actual runaway loops are
1209 now caught at runtime (see 40 above).
1210
1211 42. It seems that Perl allows any characters other than a closing parenthesis
1212 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1213 has been changed to be the same.
1214
1215 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1216 as not to get warnings when autogen.sh is called. Also changed
1217 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1218
1219 44. To help people who use pcregrep to scan files containing exceedingly long
1220 lines, the following changes have been made:
1221
1222 (a) The default value of the buffer size parameter has been increased from
1223 8K to 20K. (The actual buffer used is three times this size.)
1224
1225 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1226 PCRE is built.
1227
1228 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1229 to be set at run time.
1230
1231 (d) Numerical values in pcregrep options can be followed by K or M, for
1232 example --buffer-size=50K.
1233
1234 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1235 given and the return code is set to 2.
1236
1237 45. Add a pointer to the latest mark to the callout data block.
1238
1239 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1240 partial match of an empty string instead of no match. This was specific to
1241 the use of ".".
1242
1243 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1244 complete match instead of a partial match. This bug was dependent on both
1245 the PCRE_UTF8 and PCRE_DOTALL options being set.
1246
1247 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1248 starting byte set, because \b was not being ignored.
1249
1250
1251 Version 8.12 15-Jan-2011
1252 ------------------------
1253
1254 1. Fixed some typos in the markup of the man pages, and wrote a script that
1255 checks for such things as part of the documentation building process.
1256
1257 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1258 --match-limit and --recursion-limit options (added for 8.11). In
1259 particular, this made one of the standard tests fail. (The integer value
1260 went into the wrong half of a long int.)
1261
1262 3. If the --colour option was given to pcregrep with -v (invert match), it
1263 did strange things, either producing crazy output, or crashing. It should,
1264 of course, ignore a request for colour when reporting lines that do not
1265 match.
1266
1267 4. Another pcregrep bug caused similar problems if --colour was specified with
1268 -M (multiline) and the pattern match finished with a line ending.
1269
1270 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1271 matched in multiline mode, the following line was shown as part of the
1272 match. This seems wrong, so I have changed it.
1273
1274 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1275 the check for further matches in the same line (so they could be coloured)
1276 to overrun the end of the current line. If another match was found, it was
1277 incorrectly shown (and then shown again when found in the next line).
1278
1279 7. If pcregrep was compiled under Windows, there was a reference to the
1280 function pcregrep_exit() before it was defined. I am assuming this was
1281 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1282 reported by a user. I've moved the definition above the reference.
1283
1284
1285 Version 8.11 10-Dec-2010
1286 ------------------------
1287
1288 1. (*THEN) was not working properly if there were untried alternatives prior
1289 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1290 backtracked to try for "b" instead of moving to the next alternative branch
1291 at the same level (in this case, to look for "c"). The Perl documentation
1292 is clear that when (*THEN) is backtracked onto, it goes to the "next
1293 alternative in the innermost enclosing group".
1294
1295 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1296 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1297 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1298 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1299 (*THEN).
1300
1301 3. If \s appeared in a character class, it removed the VT character from
1302 the class, even if it had been included by some previous item, for example
1303 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1304 of \s, but is part of the POSIX "space" class.)
1305
1306 4. A partial match never returns an empty string (because you can always
1307 match an empty string at the end of the subject); however the checking for
1308 an empty string was starting at the "start of match" point. This has been
1309 changed to the "earliest inspected character" point, because the returned
1310 data for a partial match starts at this character. This means that, for
1311 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1312 (previously it gave "no match").
1313
1314 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1315 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1316 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1317 has an implication that the given string is incomplete (because a partial
1318 match is preferred over a full match). For this reason, these items now
1319 give a partial match in this situation. [Aside: previously, the one case
1320 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1321 match rather than a full match, which was wrong by the old rules, but is
1322 now correct.]
1323
1324 6. There was a bug in the handling of #-introduced comments, recognized when
1325 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1326 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1327 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1328 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1329 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1330 places in pcre_compile().
1331
1332 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1333 comments when looking ahead for named forward references to subpatterns,
1334 the only newline sequence it recognized was NL. It now handles newlines
1335 according to the set newline convention.
1336
1337 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1338 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1339 cater for a lack of strerror(). These oversights have been fixed.
1340
1341 9. Added --match-limit and --recursion-limit to pcregrep.
1342
1343 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1344
1345 11. When the -o option was used, pcregrep was setting a return code of 1, even
1346 when matches were found, and --line-buffered was not being honoured.
1347
1348 12. Added an optional parentheses number to the -o and --only-matching options
1349 of pcregrep.
1350
1351 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1352 can match an empty string. The code to do it in pcretest and pcredemo
1353 needed fixing:
1354
1355 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1356 only one byte after an empty string match just before CRLF (this case
1357 just got forgotten; "any" and "anycrlf" were OK).
1358
1359 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1360 mode when an empty string match preceded an ASCII character followed by
1361 a non-ASCII character. (The code for advancing by one character rather
1362 than one byte was nonsense.)
1363
1364 (c) The pcredemo.c sample program did not have any code at all to handle
1365 the cases when CRLF is a valid newline sequence.
1366
1367 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1368 as a starting offset was within the subject string. There is now a new
1369 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1370 negative or greater than the length of the string. In order to test this,
1371 pcretest is extended to allow the setting of negative starting offsets.
1372
1373 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1374 starting offset points to the beginning of a UTF-8 character was
1375 unnecessarily clumsy. I tidied it up.
1376
1377 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1378 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1379
1380 17. Nobody had reported that the --include_dir option, which was added in
1381 release 7.7 should have been called --include-dir (hyphen, not underscore)
1382 for compatibility with GNU grep. I have changed it to --include-dir, but
1383 left --include_dir as an undocumented synonym, and the same for
1384 --exclude-dir, though that is not available in GNU grep, at least as of
1385 release 2.5.4.
1386
1387 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1388 characters from a string of bytes) have been redefined so as not to use
1389 loops, in order to improve performance in some environments. At the same
1390 time, I abstracted some of the common code into auxiliary macros to save
1391 repetition (this should not affect the compiled code).
1392
1393 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1394 compile-time error is now given if \c is not followed by an ASCII
1395 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1396 different, and any byte value is allowed.)
1397
1398 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1399 START_OPTIMIZE option, which is now allowed at compile time - but just
1400 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1401 to pcregrep and other applications that have no direct access to PCRE
1402 options. The new /Y option in pcretest sets this option when calling
1403 pcre_compile().
1404
1405 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1406 back references. Groups containing recursive back references were forced to
1407 be atomic by that change, but in the case of named groups, the amount of
1408 memory required was incorrectly computed, leading to "Failed: internal
1409 error: code overflow". This has been fixed.
1410
1411 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1412 pcretest.c, to avoid build problems in some Borland environments.
1413
1414
1415 Version 8.10 25-Jun-2010
1416 ------------------------
1417
1418 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1419 THEN.
1420
1421 2. (*ACCEPT) was not working when inside an atomic group.
1422
1423 3. Inside a character class, \B is treated as a literal by default, but
1424 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1425 causes the error). The code is unchanged, but I tidied the documentation.
1426
1427 4. Inside a character class, PCRE always treated \R and \X as literals,
1428 whereas Perl faults them if its -w option is set. I have changed PCRE so
1429 that it faults them when PCRE_EXTRA is set.
1430
1431 5. Added support for \N, which always matches any character other than
1432 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1433
1434 6. When compiling pcregrep with newer versions of gcc which may have
1435 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1436 declared with attribute warn_unused_result" were given. Just casting the
1437 result to (void) does not stop the warnings; a more elaborate fudge is
1438 needed. I've used a macro to implement this.
1439
1440 7. Minor change to pcretest.c to avoid a compiler warning.
1441
1442 8. Added four artifical Unicode properties to help with an option to make
1443 \s etc use properties (see next item). The new properties are: Xan
1444 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1445
1446 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1447 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1448 this option. Modified pcretest to add /W to test this facility. Added
1449 REG_UCP to make it available via the POSIX interface.
1450
1451 10. Added --line-buffered to pcregrep.
1452
1453 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1454 studied, and the match started with a letter with a code point greater than
1455 127 whose first byte was different to the first byte of the other case of
1456 the letter, the other case of this starting letter was not recognized
1457 (#976).
1458
1459 12. If a pattern that was studied started with a repeated Unicode property
1460 test, for example, \p{Nd}+, there was the theoretical possibility of
1461 setting up an incorrect bitmap of starting bytes, but fortunately it could
1462 not have actually happened in practice until change 8 above was made (it
1463 added property types that matched character-matching opcodes).
1464
1465 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1466 possible starting bytes for non-anchored patterns.
1467
1468 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1469 \R, and also a number of cases that involve Unicode properties, both
1470 explicit and implicit when PCRE_UCP is set.
1471
1472 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1473 input, it could crash or give wrong results if characters with values
1474 greater than 0xc0 were present in the subject string. (Detail: it assumed
1475 UTF-8 input when processing these items.)
1476
1477 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1478 size_t is 64-bit (#991).
1479
1480 17. Added a check for running out of memory when PCRE is compiled with
1481 --disable-stack-for-recursion (#990).
1482
1483 18. If the last data line in a file for pcretest does not have a newline on
1484 the end, a newline was missing in the output.
1485
1486 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1487 less than 128) in its various bitmaps. However, there is a facility for
1488 generating tables according to the current locale when PCRE is compiled. It
1489 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1490 characters, are recognized by isspace() and therefore were getting set in
1491 these tables, and indeed these tables seem to approximate to ISO 8859. This
1492 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1493 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1494 which of course cannot start UTF-8 characters. I have changed the code so
1495 that only real ASCII characters (less than 128) and the correct starting
1496 bytes for UTF-8 encodings are set for characters greater than 127 when in
1497 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1498 altogether.)
1499
1500 20. Added the /T option to pcretest so as to be able to run tests with non-
1501 standard character tables, thus making it possible to include the tests
1502 used for 19 above in the standard set of tests.
1503
1504 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1505 reference to a subpattern the other side of a comment that contains an
1506 opening parenthesis caused either an internal compiling error, or a
1507 reference to the wrong subpattern.
1508
1509
1510 Version 8.02 19-Mar-2010
1511 ------------------------
1512
1513 1. The Unicode data tables have been updated to Unicode 5.2.0.
1514
1515 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1516 configured.
1517
1518 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1519 original author of that file, following a query about its status.
1520
1521 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1522 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1523
1524 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1525 quantifier applied to a forward-referencing subroutine call, could compile
1526 incorrect code or give the error "internal error: previously-checked
1527 referenced subpattern not found".
1528
1529 6. Both MS Visual Studio and Symbian OS have problems with initializing
1530 variables to point to external functions. For these systems, therefore,
1531 pcre_malloc etc. are now initialized to local functions that call the
1532 relevant global functions.
1533
1534 7. There were two entries missing in the vectors called coptable and poptable
1535 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1536 I've fixed the data, and added a kludgy way of testing at compile time that
1537 the lengths are correct (equal to the number of opcodes).
1538
1539 8. Following on from 7, I added a similar kludge to check the length of the
1540 eint vector in pcreposix.c.
1541
1542 9. Error texts for pcre_compile() are held as one long string to avoid too
1543 much relocation at load time. To find a text, the string is searched,
1544 counting zeros. There was no check for running off the end of the string,
1545 which could happen if a new error number was added without updating the
1546 string.
1547
1548 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1549
1550 11. \K was not working if it appeared in an atomic group or in a group that
1551 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1552 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1553 the assertion is positive, but not if it is negative.
1554
1555 12. Change 11 fortuitously reduced the size of the stack frame used in the
1556 "match()" function of pcre_exec.c by one pointer. Forthcoming
1557 implementation of support for (*MARK) will need an extra pointer on the
1558 stack; I have reserved it now, so that the stack frame size does not
1559 decrease.
1560
1561 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1562 item in branch that calls a recursion is a subroutine call - as in the
1563 second branch in the above example - was incorrectly given the compile-
1564 time error "recursive call could loop indefinitely" because pcre_compile()
1565 was not correctly checking the subroutine for matching a non-empty string.
1566
1567 14. The checks for overrunning compiling workspace could trigger after an
1568 overrun had occurred. This is a "should never occur" error, but it can be
1569 triggered by pathological patterns such as hundreds of nested parentheses.
1570 The checks now trigger 100 bytes before the end of the workspace.
1571
1572 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1573
1574
1575 Version 8.01 19-Jan-2010
1576 ------------------------
1577
1578 1. If a pattern contained a conditional subpattern with only one branch (in
1579 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1580 computed the wrong minimum data length (which is of course zero for such
1581 subpatterns). This could cause incorrect "no match" results.
1582
1583 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1584 the pattern is reset in the first branch, pcre_compile() failed with
1585 "internal error: code overflow at offset...". This happened only when
1586 the reset was to the original external option setting. (An optimization
1587 abstracts leading options settings into an external setting, which was the
1588 cause of this.)
1589
1590 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1591 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1592 assertion pattern did not match (meaning that the assertion was true), it
1593 was incorrectly treated as false if the SKIP had been reached during the
1594 matching. This also applied to assertions used as conditions.
1595
1596 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1597 assertion subpattern, including such a pattern used as a condition,
1598 unpredictable results occurred, instead of the error return
1599 PCRE_ERROR_DFA_UITEM.
1600
1601 5. The C++ GlobalReplace function was not working like Perl for the special
1602 situation when an empty string is matched. It now does the fancy magic
1603 stuff that is necessary.
1604
1605 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1606 removed. (These were left over from very, very early versions of PCRE.)
1607
1608 7. Some cosmetic changes to the code to make life easier when compiling it
1609 as part of something else:
1610
1611 (a) Change DEBUG to PCRE_DEBUG.
1612
1613 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1614 called "current" as "current_branch", to prevent a collision with the
1615 Linux macro when compiled as a kernel module.
1616
1617 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1618 prevent a collision with the Linux macro when compiled as a kernel
1619 module.
1620
1621 8. In pcre_compile() there are some checks for integer overflows that used to
1622 cast potentially large values to (double). This has been changed to that
1623 when building, a check for int64_t is made, and if it is found, it is used
1624 instead, thus avoiding the use of floating point arithmetic. (There is no
1625 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1626 double.
1627
1628 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1629 2005 (difference between two addresses compared to an unsigned value).
1630
1631 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1632 custom one, because of the following reported problem in Windows:
1633
1634 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1635 under Win32.
1636 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1637 therefore missing the function definition.
1638 - The compiler thus generates a "C" signature for the test function.
1639 - The linker fails to find the "C" function.
1640 - PCRE fails to configure if asked to do so against libbz2.
1641
1642 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1643 messages were output:
1644
1645 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1646 rerunning libtoolize, to keep the correct libtool macros in-tree.
1647 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1648
1649 I have done both of these things.
1650
1651 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1652 most of the time, it *can* run out if it is given a pattern that contains a
1653 runaway infinite recursion. I updated the discussion in the pcrestack man
1654 page.
1655
1656 13. Now that we have gone to the x.xx style of version numbers, the minor
1657 version may start with zero. Using 08 or 09 is a bad idea because users
1658 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1659 interpreted as invalid octal numbers. I've updated the previous comment in
1660 configure.ac, and also added a check that gives an error if 08 or 09 are
1661 used.
1662
1663 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1664 causing partial matching to fail when the end of the subject matched \W
1665 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1666
1667 15. There were some discrepancies between the declarations in pcre_internal.h
1668 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1669 their definitions. The declarations used "const uschar *" and the
1670 definitions used USPTR. Even though USPTR is normally defined as "const
1671 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1672 reported that: "This difference in casting confuses some C++ compilers, for
1673 example, SunCC recognizes above declarations as different functions and
1674 generates broken code for hbpcre." I have changed the declarations to use
1675 USPTR.
1676
1677 16. GNU libtool is named differently on some systems. The autogen.sh script now
1678 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1679 (FreeBSD).
1680
1681 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1682 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1683 comment: "Figure out how to create a longlong from a string: strtoll and
1684 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1685 instance, but it only takes 2 args instead of 3!"
1686
1687 18. A subtle bug concerned with back references has been fixed by a change of
1688 specification, with a corresponding code fix. A pattern such as
1689 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1690 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1691 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1692 same bug. Such groups have to be quantified to be useful, or contained
1693 inside another quantified group. (If there's no repetition, the reference
1694 can never match.) The problem arises because, having left the group and
1695 moved on to the rest of the pattern, a later failure that backtracks into
1696 the group uses the captured value from the final iteration of the group
1697 rather than the correct earlier one. I have fixed this in PCRE by forcing
1698 any group that contains a reference to itself to be an atomic group; that
1699 is, there cannot be any backtracking into it once it has completed. This is
1700 similar to recursive and subroutine calls.
1701
1702
1703 Version 8.00 19-Oct-09
1704 ----------------------
1705
1706 1. The table for translating pcre_compile() error codes into POSIX error codes
1707 was out-of-date, and there was no check on the pcre_compile() error code
1708 being within the table. This could lead to an OK return being given in
1709 error.
1710
1711 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1712 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1713 in a Windows environment.
1714
1715 3. The pcregrep --count option prints the count for each file even when it is
1716 zero, as does GNU grep. However, pcregrep was also printing all files when
1717 --files-with-matches was added. Now, when both options are given, it prints
1718 counts only for those files that have at least one match. (GNU grep just
1719 prints the file name in this circumstance, but including the count seems
1720 more useful - otherwise, why use --count?) Also ensured that the
1721 combination -clh just lists non-zero counts, with no names.
1722
1723 4. The long form of the pcregrep -F option was incorrectly implemented as
1724 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1725 but it seems right to fix it, and I didn't think it was worth preserving
1726 the old behaviour.
1727
1728 5. The command line items --regex=pattern and --regexp=pattern were not
1729 recognized by pcregrep, which required --regex pattern or --regexp pattern
1730 (with a space rather than an '='). The man page documented the '=' forms,
1731 which are compatible with GNU grep; these now work.
1732
1733 6. No libpcreposix.pc file was created for pkg-config; there was just
1734 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1735
1736 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1737 when UCP support is not needed, by modifying the Python script that
1738 generates it from Unicode data files. This should not matter if the module
1739 is correctly used as a library, but I received one complaint about 50K of
1740 unwanted data. My guess is that the person linked everything into his
1741 program rather than using a library. Anyway, it does no harm.
1742
1743 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1744 was a minimum greater than 1 for a wide character in a possessive
1745 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1746 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1747 character. Chaos in the form of incorrect output or a compiling loop could
1748 result.
1749
1750 9. The restrictions on what a pattern can contain when partial matching is
1751 requested for pcre_exec() have been removed. All patterns can now be
1752 partially matched by this function. In addition, if there are at least two
1753 slots in the offset vector, the offset of the earliest inspected character
1754 for the match and the offset of the end of the subject are set in them when
1755 PCRE_ERROR_PARTIAL is returned.
1756
1757 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1758 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1759 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1760 and may be more useful for multi-segment matching.
1761
1762 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1763 used to be given if ever the end of the subject was reached; now it is
1764 given only if matching could not proceed because another character was
1765 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1766 string "Z", which now yields "no match" instead of "partial match". In the
1767 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1768 final character ended with (*FAIL).
1769
1770 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1771 if the pattern had a "must contain" character that was already found in the
1772 earlier partial match, unless partial matching was again requested. For
1773 example, with the pattern /dog.(body)?/, the "must contain" character is
1774 "g". If the first part-match was for the string "dog", restarting with
1775 "sbody" failed. This bug has been fixed.
1776
1777 13. The string returned by pcre_dfa_exec() after a partial match has been
1778 changed so that it starts at the first inspected character rather than the
1779 first character of the match. This makes a difference only if the pattern
1780 starts with a lookbehind assertion or \b or \B (\K is not supported by
1781 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1782 matching functions compatible, and I think it's the right thing to do.
1783
1784 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1785 so that the demonstration program is easily available in environments where
1786 PCRE has not been installed from source.
1787
1788 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1789 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1790 library.
1791
1792 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1793 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1794 is not the first non-POSIX option to be added. Clearly some people find
1795 these options useful.
1796
1797 17. If a caller to the POSIX matching function regexec() passes a non-zero
1798 value for nmatch with a NULL value for pmatch, the value of
1799 nmatch is forced to zero.
1800
1801 18. RunGrepTest did not have a test for the availability of the -u option of
1802 the diff command, as RunTest does. It now checks in the same way as
1803 RunTest, and also checks for the -b option.
1804
1805 19. If an odd number of negated classes containing just a single character
1806 interposed, within parentheses, between a forward reference to a named
1807 subpattern and the definition of the subpattern, compilation crashed with
1808 an internal error, complaining that it could not find the referenced
1809 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1810 [The bug was that it was starting one character too far in when skipping
1811 over the character class, thus treating the ] as data rather than
1812 terminating the class. This meant it could skip too much.]
1813
1814 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1815 /g option in pcretest when the pattern contains \K, which makes it possible
1816 to have an empty string match not at the start, even when the pattern is
1817 anchored. Updated pcretest and pcredemo to use this option.
1818
1819 21. If the maximum number of capturing subpatterns in a recursion was greater
1820 than the maximum at the outer level, the higher number was returned, but
1821 with unset values at the outer level. The correct (outer level) value is
1822 now given.
1823
1824 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1825 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1826 make it do so. The string so far is captured, making this feature
1827 compatible with Perl.
1828
1829 23. The tests have been re-organized, adding tests 11 and 12, to make it
1830 possible to check the Perl 5.10 features against Perl 5.10.
1831
1832 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1833 pattern matches a fixed length string. PCRE did not allow this; now it
1834 does. Neither allows recursion.
1835
1836 25. I finally figured out how to implement a request to provide the minimum
1837 length of subject string that was needed in order to match a given pattern.
1838 (It was back references and recursion that I had previously got hung up
1839 on.) This code has now been added to pcre_study(); it finds a lower bound
1840 to the length of subject needed. It is not necessarily the greatest lower
1841 bound, but using it to avoid searching strings that are too short does give
1842 some useful speed-ups. The value is available to calling programs via
1843 pcre_fullinfo().
1844
1845 26. While implementing 25, I discovered to my embarrassment that pcretest had
1846 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1847 study optimizations had never been tested with that matching function.
1848 Oops. What is worse, even when it was passed study data, there was a bug in
1849 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1850 were also very few tests of studied patterns with pcre_dfa_exec().
1851
1852 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1853 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1854 on the other side of the coin, they are no longer allowed to have different
1855 names, because these cannot be distinguished in PCRE, and this has caused
1856 confusion. (This is a difference from Perl.)
1857
1858 28. When duplicate subpattern names are present (necessarily with different
1859 numbers, as required by 27 above), and a test is made by name in a
1860 conditional pattern, either for a subpattern having been matched, or for
1861 recursion in such a pattern, all the associated numbered subpatterns are
1862 tested, and the overall condition is true if the condition is true for any
1863 one of them. This is the way Perl works, and is also more like the way
1864 testing by number works.
1865
1866
1867 Version 7.9 11-Apr-09
1868 ---------------------
1869
1870 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1871 (pcretest), all targets were linked against these libraries. This included
1872 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1873 libraries. This caused unwanted dependencies to be created. This problem
1874 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1875 pcretest is linked with readline.
1876
1877 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1878 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1879 moved outside it again, because FALSE and TRUE are already defined in AIX,
1880 but BOOL is not.
1881
1882 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1883 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1884
1885 4. The pcregrep documentation said spaces were inserted as well as colons (or
1886 hyphens) following file names and line numbers when outputting matching
1887 lines. This is not true; no spaces are inserted. I have also clarified the
1888 wording for the --colour (or --color) option.
1889
1890 5. In pcregrep, when --colour was used with -o, the list of matching strings
1891 was not coloured; this is different to GNU grep, so I have changed it to be
1892 the same.
1893
1894 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1895 each matching line was coloured. Now it goes on to look for further matches
1896 of any of the test patterns, which is the same behaviour as GNU grep.
1897
1898 7. A pattern that could match an empty string could cause pcregrep to loop; it
1899 doesn't make sense to accept an empty string match in pcregrep, so I have
1900 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1901 seems to be how GNU grep behaves. [But see later change 40 for release
1902 8.33.]
1903
1904 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1905 start or after a newline", because the conditional assertion was not being
1906 correctly handled. The rule now is that both the assertion and what follows
1907 in the first alternative must satisfy the test.
1908
1909 9. If auto-callout was enabled in a pattern with a conditional group whose
1910 condition was an assertion, PCRE could crash during matching, both with
1911 pcre_exec() and pcre_dfa_exec().
1912
1913 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1914 used for matching.
1915
1916 11. Unicode property support in character classes was not working for
1917 characters (bytes) greater than 127 when not in UTF-8 mode.
1918
1919 12. Added the -M command line option to pcretest.
1920
1921 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1922
1923 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1924
1925 16. Added comments and documentation about mis-use of no_arg in the C++
1926 wrapper.
1927
1928 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1929 from Martin Jerabek that uses macro names for all relevant character and
1930 string constants.
1931
1932 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1933 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1934 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1935 these, but not everybody uses configure.
1936
1937 19. A conditional group that had only one branch was not being correctly
1938 recognized as an item that could match an empty string. This meant that an
1939 enclosing group might also not be so recognized, causing infinite looping
1940 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1941 with the subject "ab", where knowledge that the repeated group can match
1942 nothing is needed in order to break the loop.
1943
1944 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1945 exec(), but without supplying a callout function, matching went wrong.
1946
1947 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1948 leak if the size of the offset vector was greater than 30. When the vector
1949 is smaller, the saved offsets during recursion go onto a local stack
1950 vector, but for larger vectors malloc() is used. It was failing to free
1951 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1952 error, in fact).
1953
1954 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1955 heapframe that is used only when UTF-8 support is enabled. This caused no
1956 problem, but was untidy.
1957
1958 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1959 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1960 included within another project.
1961
1962 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1963 slightly modified by me:
1964
1965 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1966 not building pcregrep.
1967
1968 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1969 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1970
1971 25. Forward references, both numeric and by name, in patterns that made use of
1972 duplicate group numbers, could behave incorrectly or give incorrect errors,
1973 because when scanning forward to find the reference group, PCRE was not
1974 taking into account the duplicate group numbers. A pattern such as
1975 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1976
1977 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1978 the feature of a custom pointer more persuasive (as requested by a user).
1979
1980 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1981 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1982 pre-defined.
1983
1984 28. Added support for (*UTF8) at the start of a pattern.
1985
1986 29. Arrange for flags added by the "release type" setting in CMake to be shown
1987 in the configuration summary.
1988
1989
1990 Version 7.8 05-Sep-08
1991 ---------------------
1992
1993 1. Replaced UCP searching code with optimized version as implemented for Ad
1994 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1995 stage table and inline lookup instead of a function, giving speed ups of 2
1996 to 5 times on some simple patterns that I tested. Permission was given to
1997 distribute the MultiStage2.py script that generates the tables (it's not in
1998 the tarball, but is in the Subversion repository).
1999
2000 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2001 scripts.
2002
2003 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2004 a group with a zero qualifier. The result of the study could be incorrect,
2005 or the function might crash, depending on the pattern.
2006
2007 4. Caseless matching was not working for non-ASCII characters in back
2008 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2009 It now works when Unicode Property Support is available.
2010
2011 5. In pcretest, an escape such as \x{de} in the data was always generating
2012 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2013 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2014 truncation.
2015
2016 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2017
2018 7. Added two (int) casts to pcregrep when printing the difference of two
2019 pointers, in case they are 64-bit values.
2020
2021 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2022 test 2 if it fails.
2023
2024 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2025 and a #define of that name to empty if it is not externally set. This is to
2026 allow users of MSVC to set it if necessary.
2027
2028 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2029 the convenience functions in the pcre_get.c source file.
2030
2031 11. An option change at the start of a pattern that had top-level alternatives
2032 could cause overwriting and/or a crash. This command provoked a crash in
2033 some environments:
2034
2035 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2036
2037 This potential security problem was recorded as CVE-2008-2371.
2038
2039 12. For a pattern where the match had to start at the beginning or immediately
2040 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2041 pcre_dfa_exec() could read past the end of the passed subject if there was
2042 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2043 pcretest so that it places the subject at the end of its malloc-ed buffer.
2044
2045 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2046 exec() might read past the end of the data buffer in UTF-8 mode.
2047
2048 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2049 the data contained the byte 0x85 as part of a UTF-8 character within its
2050 first line. This applied both to normal and DFA matching.
2051
2052 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2053 /^[^d]*?$/8 failed to match "abc".
2054
2055 16. Added a missing copyright notice to pcrecpp_internal.h.
2056
2057 17. Make it more clear in the documentation that values returned from
2058 pcre_exec() in ovector are byte offsets, not character counts.
2059
2060 18. Tidied a few places to stop certain compilers from issuing warnings.
2061
2062 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2063 supplied by Stefan Weber. I made a further small update for 7.8 because
2064 there is a change of source arrangements: the pcre_searchfuncs.c module is
2065 replaced by pcre_ucd.c.
2066
2067
2068 Version 7.7 07-May-08
2069 ---------------------
2070
2071 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2072 a string to a long long, pretend we don't even have a long long." This is
2073 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2074
2075 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2076 pre-7.6 versions, which defined a global no_arg variable instead of putting
2077 it in the RE class. (See also #8 below.)
2078
2079 3. Remove a line of dead code, identified by coverity and reported by Nuno
2080 Lopes.
2081
2082 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2083
2084 (1) The include/exclude patterns were being applied to the whole pathnames
2085 of files, instead of just to the final components.
2086
2087 (2) If there was more than one level of directory, the subdirectories were
2088 skipped unless they satisfied the include/exclude conditions. This is
2089 inconsistent with GNU grep (and could even be seen as contrary to the
2090 pcregrep specification - which I improved to make it absolutely clear).
2091 The action now is always to scan all levels of directory, and just
2092 apply the include/exclude patterns to regular files.
2093
2094 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2095 --exclude_dir in the tests to avoid scanning .svn directories.
2096
2097 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2098 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2099 doesn't support NULs in patterns.
2100
2101 7. Added some missing "const"s to declarations of static tables in
2102 pcre_compile.c and pcre_dfa_exec.c.
2103
2104 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2105 caused by fix #2 above. (Subsequently also a second patch to fix the
2106 first patch. And a third patch - this was a messy problem.)
2107
2108 9. Applied Craig's patch to remove the use of push_back().
2109
2110 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2111 matching function regexec().
2112
2113 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2114 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2115 references. PCRE supports relative numbers with this syntax (I don't think
2116 Oniguruma does).
2117
2118 12. Previously, a group with a zero repeat such as (...){0} was completely
2119 omitted from the compiled regex. However, this means that if the group
2120 was called as a subroutine from elsewhere in the pattern, things went wrong
2121 (an internal error was given). Such groups are now left in the compiled
2122 pattern, with a new opcode that causes them to be skipped at execution
2123 time.
2124
2125 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2126 to the way PCRE behaves:
2127
2128 (a) A lone ] character is dis-allowed (Perl treats it as data).
2129
2130 (b) A back reference to an unmatched subpattern matches an empty string
2131 (Perl fails the current match path).
2132
2133 (c) A data ] in a character class must be notated as \] because if the
2134 first data character in a class is ], it defines an empty class. (In
2135 Perl it is not possible to have an empty class.) The empty class []
2136 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2137 The negative empty class [^] matches any one character, independently
2138 of the DOTALL setting.
2139
2140 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2141 non-existent subpattern following a character class starting with ']' and
2142 containing () gave an internal compiling error instead of "reference to
2143 non-existent subpattern". Fortunately, when the pattern did exist, the
2144 compiled code was correct. (When scanning forwards to check for the
2145 existence of the subpattern, it was treating the data ']' as terminating
2146 the class, so got the count wrong. When actually compiling, the reference
2147 was subsequently set up correctly.)
2148
2149 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2150 it was being rejected as not supported by pcre_dfa_exec(), even though
2151 other assertions are supported. I have made pcre_dfa_exec() support
2152 (*FAIL).
2153
2154 16. The implementation of 13c above involved the invention of a new opcode,
2155 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2156 cannot be changed at match time, I realized I could make a small
2157 improvement to matching performance by compiling OP_ALLANY instead of
2158 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2159 on the OP_ANY path.
2160
2161 17. Compiling pcretest on Windows with readline support failed without the
2162 following two fixes: (1) Make the unistd.h include conditional on
2163 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2164
2165 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2166 ncurses library to be included for pcretest when ReadLine support is
2167 requested, but also to allow for it to be overridden. This patch came from
2168 Daniel Bergström.
2169
2170 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2171 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2172 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2173 spotting this.
2174
2175
2176 Version 7.6 28-Jan-08
2177 ---------------------
2178
2179 1. A character class containing a very large number of characters with
2180 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2181 overflow.
2182
2183 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2184 HAVE_LONG_LONG is not defined.
2185
2186 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2187 bring them up to date and include new features. This patch includes:
2188
2189 - Fixed PH's badly added libz and libbz2 support.
2190 - Fixed a problem with static linking.
2191 - Added pcredemo. [But later removed - see 7 below.]
2192 - Fixed dftables problem and added an option.
2193 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2194 HAVE_LONG_LONG.
2195 - Added readline support for pcretest.
2196 - Added an listing of the option settings after cmake has run.
2197
2198 4. A user submitted a patch to Makefile that makes it easy to create
2199 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2200 Makefile.am that cause it to include this special target, without
2201 affecting anything else. Note that the same mingw target plus all
2202 the other distribution libraries and programs are now supported
2203 when configuring with CMake (see 6 below) instead of with
2204 Configure/Make.
2205
2206 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2207 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2208 exported in the Windows port". It has not yet been confirmed that the patch
2209 solves the problem, but it does no harm.
2210
2211 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2212 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2213 with CMake, and also correct the comment about stack recursion.
2214
2215 7. Remove the automatic building of pcredemo from the ./configure system and
2216 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2217 of a program that users should build themselves after PCRE is installed, so
2218 building it automatically is not really right. What is more, it gave
2219 trouble in some build environments.
2220
2221 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2222
2223
2224 Version 7.5 10-Jan-08
2225 ---------------------
2226
2227 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2228 values in parens when parsing an RE using the C++ wrapper."
2229
2230 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2231 Characters greater than 255 were excluded from the class instead of being
2232 included.
2233
2234 3. The same bug as (2) above applied to negated POSIX classes such as
2235 [:^space:].
2236
2237 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2238 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2239 I have changed it.
2240
2241 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2242 first named subpattern) and a construct such as (?&a) would reference the
2243 first named subpattern whose name started with "a" (in other words, the
2244 length check was missing). Both these problems are fixed. "Subpattern name
2245 expected" is now given for (?&) (a zero-length name), and this patch also
2246 makes it give the same error for \k'' (previously it complained that that
2247 was a reference to a non-existent subpattern).
2248
2249 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2250 this is right because (?- can be followed by option settings as well as by
2251 digits. I have, however, made the messages clearer.
2252
2253 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2254 than the number used in the conditional) now cause a compile-time error.
2255 This is actually not compatible with Perl, which accepts such patterns, but
2256 treats the conditional as always being FALSE (as PCRE used to), but it
2257 seems to me that giving a diagnostic is better.
2258
2259 8. Change "alphameric" to the more common word "alphanumeric" in comments
2260 and messages.
2261
2262 9. Fix two occurrences of "backslash" in comments that should have been
2263 "backspace".
2264
2265 10. Remove two redundant lines of code that can never be obeyed (their function
2266 was moved elsewhere).
2267
2268 11. The program that makes PCRE's Unicode character property table had a bug
2269 which caused it to generate incorrect table entries for sequences of
2270 characters that have the same character type, but are in different scripts.
2271 It amalgamated them into a single range, with the script of the first of
2272 them. In other words, some characters were in the wrong script. There were
2273 thirteen such cases, affecting characters in the following ranges:
2274
2275 U+002b0 - U+002c1
2276 U+0060c - U+0060d
2277 U+0061e - U+00612
2278 U+0064b - U+0065e
2279 U+0074d - U+0076d
2280 U+01800 - U+01805
2281 U+01d00 - U+01d77
2282 U+01d9b - U+01dbf
2283 U+0200b - U+0200f
2284 U+030fc - U+030fe
2285 U+03260 - U+0327f
2286 U+0fb46 - U+0fbb1
2287 U+10450 - U+1049d
2288
2289 12. The -o option (show only the matching part of a line) for pcregrep was not
2290 compatible with GNU grep in that, if there was more than one match in a
2291 line, it showed only the first of them. It now behaves in the same way as
2292 GNU grep.
2293
2294 13. If the -o and -v options were combined for pcregrep, it printed a blank
2295 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2296 does the same. The return code can be used to tell if there were any
2297 non-matching lines.
2298
2299 14. Added --file-offsets and --line-offsets to pcregrep.
2300
2301 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2302 infinitely looping recursion. The bug was that positive lookaheads were not
2303 being skipped when checking for a possible empty match (negative lookaheads
2304 and both kinds of lookbehind were skipped).
2305
2306 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2307 inclusion of <windows.h> to before rather than after the definition of
2308 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2309
2310 17. Specifying a possessive quantifier with a specific limit for a Unicode
2311 character property caused pcre_compile() to compile bad code, which led at
2312 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2313 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2314 caused the error; without that there was no problem.
2315
2316 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2317
2318 19. Added --enable-pcretest-libreadline.
2319
2320 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2321 RE::GlobalReplace(). As a result, the number of replacements returned was
2322 double what it should be. I removed one of the increments, but Craig sent a
2323 later patch that removed the other one (the right fix) and added unit tests
2324 that check the return values (which was not done before).
2325
2326 21. Several CMake things:
2327
2328 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2329 the names libpcre and libpcreposix, not just pcre and pcreposix.
2330
2331 (2) The above change means that pcretest and pcregrep are now correctly
2332 linked with the newly-built libraries, not previously installed ones.
2333
2334 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2335
2336 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2337 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2338 UTF-8 newline character). The key issue is that the pattern starts .*;
2339 this means that the match must be either at the beginning, or after a
2340 newline. The bug was in the code for advancing after a failed match and
2341 checking that the new position followed a newline. It was not taking
2342 account of UTF-8 characters correctly.
2343
2344 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2345 character classes. PCRE was not treating the sequence [:...:] as a
2346 character class unless the ... were all letters. Perl, however, seems to
2347 allow any characters between [: and :], though of course it rejects as
2348 unknown any "names" that contain non-letters, because all the known class
2349 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2350 for example, whereas PCRE did not - it did not recognize a POSIX character
2351 class. This seemed a bit dangerous, so the code has been changed to be
2352 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2353 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2354 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2355 Perl does, and where it didn't before.
2356
2357 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2358 Windows environments %n is disabled by default.
2359
2360
2361 Version 7.4 21-Sep-07
2362 ---------------------
2363
2364 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2365 means that a class such as [\s] counted as "explicit reference to CR or
2366 LF". That isn't really right - the whole point of the change was to try to
2367 help when there was an actual mention of one of the two characters. So now
2368 the change happens only if \r or \n (or a literal CR or LF) character is
2369 encountered.
2370
2371 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2372 of both had grown to the point where there were only 3 bits left.
2373 Fortunately, there was spare space in the data structure, and so I have
2374 moved the internal flags into a new 16-bit field to free up more option
2375 bits.
2376
2377 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2378 but did not set the internal JCHANGED flag - either of these is enough to
2379 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2380 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2381 start sets both bits.
2382
2383 4. Added options (at build time, compile time, exec time) to change \R from
2384 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2385
2386 5. doc/pcresyntax.html was missing from the distribution.
2387
2388 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2389 compatibility, even though it is no longer used.
2390
2391 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2392 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2393 windows.h file is present (where different names are used). [This was
2394 reversed later after testing - see 16 below.]
2395
2396 8. Changed all #include <config.h> to #include "config.h". There were also
2397 some further <pcre.h> cases that I changed to "pcre.h".
2398
2399 9. When pcregrep was used with the --colour option, it missed the line ending
2400 sequence off the lines that it output.
2401
2402 10. It was pointed out to me that arrays of string pointers cause lots of
2403 relocations when a shared library is dynamically loaded. A technique of
2404 using a single long string with a table of offsets can drastically reduce
2405 these. I have refactored PCRE in four places to do this. The result is
2406 dramatic:
2407
2408 Originally: 290
2409 After changing UCP table: 187
2410 After changing error message table: 43
2411 After changing table of "verbs" 36
2412 After changing table of Posix names 22
2413
2414 Thanks to the folks working on Gregex for glib for this insight.
2415
2416 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2417 unicode-properties was also set.
2418
2419 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2420
2421 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2422 checked only for CRLF.
2423
2424 14. Added casts to pcretest.c to avoid compiler warnings.
2425
2426 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2427
2428 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2429 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2430 entirely. This removes changes made in 7 above.
2431
2432 17. The CMake files have been updated, and there is now more information about
2433 building with CMake in the NON-UNIX-USE document.
2434
2435
2436 Version 7.3 28-Aug-07
2437 ---------------------
2438
2439 1. In the rejigging of the build system that eventually resulted in 7.1, the
2440 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2441 brackets there is not right, since it causes compilers to look for an
2442 installed pcre.h, not the version that is in the source that is being
2443 compiled (which of course may be different). I have changed it back to:
2444
2445 #include "pcre.h"
2446
2447 I have a vague recollection that the change was concerned with compiling in
2448 different directories, but in the new build system, that is taken care of
2449 by the VPATH setting the Makefile.
2450
2451 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2452 when the subject happened to end in the byte 0x85 (e.g. if the last
2453 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2454 characters but of course it shouldn't be taken as a newline when it is part
2455 of another character. The bug was that, for an unlimited repeat of . in
2456 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2457 characters when looking for a newline.
2458
2459 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2460
2461 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2462 in debug output.
2463
2464 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2465 long printing in the pcrecpp unittest when running under MinGW.
2466
2467 6. ESC_K was left out of the EBCDIC table.
2468
2469 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2470 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2471 limit also applies to "virtual nesting" when a pattern is recursive, and in
2472 this case 1000 isn't so big. I have been able to remove this limit at the
2473 expense of backing off one optimization in certain circumstances. Normally,
2474 when pcre_exec() would call its internal match() function recursively and
2475 immediately return the result unconditionally, it uses a "tail recursion"
2476 feature to save stack. However, when a subpattern that can match an empty
2477 string has an unlimited repetition quantifier, it no longer makes this
2478 optimization. That gives it a stack frame in which to save the data for
2479 checking that an empty string has been matched. Previously this was taken
2480 from the 1000-entry workspace that had been reserved. So now there is no
2481 explicit limit, but more stack is used.
2482
2483 8. Applied Daniel's patches to solve problems with the import/export magic
2484 syntax that is required for Windows, and which was going wrong for the
2485 pcreposix and pcrecpp parts of the library. These were overlooked when this
2486 problem was solved for the main library.
2487
2488 9. There were some crude static tests to avoid integer overflow when computing
2489 the size of patterns that contain repeated groups with explicit upper
2490 limits. As the maximum quantifier is 65535, the maximum group length was
2491 set at 30,000 so that the product of these two numbers did not overflow a
2492 32-bit integer. However, it turns out that people want to use groups that
2493 are longer than 30,000 bytes (though not repeat them that many times).
2494 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2495 made it possible to implement the integer overflow checks in a much more
2496 dynamic way, which I have now done. The artificial limitation on group
2497 length has been removed - we now have only the limit on the total length of
2498 the compiled pattern, which depends on the LINK_SIZE setting.
2499
2500 10. Fixed a bug in the documentation for get/copy named substring when
2501 duplicate names are permitted. If none of the named substrings are set, the
2502 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2503 empty string.
2504
2505 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2506 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2507 because the ] is interpreted as the first data character and the
2508 terminating ] is not found. PCRE has been made compatible with Perl in this
2509 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2510 cause memory overwriting.
2511
2512 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2513 string has been matched (to stop an infinite loop). It was not recognizing
2514 a conditional subpattern that could match an empty string if that
2515 subpattern was within another subpattern. For example, it looped when
2516 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2517 condition was not nested. This bug has been fixed.
2518
2519 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2520 past the start of the subject in the presence of bytes with the top bit
2521 set, for example "\x8aBCD".
2522
2523 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2524 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2525
2526 14. Optimized (?!) to (*FAIL).
2527
2528 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2529 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2530 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2531 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2532 does: it's just the validity check that is more restrictive.
2533
2534 16. Inserted checks for integer overflows during escape sequence (backslash)
2535 processing, and also fixed erroneous offset values for syntax errors during
2536 backslash processing.
2537
2538 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2539 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2540
2541 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2542 caused an overrun.
2543
2544 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2545 something other than just ASCII characters) inside a group that had an
2546 unlimited repeat caused a loop at compile time (while checking to see
2547 whether the group could match an empty string).
2548
2549 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2550 [\P{Any}] did so. (Error in the code for printing property names.)
2551
2552 21. An orphan \E inside a character class could cause a crash.
2553
2554 22. A repeated capturing bracket such as (A)? could cause a wild memory
2555 reference during compilation.
2556
2557 23. There are several functions in pcre_compile() that scan along a compiled
2558 expression for various reasons (e.g. to see if it's fixed length for look
2559 behind). There were bugs in these functions when a repeated \p or \P was
2560 present in the pattern. These operators have additional parameters compared
2561 with \d, etc, and these were not being taken into account when moving along
2562 the compiled data. Specifically:
2563
2564 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2565 length.
2566
2567 (b) An item such as \pL+ within a repeated group could cause crashes or
2568 loops.
2569
2570 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2571 "reference to non-existent subpattern" error.
2572
2573 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2574
2575 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2576 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2577
2578 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2579
2580 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2581 character were causing crashes (broken optimization).
2582
2583 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2584 \p or \P) caused a compile-time loop.
2585
2586 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2587 break. For example, the unstudied pattern [\r\n]A does not match the string
2588 "\r\nA" because change 7.0/46 below moves the current point on by two
2589 characters after failing to match at the start. However, the pattern \nA
2590 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2591 the same is true. There doesn't seem any very clean way out of this, but
2592 what I have chosen to do makes the common cases work: PCRE now takes note
2593 of whether there can be an explicit match for \r or \n anywhere in the
2594 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2595 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2596 pattern has explicit CR or LF references.
2597
2598 29. Added (*CR) etc for changing newline setting at start of pattern.
2599
2600
2601 Version 7.2 19-Jun-07
2602 ---------------------
2603
2604 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2605 which is apparently normally available under Windows.
2606
2607 2. Re-jig the pcregrep tests with different newline settings in an attempt
2608 to make them independent of the local environment's newline setting.
2609
2610 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2611
2612 4. Some of the "internals" tests were previously cut out when the link size
2613 was not 2, because the output contained actual offsets. The recent new
2614 "Z" feature of pcretest means that these can be cut out, making the tests
2615 usable with all link sizes.
2616
2617 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2618 stack recursion. This gives a massive performance boost under BSD, but just
2619 a small improvement under Linux. However, it saves one field in the frame
2620 in all cases.
2621
2622 6. Added more features from the forthcoming Perl 5.10:
2623
2624 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2625 recursion call. It refers to the nth most recently opened parentheses.
2626
2627 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2628 to be opened parentheses.
2629
2630 (c) Conditions that refer to capturing parentheses can be specified
2631 relatively, for example, (?(-2)... or (?(+3)...
2632
2633 (d) \K resets the start of the current match so that everything before
2634 is not part of it.
2635
2636 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2637
2638 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2639 reference syntax.
2640
2641 (g) (?| introduces a group in which the numbering of parentheses in each
2642 alternative starts with the same number.
2643
2644 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2645
2646 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2647 PCRE_INFO_JCHANGED.
2648
2649 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2650 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2651 for detecting groups that can match an empty string.
2652
2653 9. A pattern with a very large number of alternatives (more than several
2654 hundred) was running out of internal workspace during the pre-compile
2655 phase, where pcre_compile() figures out how much memory will be needed. A
2656 bit of new cunning has reduced the workspace needed for groups with
2657 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2658 workspace instead of running out of the 4096 that are available.
2659
2660 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2661
2662 11. Applied patch from Google to remove an optimization that didn't quite work.
2663 The report of the bug said:
2664
2665 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2666 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2667 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2668
2669 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2670 it matched the wrong number of bytes.
2671
2672
2673 Version 7.1 24-Apr-07
2674 ---------------------
2675
2676 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2677 that is more "standard", making use of automake and other Autotools. There
2678 is some re-arrangement of the files and adjustment of comments consequent
2679 on this.
2680
2681 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2682 for recursive directory scanning broke on some systems because the files
2683 are not scanned in any specific order and on different systems the order
2684 was different. A call to "sort" has been inserted into RunGrepTest for the
2685 approprate test as a short-term fix. In the longer term there may be an
2686 alternative.
2687
2688 3. I had an email from Eric Raymond about problems translating some of PCRE's
2689 man pages to HTML (despite the fact that I distribute HTML pages, some
2690 people do their own conversions for various reasons). The problems
2691 concerned the use of low-level troff macros .br and .in. I have therefore
2692 removed all such uses from the man pages (some were redundant, some could
2693 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2694 HTML has been updated to handle .nf/.fi and to complain if it encounters
2695 .br or .in.
2696
2697 4. Updated comments in configure.ac that get placed in config.h.in and also
2698 arranged for config.h to be included in the distribution, with the name
2699 config.h.generic, for the benefit of those who have to compile without
2700 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2701
2702 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2703 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2704 makevp.bat for the current PCRE, using the additional files
2705 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2706
2707 6. A Windows user reported a minor discrepancy with test 2, which turned out
2708 to be caused by a trailing space on an input line that had got lost in his
2709 copy. The trailing space was an accident, so I've just removed it.
2710
2711 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2712 that is needed.
2713
2714 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2715 as "const" (a) because they are and (b) because it helps the PHP
2716 maintainers who have recently made a script to detect big data structures
2717 in the php code that should be moved to the .rodata section. I remembered
2718 to update Builducptable as well, so it won't revert if ucptable.h is ever
2719 re-created.
2720
2721 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2722 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2723 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2724 support is not required. This saves 1.5-2K of code, which is important in
2725 some applications.
2726
2727 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2728 so as not to refer to the tables, even though these functions will never be
2729 called when UTF-8 support is disabled. Otherwise there are problems with a
2730 shared library.
2731
2732 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2733
2734 (a) It was defining its arguments as char * instead of void *.
2735
2736 (b) It was assuming that all moves were upwards in memory; this was true
2737 a long time ago when I wrote it, but is no longer the case.
2738
2739 The emulated memove() is provided for those environments that have neither
2740 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2741 is clearly not the case, as these two bugs were recently reported.
2742
2743 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2744 and Detrail to create the HTML documentation, the .txt form of the man
2745 pages, and it removes trailing spaces from listed files. It also creates
2746 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2747 case, it wraps all the #defines with #ifndefs. This script should be run
2748 before "make dist".
2749
2750 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2751 with Unicode property support.
2752
2753 (a) For a maximizing quantifier, if the two different cases of the
2754 character were of different lengths in their UTF-8 codings (there are
2755 some cases like this - I found 11), and the matching function had to
2756 back up over a mixture of the two cases, it incorrectly assumed they
2757 were both the same length.
2758
2759 (b) When PCRE was configured to use the heap rather than the stack for
2760 recursion during matching, it was not correctly preserving the data for
2761 the other case of a UTF-8 character when checking ahead for a match
2762 while processing a minimizing repeat. If the check also involved
2763 matching a wide character, but failed, corruption could cause an
2764 erroneous result when trying to check for a repeat of the original
2765 character.
2766
2767 13. Some tidying changes to the testing mechanism:
2768
2769 (a) The RunTest script now detects the internal link size and whether there
2770 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2771 values substituted by "configure". (The RunGrepTest script already did
2772 this for UTF-8.) The configure.ac script no longer substitutes the
2773 relevant variables.
2774
2775 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2776 with length and offset values. This means that the output is different
2777 for different internal link sizes. Test 2 is skipped for link sizes
2778 other than 2 because of this, bypassing the problem. Unfortunately,
2779 there was also a test in test 3 (the locale tests) that used /B and
2780 failed for link sizes other than 2. Rather than cut the whole test out,
2781 I have added a new /Z option to pcretest that replaces the length and
2782 offset values with spaces. This is now used to make test 3 independent
2783 of link size. (Test 2 will be tidied up later.)
2784
2785 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2786 segmentation fault instead of returning the appropriate error message.
2787
2788 15. In multiline mode when the newline sequence was set to "any", the pattern
2789 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2790 This doesn't seem right; it now treats the CRLF combination as the line
2791 ending, and so does not match in that case. It's only a pattern such as ^$
2792 that would hit this one: something like ^ABC$ would have failed after \r
2793 and then tried again after \r\n.
2794
2795 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2796 in an attempt to make files that differ only in their line terminators
2797 compare equal. This works on Linux.
2798
2799 17. Under certain error circumstances pcregrep might try to free random memory
2800 as it exited. This is now fixed, thanks to valgrind.
2801
2802 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2803 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2804 was because its rules for how to advance for /g after matching an empty
2805 string at the end of a line did not allow for this case. They now check for
2806 it specially.
2807
2808 20. pcretest is supposed to handle patterns and data of any length, by
2809 extending its buffers when necessary. It was getting this wrong when the
2810 buffer for a data line had to be extended.
2811
2812 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2813 CRLF as a newline sequence.
2814
2815 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2816 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2817 I have nevertheless tidied it up.
2818
2819 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2820
2821 24. Added a man page for pcre-config.
2822
2823
2824 Version 7.0 19-Dec-06
2825 ---------------------
2826
2827 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2828 moving to gcc 4.1.1.
2829
2830 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2831 sys/time.h, which is documented as needed for this function. It doesn't
2832 seem to matter on Linux, but it showed up on some releases of OS X.
2833
2834 3. It seems that there are systems where bytes whose values are greater than
2835 127 match isprint() in the "C" locale. The "C" locale should be the
2836 default when a C program starts up. In most systems, only ASCII printing
2837 characters match isprint(). This difference caused the output from pcretest
2838 to vary, making some of the tests fail. I have changed pcretest so that:
2839
2840 (a) When it is outputting text in the compiled version of a pattern, bytes
2841 other than 32-126 are always shown as hex escapes.
2842
2843 (b) When it is outputting text that is a matched part of a subject string,
2844 it does the same, unless a different locale has been set for the match
2845 (using the /L modifier). In this case, it uses isprint() to decide.
2846
2847 4. Fixed a major bug that caused incorrect computation of the amount of memory
2848 required for a compiled pattern when options that changed within the
2849 pattern affected the logic of the preliminary scan that determines the
2850 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2851 that the computed length was too small. The symptoms of this bug were
2852 either the PCRE error "internal error: code overflow" from pcre_compile(),
2853 or a glibc crash with a message such as "pcretest: free(): invalid next
2854 size (fast)". Examples of patterns that provoked this bug (shown in
2855 pcretest format) are:
2856
2857 /(?-x: )/x
2858 /(?x)(?-x: \s*#\s*)/
2859 /((?i)[\x{c0}])/8
2860 /(?i:[\x{c0}])/8
2861
2862 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2863 is now done differently.
2864
2865 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2866 wrapper classes; (b) implement a new function in the C++ scanner that is
2867 more efficient than the old way of doing things because it avoids levels of
2868 recursion in the regex matching; (c) add a paragraph to the documentation
2869 for the FullMatch() function.
2870
2871 6. The escape sequence \n was being treated as whatever was defined as
2872 "newline". Not only was this contrary to the documentation, which states
2873 that \n is character 10 (hex 0A), but it also went horribly wrong when
2874 "newline" was defined as CRLF. This has been fixed.
2875
2876 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2877 was being set to -1 for the "end of line" case (supposedly a value that no
2878 character can have). Though this value is never used (the check for end of
2879 line is "zero bytes in current character"), it caused compiler complaints.
2880 I've changed it to 0xffffffff.
2881
2882 8. In pcre_version.c, the version string was being built by a sequence of
2883 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2884 string (as it is for production releases) called a macro with an empty
2885 argument. The C standard says the result of this is undefined. The gcc
2886 compiler treats it as an empty string (which was what was wanted) but it is
2887 reported that Visual C gives an error. The source has been hacked around to
2888 avoid this problem.
2889
2890 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2891 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2892 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2893 of them did).
2894
2895 10. Originally, pcretest opened its input and output without "b"; then I was
2896 told that "b" was needed in some environments, so it was added for release
2897 5.0 to both the input and output. (It makes no difference on Unix-like
2898 systems.) Later I was told that it is wrong for the input on Windows. I've
2899 now abstracted the modes into two macros, to make it easier to fiddle with
2900 them, and removed "b" from the input mode under Windows.
2901
2902 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2903
2904 12. Added -help and --help to pcretest as an official way of being reminded
2905 of the options.
2906
2907 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2908 and pcrecpp.cc because they annoy compilers at high warning levels.
2909
2910 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2911
2912 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2913 scripts are not C programs :-) and which was not noticed because it works
2914 on Linux.
2915
2916 16. pcretest is supposed to handle any length of pattern and data line (as one
2917 line or as a continued sequence of lines) by extending its input buffer if
2918 necessary. This feature was broken for very long pattern lines, leading to
2919 a string of junk being passed to pcre_compile() if the pattern was longer
2920 than about 50K.
2921
2922 17. I have done a major re-factoring of the way pcre_compile() computes the
2923 amount of memory needed for a compiled pattern. Previously, there was code
2924 that made a preliminary scan of the pattern in order to do this. That was
2925 OK when PCRE was new, but as the facilities have expanded, it has become
2926 harder and harder to keep it in step with the real compile phase, and there
2927 have been a number of bugs (see for example, 4 above). I have now found a
2928 cunning way of running the real compile function in a "fake" mode that
2929 enables it to compute how much memory it would need, while actually only
2930 ever using a few hundred bytes of working memory and without too many
2931 tests of the mode. This should make future maintenance and development
2932 easier. A side effect of this work is that the limit of 200 on the nesting
2933 depth of parentheses has been removed (though this was never a serious
2934 limitation, I suspect). However, there is a downside: pcre_compile() now
2935 runs more slowly than before (30% or more, depending on the pattern). I
2936 hope this isn't a big issue. There is no effect on runtime performance.
2937
2938 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2939 newline (only possible for the last line of a file) and it was a
2940 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2941
2942 19. Added additional timing features to pcretest. (1) The -tm option now times
2943 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2944 separate command line item, by a number that specifies the number of
2945 repeats to use when timing. The default is 50000; this gives better
2946 precision, but takes uncomfortably long for very large patterns.
2947
2948 20. Extended pcre_study() to be more clever in cases where a branch of a
2949 subpattern has no definite first character. For example, (a*|b*)[cd] would
2950 previously give no result from pcre_study(). Now it recognizes that the
2951 first character must be a, b, c, or d.
2952
2953 21. There was an incorrect error "recursive call could loop indefinitely" if
2954 a subpattern (or the entire pattern) that was being tested for matching an
2955 empty string contained only one non-empty item after a nested subpattern.
2956 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2957 incorrectly, because the \d was being skipped in the check.
2958
2959 22. The pcretest program now has a new pattern option /B and a command line
2960 option -b, which is equivalent to adding /B to every pattern. This causes
2961 it to show the compiled bytecode, without the additional information that
2962 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2963 is the same as /B/I).
2964
2965 23. A new optimization is now able automatically to treat some sequences such
2966 as a*b as a*+b. More specifically, if something simple (such as a character
2967 or a simple class like \d) has an unlimited quantifier, and is followed by
2968 something that cannot possibly match the quantified thing, the quantifier
2969 is automatically "possessified".
2970
2971 24. A recursive reference to a subpattern whose number was greater than 39
2972 went wrong under certain circumstances in UTF-8 mode. This bug could also
2973 have affected the operation of pcre_study().
2974
2975 25. Realized that a little bit of performance could be had by replacing
2976 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2977
2978 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2979
2980 27. Possessive quantifiers such as a++ were previously implemented by turning
2981 them into atomic groups such as ($>a+). Now they have their own opcodes,
2982 which improves performance. This includes the automatically created ones
2983 from 23 above.
2984
2985 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2986 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2987 the first matched character to be a colon. This applied both to named and
2988 numbered groups.
2989
2990 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2991
2992 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2993 building PCRE on Windows easier, so I have included it in the distribution.
2994
2995 31. There is now a check in pcretest against a ridiculously large number being
2996 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2997 loop, the loop is abandoned.
2998
2999 32. Forward references to subpatterns in conditions such as (?(2)...) where
3000 subpattern 2 is defined later cause pcre_compile() to search forwards in
3001 the pattern for the relevant set of parentheses. This search went wrong
3002 when there were unescaped parentheses in a character class, parentheses
3003 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3004
3005 33. "Subroutine" calls and backreferences were previously restricted to
3006 referencing subpatterns earlier in the regex. This restriction has now
3007 been removed.
3008
3009 34. Added a number of extra features that are going to be in Perl 5.10. On the
3010 whole, these are just syntactic alternatives for features that PCRE had
3011 previously implemented using the Python syntax or my own invention. The
3012 other formats are all retained for compatibility.
3013
3014 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3015 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3016 also .NET compatible.
3017
3018 (b) A recursion or subroutine call to a named group can now be defined as
3019 (?&name) as well as (?P>name).
3020
3021 (c) A backreference to a named group can now be defined as \k<name> or
3022 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3023 5.10, are also .NET compatible.
3024
3025 (d) A conditional reference to a named group can now use the syntax
3026 (?(<name>) or (?('name') as well as (?(name).
3027
3028 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3029 groups (named and numbered) that are never evaluated inline, but can be
3030 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3031 is always false. There may be only one alternative in such a group.
3032
3033 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3034 as the simple (?(R). The condition is true only if the most recent
3035 recursion is that of the given number or name. It does not search out
3036 through the entire recursion stack.
3037
3038 (g) The escape \gN or \g{N} has been added, where N is a positive or
3039 negative number, specifying an absolute or relative reference.
3040
3041 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3042 some "unreachable code" warnings.
3043
3044 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3045 things, this adds five new scripts.
3046
3047 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3048 There were also incompatibilities regarding the handling of \Q..\E inside
3049 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3050 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3051
3052 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3053 matches an empty string, and forcibly breaks the loop. There were bugs in
3054 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3055 against aaaa the result was just "a" rather than "aaaa", for example. Two
3056 separate and independent bugs (that affected different cases) have been
3057 fixed.
3058
3059 39. Refactored the code to abolish the use of different opcodes for small
3060 capturing bracket numbers. This is a tidy that I avoided doing when I
3061 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3062 The new approach is not only tidier, it makes it possible to reduce the
3063 memory needed to fix the previous bug (38).
3064
3065 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3066 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3067 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3068 mode.
3069
3070 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3071 report.
3072
3073 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3074 copy construction and assignment in the C++ wrapper.
3075
3076 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3077 couple of bugs that could have given wrong results in the "--newline=crlf"
3078 case.
3079
3080 44. Added a number of casts and did some reorganization of signed/unsigned int
3081 variables following suggestions from Dair Grant. Also renamed the variable
3082 "this" as "item" because it is a C++ keyword.
3083
3084 45. Arranged for dftables to add
3085
3086 #include "pcre_internal.h"
3087
3088 to pcre_chartables.c because without it, gcc 4.x may remove the array
3089 definition from the final binary if PCRE is built into a static library and
3090 dead code stripping is activated.
3091
3092 46. For an unanchored pattern, if a match attempt fails at the start of a
3093 newline sequence, and the newline setting is CRLF or ANY, and the next two
3094 characters are CRLF, advance by two characters instead of one.
3095
3096
3097 Version 6.7 04-Jul-06
3098 ---------------------
3099
3100 1. In order to handle tests when input lines are enormously long, pcretest has
3101 been re-factored so that it automatically extends its buffers when
3102 necessary. The code is crude, but this _is_ just a test program. The
3103 default size has been increased from 32K to 50K.
3104
3105 2. The code in pcre_study() was using the value of the re argument before
3106 testing it for NULL. (Of course, in any sensible call of the function, it
3107 won't be NULL.)
3108
3109 3. The memmove() emulation function in pcre_internal.h, which is used on
3110 systems that lack both memmove() and bcopy() - that is, hardly ever -
3111 was missing a "static" storage class specifier.
3112
3113 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3114 containing an extended class (one that cannot be represented by a bitmap
3115 because it contains high-valued characters or Unicode property items, e.g.
3116 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3117 pattern, but PCRE should not loop if you do not (it no longer does).
3118 [Detail: two cases were found: (a) a repeated subpattern containing an
3119 extended class; (b) a recursive reference to a subpattern that followed a
3120 previous extended class. It wasn't skipping over the extended class
3121 correctly when UTF-8 mode was not set.]
3122
3123 5. A negated single-character class was not being recognized as fixed-length
3124 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3125 compile error "lookbehind assertion is not fixed length".
3126
3127 6. The RunPerlTest auxiliary script was showing an unexpected difference
3128 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3129 write a Perl script that can interpret lines of an input file either as
3130 byte characters or as UTF-8, which is what "perltest" was being required to
3131 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3132 can't do is switch easily at run time between having the "use utf8;" pragma
3133 or not. In the end, I fudged it by using the RunPerlTest script to insert
3134 "use utf8;" explicitly for the UTF-8 tests.
3135
3136 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3137 the end of the subject string, contrary to the documentation and to what
3138 Perl does. This was true of both matching functions. Now it matches only at
3139 the start of the subject and immediately after *internal* newlines.
3140
3141 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3142 a pointer to an int instead of a pointer to an unsigned long int. This
3143 caused problems on 64-bit systems.
3144
3145 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3146 instance of the 'standard' template library not being so standard".
3147
3148 10. There was no check on the number of named subpatterns nor the maximum
3149 length of a subpattern name. The product of these values is used to compute
3150 the size of the memory block for a compiled pattern. By supplying a very
3151 long subpattern name and a large number of named subpatterns, the size
3152 computation could be caused to overflow. This is now prevented by limiting
3153 the length of names to 32 characters, and the number of named subpatterns
3154 to 10,000.
3155
3156 11. Subpatterns that are repeated with specific counts have to be replicated in
3157 the compiled pattern. The size of memory for this was computed from the
3158 length of the subpattern and the repeat count. The latter is limited to
3159 65535, but there was no limit on the former, meaning that integer overflow
3160 could in principle occur. The compiled length of a repeated subpattern is
3161 now limited to 30,000 bytes in order to prevent this.
3162
3163 12. Added the optional facility to have named substrings with the same name.
3164
3165 13. Added the ability to use a named substring as a condition, using the
3166 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3167 are numbers (not recommended). Forward references are permitted.
3168
3169 14. Added forward references in named backreferences (if you see what I mean).
3170
3171 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3172 pattern could run off the end of the subject. For example, the pattern
3173 "(?s)(.{1,5})"8 did this with the subject "ab".
3174
3175 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3176 PCRE_CASELESS was set when matching characters that were quantified with ?
3177 or *.
3178
3179 17. A character class other than a single negated character that had a minimum
3180 but no maximum quantifier - for example [ab]{6,} - was not handled
3181 correctly by pce_dfa_exec(). It would match only one character.
3182
3183 18. A valid (though odd) pattern that looked like a POSIX character
3184 class but used an invalid character after [ (for example [[,abc,]]) caused
3185 pcre_compile() to give the error "Failed: internal error: code overflow" or
3186 in some cases to crash with a glibc free() error. This could even happen if
3187 the pattern terminated after [[ but there just happened to be a sequence of
3188 letters, a binary zero, and a closing ] in the memory that followed.
3189
3190 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3191 over the years. Originally (before any Unicode support), just the bottom 8
3192 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3193 output from "man perlunicode" includes this:
3194
3195 The regular expression compiler produces polymorphic opcodes. That
3196 is, the pattern adapts to the data and automatically switches to
3197 the Unicode character scheme when presented with Unicode data--or
3198 instead uses a traditional byte scheme when presented with byte
3199 data.
3200
3201 Sadly, a wide octal escape does not cause a switch, and in a string with
3202 no other multibyte characters, these octal escapes are treated as before.
3203 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3204 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3205 Unicode string.
3206
3207 I have not perpetrated such confusion in PCRE. Up till now, it took just
3208 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3209 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3210 translate to the appropriate multibyte character.
3211
3212 29. Applied some refactoring to reduce the number of warnings from Microsoft
3213 and Borland compilers. This has included removing the fudge introduced
3214 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3215 a warning about an unused variable.
3216
3217 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3218 characters since release 4.0, because Perl (from release 5.004) does not.
3219 [Or at least, is documented not to: some releases seem to be in conflict
3220 with the documentation.] However, when a pattern was studied with
3221 pcre_study() and all its branches started with \s, PCRE still included VT
3222 as a possible starting character. Of course, this did no harm; it just
3223 caused an unnecessary match attempt.
3224
3225 22. Removed a now-redundant internal flag bit that recorded the fact that case
3226 dependency changed within the pattern. This was once needed for "required
3227 byte" processing, but is no longer used. This recovers a now-scarce options
3228 bit. Also moved the least significant internal flag bit to the most-
3229 significant bit of the word, which was not previously used (hangover from
3230 the days when it was an int rather than a uint) to free up another bit for
3231 the future.
3232
3233 23. Added support for CRLF line endings as well as CR and LF. As well as the
3234 default being selectable at build time, it can now be changed at runtime
3235 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3236 specify that it is scanning data with non-default line endings.
3237
3238 24. Changed the definition of CXXLINK to make it agree with the definition of
3239 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3240
3241 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3242 recursions. This makes a big different to stack usage for some patterns.
3243
3244 26. If a subpattern containing a named recursion or subroutine reference such
3245 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3246 the space required for the compiled pattern went wrong and gave too small a
3247 value. Depending on the environment, this could lead to "Failed: internal
3248 error: code overflow at offset 49" or "glibc detected double free or
3249 corruption" errors.
3250
3251 27. Applied patches from Google (a) to support the new newline modes and (b) to
3252 advance over multibyte UTF-8 characters in GlobalReplace.
3253
3254 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3255 difference for some implementation of PCRE in some Windows version.
3256
3257 29. Added some extra testing facilities to pcretest:
3258
3259 \q<number> in a data line sets the "match limit" value
3260 \Q<number> in a data line sets the "match recursion limt" value
3261 -S <number> sets the stack size, where <number> is in megabytes
3262
3263 The -S option isn't available for Windows.
3264
3265
3266 Version 6.6 06-Feb-06
3267 ---------------------
3268
3269 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3270 in pcreposix.h. I have copied the definition from pcre.h.
3271
3272 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3273 because pcre.h is no longer a built file.
3274
3275 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3276 not normally included in the compiled code.
3277
3278
3279 Version 6.5 01-Feb-06
3280 ---------------------
3281
3282 1. When using the partial match feature with pcre_dfa_exec(), it was not
3283 anchoring the second and subsequent partial matches at the new starting
3284 point. This could lead to incorrect results. For example, with the pattern
3285 /1234/, partially matching against "123" and then "a4" gave a match.
3286
3287 2. Changes to pcregrep:
3288
3289 (a) All non-match returns from pcre_exec() were being treated as failures
3290 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3291 error message is output. Some extra information is given for the
3292 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3293 probably the only errors that are likely to be caused by users (by
3294 specifying a regex that has nested indefinite repeats, for instance).
3295 If there are more than 20 of these errors, pcregrep is abandoned.
3296
3297 (b) A binary zero was treated as data while matching, but terminated the
3298 output line if it was written out. This has been fixed: binary zeroes
3299 are now no different to any other data bytes.
3300
3301 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3302 used to set a locale for matching. The --locale=xxxx long option has
3303 been added (no short equivalent) to specify a locale explicitly on the
3304 pcregrep command, overriding the environment variables.
3305
3306 (d) When -B was used with -n, some line numbers in the output were one less
3307 than they should have been.
3308
3309 (e) Added the -o (--only-matching) option.
3310
3311 (f) If -A or -C was used with -c (count only), some lines of context were
3312 accidentally printed for the final match.
3313
3314 (g) Added the -H (--with-filename) option.
3315
3316 (h) The combination of options -rh failed to suppress file names for files
3317 that were found from directory arguments.
3318
3319 (i) Added the -D (--devices) and -d (--directories) options.
3320
3321 (j) Added the -F (--fixed-strings) option.
3322
3323 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3324
3325 (l) Added the --colo(u)r option.
3326
3327 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3328 is not present by default.
3329
3330 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3331 items such as (?R) or (?1), when the recursion could match a number of
3332 alternatives. If it matched one of the alternatives, but subsequently,
3333 outside the recursion, there was a failure, the code tried to back up into
3334 the recursion. However, because of the way PCRE is implemented, this is not
3335 possible, and the result was an incorrect result from the match.
3336
3337 In order to prevent this happening, the specification of recursion has
3338 been changed so that all such subpatterns are automatically treated as
3339 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3340
3341 4. I had overlooked the fact that, in some locales, there are characters for
3342 which isalpha() is true but neither isupper() nor islower() are true. In
3343 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3344 and ordfeminine) are like this. This affected the treatment of \w and \W
3345 when they appeared in character classes, but not when they appeared outside
3346 a character class. The bit map for "word" characters is now created
3347 separately from the results of isalnum() instead of just taking it from the
3348 upper, lower, and digit maps. (Plus the underscore character, of course.)
3349
3350 5. The above bug also affected the handling of POSIX character classes such as
3351 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3352 permanent tables. Instead, the bit maps for such a class were previously
3353 created as the appropriate unions of the upper, lower, and digit bitmaps.
3354 Now they are created by subtraction from the [[:word:]] class, which has
3355 its own bitmap.
3356
3357 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3358 It is created by subtracting the vertical space characters (\x09, \x0a,
3359 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3360 subtraction was done in the overall bitmap for a character class, meaning
3361 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3362 be recognized. This bug has been fixed.
3363
3364 7. Patches from the folks at Google:
3365
3366 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3367 real life, but is still worth protecting against".
3368
3369 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3370 regular expressions".
3371
3372 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3373 have it.
3374
3375 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3376 "configure" and the latter not, in order to fix a problem somebody had
3377 with compiling the Arg class on HP-UX.
3378
3379 (e) Improve the error-handling of the C++ wrapper a little bit.
3380
3381 (f) New tests for checking recursion limiting.
3382
3383 8. The pcre_memmove() function, which is used only if the environment does not
3384 have a standard memmove() function (and is therefore rarely compiled),
3385 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3386 returning a result (though PCRE never actually uses the result).
3387
3388 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3389 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3390 returned instead of calling malloc() with an overflowing number that would
3391 most likely cause subsequent chaos.
3392
3393 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3394
3395 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3396 with this option is matched, the nmatch and pmatch options of regexec() are
3397 ignored.
3398
3399 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3400 provided in case anyone wants to the the POSIX interface with UTF-8
3401 strings.
3402
3403 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3404 C++ linking (needed for some HP-UX environments).
3405
3406 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3407 (unused parameter) and in the pcre_printint() function (omitted "default"
3408 switch label when the default is to do nothing).
3409
3410 15. Added some code to make it possible, when PCRE is compiled as a C++
3411 library, to replace subject pointers for pcre_exec() with a smart pointer
3412 class, thus making it possible to process discontinuous strings.
3413
3414 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3415 much the same function. They were added by different people who were trying
3416 to make PCRE easy to compile on non-Unix systems. It has been suggested
3417 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3418 for compiling on Windows systems. I have therefore replaced it with
3419 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3420 defaults to "extern" for C or "extern C" for C++, which works fine on
3421 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3422 SCOPE with something explicit in config.h. In addition:
3423
3424 (a) pcreposix.h still had just "extern" instead of either of these macros;
3425 I have replaced it with PCRE_DATA_SCOPE.
3426
3427 (b) Functions such as _pcre_xclass(), which are internal to the library,
3428 but external in the C sense, all had PCRE_EXPORT in their definitions.
3429 This is apparently wrong for the Windows case, so I have removed it.
3430 (It makes no difference on Unix-like systems.)
3431
3432 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3433 of recursive calls to match(). This is different to MATCH_LIMIT because
3434 that limits the total number of calls to match(), not all of which increase
3435 the depth of recursion. Limiting the recursion depth limits the amount of
3436 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3437 when PCRE is compiled, and changed at run time. A patch from Google adds
3438 this functionality to the C++ interface.
3439
3440 18. Changes to the handling of Unicode character properties:
3441
3442 (a) Updated the table to Unicode 4.1.0.
3443
3444 (b) Recognize characters that are not in the table as "Cn" (undefined).
3445
3446 (c) I revised the way the table is implemented to a much improved format
3447 which includes recognition of ranges. It now supports the ranges that
3448 are defined in UnicodeData.txt, and it also amalgamates other
3449 characters into ranges. This has reduced the number of entries in the
3450 table from around 16,000 to around 3,000, thus reducing its size
3451 considerably. I realized I did not need to use a tree structure after
3452 all - a binary chop search is just as efficient. Having reduced the
3453 number of entries, I extended their size from 6 bytes to 8 bytes to
3454 allow for more data.
3455
3456 (d) Added support for Unicode script names via properties such as \p{Han}.
3457
3458 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3459 matching that character.
3460
3461 20. When matching a repeated Unicode property with a minimum greater than zero,
3462 (for example \pL{2,}), PCRE could look past the end of the subject if it
3463 reached it while seeking the minimum number of characters. This could
3464 happen only if some of the characters were more than one byte long, because
3465 there is a check for at least the minimum number of bytes.
3466
3467 21. Refactored the implementation of \p and \P so as to be more general, to
3468 allow for more different types of property in future. This has changed the
3469 compiled form incompatibly. Anybody with saved compiled patterns that use
3470 \p or \P will have to recompile them.
3471
3472 22. Added "Any" and "L&" to the supported property types.
3473
3474 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3475 but give a compile time error if the value is greater than 0xff.
3476
3477 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3478 accidentally not being installed or uninstalled.
3479
3480 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3481 made were to insert the current release number. This seemed silly, because
3482 it made things harder for people building PCRE on systems that don't run
3483 "configure". I have turned pcre.h into a distributed file, no longer built
3484 by "configure", with the version identification directly included. There is
3485 no longer a pcre.h.in file.
3486
3487 However, this change necessitated a change to the pcre-config script as
3488 well. It is built from pcre-config.in, and one of the substitutions was the
3489 release number. I have updated configure.ac so that ./configure now finds
3490 the release number by grepping pcre.h.
3491
3492 26. Added the ability to run the tests under valgrind.
3493
3494
3495 Version 6.4 05-Sep-05
3496 ---------------------
3497
3498 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3499 "--" to be printed when multiple files were scanned, even when none of the
3500 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3501 consider it to be a bug, and have restored the previous behaviour.
3502
3503 2. A couple of code tidies to get rid of compiler warnings.
3504
3505 3. The pcretest program used to cheat by referring to symbols in the library
3506 whose names begin with _pcre_. These are internal symbols that are not
3507 really supposed to be visible externally, and in some environments it is
3508 possible to suppress them. The cheating is now confined to including
3509 certain files from the library's source, which is a bit cleaner.
3510
3511 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3512 file's purpose clearer.
3513
3514 5. Reorganized pcre_ucp_findchar().
3515
3516
3517 Version 6.3 15-Aug-05
3518 ---------------------
3519
3520 1. The file libpcre.pc.in did not have general read permission in the tarball.
3521
3522 2. There were some problems when building without C++ support:
3523
3524 (a) If C++ support was not built, "make install" and "make test" still
3525 tried to test it.
3526
3527 (b) There were problems when the value of CXX was explicitly set. Some
3528 changes have been made to try to fix these, and ...
3529
3530 (c) --disable-cpp can now be used to explicitly disable C++ support.
3531
3532 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3533 backslash in a target when C++ was disabled. This confuses some
3534 versions of "make", apparently. Using an intermediate variable solves
3535 this. (Same for CPP_LOBJ.)
3536
3537 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3538 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3539 necessary on certain architectures.
3540
3541 4. Added a setting of -export-symbols-regex to the link command to remove
3542 those symbols that are exported in the C sense, but actually are local
3543 within the library, and not documented. Their names all begin with
3544 "_pcre_". This is not a perfect job, because (a) we have to except some
3545 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3546 available (and never for static libraries). I have made a note to try to
3547 find a way round (a) in the future.
3548
3549
3550 Version 6.2 01-Aug-05
3551 ---------------------
3552
3553 1. There was no test for integer overflow of quantifier values. A construction
3554 such as {1111111111111111} would give undefined results. What is worse, if
3555 a minimum quantifier for a parenthesized subpattern overflowed and became
3556 negative, the calculation of the memory size went wrong. This could have
3557 led to memory overwriting.
3558
3559 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3560
3561 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3562 operating environments where this matters.
3563
3564 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3565 PCRE options from within the C++ wrapper.
3566
3567 5. Named capturing subpatterns were not being correctly counted when a pattern
3568 was compiled. This caused two problems: (a) If there were more than 100
3569 such subpatterns, the calculation of the memory needed for the whole
3570 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3571 back references of the form \12, where the number was greater than 9, were
3572 not recognized as back references, even though there were sufficient
3573 previous subpatterns.
3574
3575 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3576 versions of gcc, e.g. 2.95.4.
3577
3578
3579 Version 6.1 21-Jun-05
3580 ---------------------
3581
3582 1. There was one reference to the variable "posix" in pcretest.c that was not
3583 surrounded by "#if !defined NOPOSIX".
3584
3585 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3586 the cross-check on the old pcre_info() function, for the benefit of the
3587 cut-down version of PCRE that is currently imported into Exim.
3588
3589 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3590 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3591 patterns that start (?i) but I don't think that matters. The use of (?i) is
3592 just an example; this all applies to the other options as well.
3593
3594 4. Since libtool seems to echo the compile commands it is issuing, the output
3595 from "make" can be reduced a bit by putting "@" in front of each libtool
3596 compile command.
3597
3598 5. Patch from the folks at Google for configure.in to be a bit more thorough
3599 in checking for a suitable C++ installation before trying to compile the
3600 C++ stuff. This should fix a reported problem when a compiler was present,
3601 but no suitable headers.
3602
3603 6. The man pages all had just "PCRE" as their title. I have changed them to
3604 be the relevant file name. I have also arranged that these names are
3605 retained in the file doc/pcre.txt, which is a concatenation in text format
3606 of all the man pages except the little individual ones for each function.
3607
3608 7. The NON-UNIX-USE file had not been updated for the different set of source
3609 files that come with release 6. I also added a few comments about the C++
3610 wrapper.
3611
3612
3613 Version 6.0 07-Jun-05
3614 ---------------------
3615
3616 1. Some minor internal re-organization to help with my DFA experiments.
3617
3618 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3619 didn't matter for the library itself when fully configured, but did matter
3620 when compiling without UCP support, or within Exim, where the ucp files are
3621 not imported.
3622
3623 3. Refactoring of the library code to split up the various functions into
3624 different source modules. The addition of the new DFA matching code (see
3625 below) to a single monolithic source would have made it really too
3626 unwieldy, quite apart from causing all the code to be include in a
3627 statically linked application, when only some functions are used. This is
3628 relevant even without the DFA addition now that patterns can be compiled in
3629 one application and matched in another.
3630
3631 The downside of splitting up is that there have to be some external
3632 functions and data tables that are used internally in different modules of
3633 the library but which are not part of the API. These have all had their
3634 names changed to start with "_pcre_" so that they are unlikely to clash
3635 with other external names.
3636
3637 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3638 a different (DFA) algorithm. Although it is slower than the original
3639 function, it does have some advantages for certain types of matching
3640 problem.
3641
3642 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3643 including restarting after a partial match.
3644
3645 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3646 defined when compiling for Windows was sent to me. I have put it into the
3647 code, though I have no means of testing or verifying it.
3648
3649 7. Added the pcre_refcount() auxiliary function.
3650
3651 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3652 match before or at the first newline in the subject string. In pcretest,
3653 the /f option on a pattern can be used to set this.
3654
3655 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3656 would behave wrongly. This has been present in PCRE since release 4.0.
3657
3658 10. A number of changes to the pcregrep command:
3659
3660 (a) Refactored how -x works; insert ^(...)$ instead of setting
3661 PCRE_ANCHORED and checking the length, in preparation for adding
3662 something similar for -w.
3663
3664 (b) Added the -w (match as a word) option.
3665
3666 (c) Refactored the way lines are read and buffered so as to have more
3667 than one at a time available.
3668
3669 (d) Implemented a pcregrep test script.
3670
3671 (e) Added the -M (multiline match) option. This allows patterns to match
3672 over several lines of the subject. The buffering ensures that at least
3673 8K, or the rest of the document (whichever is the shorter) is available
3674 for matching (and similarly the previous 8K for lookbehind assertions).
3675
3676 (f) Changed the --help output so that it now says
3677
3678 -w, --word-regex(p)
3679
3680 instead of two lines, one with "regex" and the other with "regexp"
3681 because that confused at least one person since the short forms are the
3682 same. (This required a bit of code, as the output is generated
3683 automatically from a table. It wasn't just a text change.)
3684
3685 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3686 option but starts with a hyphen. Could be a pattern or a path name
3687 starting with a hyphen, for instance.
3688
3689 (h) "-" can be given as a file name to represent stdin.
3690
3691 (i) When file names are being printed, "(standard input)" is used for
3692 the standard input, for compatibility with GNU grep. Previously
3693 "<stdin>" was used.
3694
3695 (j) The option --label=xxx can be used to supply a name to be used for
3696 stdin when file names are being printed. There is no short form.
3697
3698 (k) Re-factored the options decoding logic because we are going to add
3699 two more options that take data. Such options can now be given in four
3700 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3701
3702 (l) Added the -A, -B, and -C options for requesting that lines of context
3703 around matches be printed.
3704
3705 (m) Added the -L option to print the names of files that do not contain
3706 any matching lines, that is, the complement of -l.
3707
3708 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3709 continue to scan other files.
3710
3711 (o) The -s option was incorrectly implemented. For compatibility with other
3712 greps, it now suppresses the error message for a non-existent or non-
3713 accessible file (but not the return code). There is a new option called
3714 -q that suppresses the output of matching lines, which was what -s was
3715 previously doing.
3716
3717 (p) Added --include and --exclude options to specify files for inclusion
3718 and exclusion when recursing.
3719
3720 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3721 Hopefully, it now does.
3722
3723 12. Missing cast in pcre_study().
3724
3725 13. Added an "uninstall" target to the makefile.
3726
3727 14. Replaced "extern" in the function prototypes in Makefile.in with
3728 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3729 world, but is set differently for Windows.
3730
3731 15. Added a second compiling function called pcre_compile2(). The only
3732 difference is that it has an extra argument, which is a pointer to an
3733 integer error code. When there is a compile-time failure, this is set
3734 non-zero, in addition to the error test pointer being set to point to an
3735 error message. The new argument may be NULL if no error number is required
3736 (but then you may as well call pcre_compile(), which is now just a
3737 wrapper). This facility is provided because some applications need a
3738 numeric error indication, but it has also enabled me to tidy up the way
3739 compile-time errors are handled in the POSIX wrapper.
3740
3741 16. Added VPATH=.libs to the makefile; this should help when building with one
3742 prefix path and installing with another. (Or so I'm told by someone who
3743 knows more about this stuff than I do.)
3744
3745 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3746 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3747 match everything, including newlines. This is not POSIX-compatible, but
3748 somebody wanted the feature. From pcretest it can be activated by using
3749 both the P and the s flags.
3750
3751 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3752
3753 19. libpcre.pc was being incorrectly installed as executable.
3754
3755 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3756 it now also looks for '\r' so that it will work unmodified on Windows.
3757
3758 21. Added Google's contributed C++ wrapper to the distribution.
3759
3760 22. Added some untidy missing memory free() calls in pcretest, to keep
3761 Electric Fence happy when testing.
3762
3763
3764
3765 Version 5.0 13-Sep-04
3766 ---------------------
3767
3768 1. Internal change: literal characters are no longer packed up into items
3769 containing multiple characters in a single byte-string. Each character
3770 is now matched using a separate opcode. However, there may be more than one
3771 byte in the character in UTF-8 mode.
3772
3773 2. The pcre_callout_block structure has two new fields: pattern_position and
3774 next_item_length. These contain the offset in the pattern to the next match
3775 item, and its length, respectively.
3776
3777 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3778 insertion of callouts before each pattern item. Added the /C option to
3779 pcretest to make use of this.
3780
3781 4. On the advice of a Windows user, the lines
3782
3783 #if defined(_WIN32) || defined(WIN32)
3784 _setmode( _fileno( stdout ), 0x8000 );
3785 #endif /* defined(_WIN32) || defined(WIN32) */
3786
3787 have been added to the source of pcretest. This apparently does useful
3788 magic in relation to line terminators.
3789
3790 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3791 for the benefit of those environments where the "b" makes a difference.
3792
3793 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3794 to know about it. I have put a hack into configure.in that adds in code
3795 to set GCC=yes if CC=icc. This seems to end up at a point in the
3796 generated configure script that is early enough to affect the setting of
3797 compiler options, which is what is needed, but I have no means of testing
3798 whether it really works. (The user who reported this had patched the
3799 generated configure script, which of course I cannot do.)
3800
3801 LATER: After change 22 below (new libtool files), the configure script
3802 seems to know about icc (and also ecc). Therefore, I have commented out
3803 this hack in configure.in.
3804
3805 7. Added support for pkg-config (2 patches were sent in).
3806
3807 8. Negated POSIX character classes that used a combination of internal tables
3808 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3809 [[:^ascii]]. Typically, they would match almost any characters. The other
3810 POSIX classes were not broken in this way.
3811
3812 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3813 to find the match, as PCRE was deluded into thinking that the match had to
3814 start at the start point or following a newline. The same bug applied to
3815 patterns with negative forward assertions or any backward assertions
3816 preceding ".*" at the start, unless the pattern required a fixed first
3817 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3818
3819 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3820 starting at the last subject character, bytes beyond the end of the subject
3821 string were read.
3822
3823 11. Renamed the variable "class" as "classbits" to make life easier for C++
3824 users. (Previously there was a macro definition, but it apparently wasn't
3825 enough.)
3826
3827 12. Added the new field "tables" to the extra data so that tables can be passed
3828 in at exec time, or the internal tables can be re-selected. This allows
3829 a compiled regex to be saved and re-used at a later time by a different
3830 program that might have everything at different addresses.
3831
3832 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3833 -R library as well as a -L library.
3834
3835 14. The debugging options of pcretest (-d on the command line or D on a
3836 pattern) showed incorrect output for anything following an extended class
3837 that contained multibyte characters and which was followed by a quantifier.
3838
3839 15. Added optional support for general category Unicode character properties
3840 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3841 support. It adds about 90K to the size of the library. The meanings of the
3842 inbuilt class escapes such as \d and \s have NOT been changed.
3843
3844 16. Updated pcredemo.c to include calls to free() to release the memory for the
3845 compiled pattern.
3846
3847 17. The generated file chartables.c was being created in the source directory
3848 instead of in the building directory. This caused the build to fail if the
3849 source directory was different from the building directory, and was
3850 read-only.
3851
3852 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3853 file. No doubt somebody will tell me if they don't make sense... Also added
3854 Dan Mooney's comments about building on OpenVMS.
3855
3856 19. Added support for partial matching via the PCRE_PARTIAL option for
3857 pcre_exec() and the \P data escape in pcretest.
3858
3859 20. Extended pcretest with 3 new pattern features:
3860
3861 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3862 write the compiled pattern to the file whose name is "rest-of-line".
3863 This is a straight binary dump of the data, with the saved pointer to
3864 the character tables forced to be NULL. The study data, if any, is
3865 written too. After writing, pcretest reads a new pattern.
3866
3867 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3868 compiled pattern from the given file. There must not be any
3869 occurrences of "<" in the file name (pretty unlikely); if there are,
3870 pcretest will instead treat the initial "<" as a pattern delimiter.
3871 After reading in the pattern, pcretest goes on to read data lines as
3872 usual.
3873
3874 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3875 and 16-bit fields in a compiled pattern, to simulate a pattern that
3876 was compiled on a host of opposite endianness.
3877
3878 21. The pcre-exec() function can now cope with patterns that were compiled on
3879 hosts of opposite endianness, with this restriction:
3880
3881 As for any compiled expression that is saved and used later, the tables
3882 pointer field cannot be preserved; the extra_data field in the arguments
3883 to pcre_exec() should be used to pass in a tables address if a value
3884 other than the default internal tables were used at compile time.
3885
3886 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3887 now diagnosed as an error. Previously, most of the time, a negative number
3888 would have been treated as zero, but if in addition "ovector" was passed as
3889 NULL, a crash could occur.
3890
3891 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3892 new versions from the libtool 1.5 distribution (the last one is a copy of
3893 a file called libtool.m4). This seems to have fixed the need to patch
3894 "configure" to support Darwin 1.3 (which I used to do). However, I still
3895 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3896 workstation).
3897
3898 24. Changed the PCRE licence to be the more standard "BSD" licence.
3899
3900
3901 Version 4.5 01-Dec-03
3902 ---------------------
3903
3904 1. There has been some re-arrangement of the code for the match() function so
3905 that it can be compiled in a version that does not call itself recursively.
3906 Instead, it keeps those local variables that need separate instances for
3907 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3908 needs to "recurse". Keeping track of where control must go is done by means
3909 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3910 hide most of the details from the main code, and operates only if
3911 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3912 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3913 operating.
3914
3915 To make it easier for callers to provide specially tailored get/free
3916 functions for this usage, two new functions, pcre_stack_malloc, and
3917 pcre_stack_free, are used. They are always called in strict stacking order,
3918 and the size of block requested is always the same.
3919
3920 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3921 PCRE has been compiled to use the stack or the heap for recursion. The
3922 -C option of pcretest uses this to show which version is compiled.
3923
3924 A new data escape \S, is added to pcretest; it causes the amounts of store
3925 obtained and freed by both kinds of malloc/free at match time to be added
3926 to the output.
3927
3928 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3929 what's available on my current Linux desktop machine.
3930
3931 3. When matching a UTF-8 string, the test for a valid string at the start has
3932 been extended. If start_offset is not zero, PCRE now checks that it points
3933 to a byte that is the start of a UTF-8 character. If not, it returns
3934 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3935 this is necessary because there may be backward assertions in the pattern.
3936 When matching the same subject several times, it may save resources to use
3937 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3938
3939 4. The code for checking the validity of UTF-8 strings has been tightened so
3940 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3941 containing "overlong sequences".
3942
3943 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3944 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3945 should have been "&", but it just so happened that all the cases this let
3946 through by mistake were picked up later in the function.
3947
3948 6. I had used a variable called "isblank" - this is a C99 function, causing
3949 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3950
3951 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3952 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3953 so the version is shown; (c) stop "make test" echoing "./RunTest".
3954
3955 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3956
3957 9. The prototype for memmove() for systems that don't have it was using
3958 size_t, but the inclusion of the header that defines size_t was later. I've
3959 moved the #includes for the C headers earlier to avoid this.
3960
3961 10. Added some adjustments to the code to make it easier to compiler on certain
3962 special systems:
3963
3964 (a) Some "const" qualifiers were missing.
3965 (b) Added the macro EXPORT before all exported functions; by default this
3966 is defined to be empty.
3967 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3968 that it reads its output file name as an argument instead of writing
3969 to the standard output and assuming this can be redirected.
3970
3971 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3972 class containing characters with values greater than 255, PCRE compilation
3973 went into a loop.
3974
3975 12. A recursive reference to a subpattern that was within another subpattern
3976 that had a minimum quantifier of zero caused PCRE to crash. For example,
3977 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3978 recursion. If the recursively-called subpattern itself had a zero repeat,
3979 that was OK.
3980
3981 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3982 buffer into which it was copied (for escape processing) was still set at
3983 1024, so long lines caused crashes.
3984
3985 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3986 "internal error: code overflow...". This applied to any character class
3987 that was followed by a possessive quantifier.
3988
3989 15. Modified the Makefile to add libpcre.la as a prerequisite for
3990 libpcreposix.la because I was told this is needed for a parallel build to
3991 work.
3992
3993 16. If a pattern that contained .* following optional items at the start was
3994 studied, the wrong optimizing data was generated, leading to matching
3995 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3996 matching string must start with a or b or c. The correct conclusion for
3997 this pattern is that a match can start with any character.
3998
3999
4000 Version 4.4 13-Aug-03
4001 ---------------------
4002
4003 1. In UTF-8 mode, a character class containing characters with values between
4004 127 and 255 was not handled correctly if the compiled pattern was studied.
4005 In fixing this, I have also improved the studying algorithm for such
4006 classes (slightly).
4007
4008 2. Three internal functions had redundant arguments passed to them. Removal
4009 might give a very teeny performance improvement.
4010
4011 3. Documentation bug: the value of the capture_top field in a callout is *one
4012 more than* the number of the hightest numbered captured substring.
4013
4014 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4015 in incorrectly linking with a previously installed version. They now link
4016 explicitly with libpcre.la.
4017
4018 5. configure.in no longer needs to recognize Cygwin specially.
4019
4020 6. A problem in pcre.in for Windows platforms is fixed.
4021
4022 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4023 pcretest, it used to include the size of the study block as part of its
4024 output. Unfortunately, the structure contains a field that has a different
4025 size on different hardware architectures. This meant that the tests that
4026 showed this size failed. As the block is currently always of a fixed size,
4027 this information isn't actually particularly useful in pcretest output, so
4028 I have just removed it.
4029
4030 8. Three pre-processor statements accidentally did not start in column 1.
4031 Sadly, there are *still* compilers around that complain, even though
4032 standard C has not required this for well over a decade. Sigh.
4033
4034 9. In pcretest, the code for checking callouts passed small integers in the
4035 callout_data field, which is a void * field. However, some picky compilers
4036 complained about the casts involved for this on 64-bit systems. Now
4037 pcretest passes the address of the small integer instead, which should get
4038 rid of the warnings.
4039
4040 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4041 both compile and run time, and gives an error if an invalid UTF-8 sequence
4042 is found. There is a option for disabling this check in cases where the
4043 string is known to be correct and/or the maximum performance is wanted.
4044
4045 11. In response to a bug report, I changed one line in Makefile.in from
4046
4047 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4048 to
4049 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4050
4051 to look similar to other lines, but I have no way of telling whether this
4052 is the right thing to do, as I do not use Windows. No doubt I'll get told
4053 if it's wrong...
4054
4055
4056 Version 4.3 21-May-03
4057 ---------------------
4058
4059 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4060 Makefile.
4061
4062 2. Some refactoring to improve the quality of the code:
4063
4064 (i) The utf8_table... variables are now declared "const".
4065
4066 (ii) The code for \cx, which used the "case flipping" table to upper case
4067 lower case letters, now just substracts 32. This is ASCII-specific,
4068 but the whole concept of \cx is ASCII-specific, so it seems
4069 reasonable.
4070
4071 (iii) PCRE was using its character types table to recognize decimal and
4072 hexadecimal digits in the pattern. This is silly, because it handles
4073 only 0-9, a-f, and A-F, but the character types table is locale-
4074 specific, which means strange things might happen. A private
4075 table is now used for this - though it costs 256 bytes, a table is
4076 much faster than multiple explicit tests. Of course, the standard
4077 character types table is still used for matching digits in subject
4078 strings against \d.
4079
4080 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4081 ending in _t are). So I've renamed it as ESC_tee.
4082
4083 3. The first argument for regexec() in the POSIX wrapper should have been
4084 defined as "const".
4085
4086 4. Changed pcretest to use malloc() for its buffers so that they can be
4087 Electric Fenced for debugging.
4088
4089 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4090 to read one or more bytes before the start of the subject string. Often this
4091 had no effect on PCRE's behaviour, but in some circumstances it could
4092 provoke a segmentation fault.
4093
4094 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4095 to try to read one or more bytes before the start of the subject string.
4096
4097 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4098 UTF-8 support could misbehave in various ways if the subject string
4099 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4100 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4101 back over UTF-8 characters.)
4102
4103
4104 Version 4.2 14-Apr-03
4105 ---------------------
4106
4107 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4108
4109 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4110 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4111 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4112 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4113 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4114 and BUILD_EXEEXT
4115 Note: automatic setting of the BUILD variables is not yet working
4116 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4117 compile-time but not at link-time
4118 [LINK]: use for linking executables only
4119 make different versions for Windows and non-Windows
4120 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4121 libraries
4122 [LINK_FOR_BUILD]: new variable
4123 [OBJEXT]: use throughout
4124 [EXEEXT]: use throughout
4125 <winshared>: new target
4126 <wininstall>: new target
4127 <dftables.o>: use native compiler
4128 <dftables>: use native linker
4129 <install>: handle Windows platform correctly
4130 <clean>: ditto
4131 <check>: ditto
4132 copy DLL to top builddir before testing
4133
4134 As part of these changes, -no-undefined was removed again. This was reported
4135 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4136 in any case.
4137
4138 3. Some tidies to get rid of compiler warnings:
4139
4140 . In the match_data structure, match_limit was an unsigned long int, whereas
4141 match_call_count was an int. I've made them both unsigned long ints.
4142
4143 . In pcretest the fact that a const uschar * doesn't automatically cast to
4144 a void * provoked a warning.
4145
4146 . Turning on some more compiler warnings threw up some "shadow" variables
4147 and a few more missing casts.
4148
4149 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4150 option, a class that contained a single character with a value between 128
4151 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4152
4153 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4154 option, a class that contained several characters, but with at least one
4155 whose value was between 128 and 255 caused PCRE to crash.
4156
4157
4158 Version 4.1 12-Mar-03
4159 ---------------------
4160
4161 1. Compiling with gcc -pedantic found a couple of places where casts were
4162 needed, and a string in dftables.c that was longer than standard compilers are
4163 required to support.
4164
4165 2. Compiling with Sun's compiler found a few more places where the code could
4166 be tidied up in order to avoid warnings.
4167
4168 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4169 first of these names is deprecated in the latest Autoconf in favour of the name
4170 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4171 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4172 analogy I have changed it to CFLAGS_FOR_BUILD.
4173
4174 4. Added -no-undefined to the linking command in the Makefile, because this is
4175 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4176 linking step for the pcreposix library.
4177
4178 5. PCRE was failing to diagnose the case of two named groups with the same
4179 name.
4180
4181 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4182 literal character that is needed in the subject for a match, and scans along to
4183 ensure that it is present before embarking on the full matching process. This
4184 saves time in cases of nested unlimited repeats that are never going to match.
4185 Problem: the scan can take a lot of time if the subject is very long (e.g.
4186 megabytes), thus penalizing straightforward matches. It is now done only if the
4187 amount of subject to be scanned is less than 1000 bytes.
4188
4189 7. A lesser problem with the same optimization is that it was recording the
4190 first character of an anchored pattern as "needed", thus provoking a search
4191 right along the subject, even when the first match of the pattern was going to
4192 fail. The "needed" character is now not set for anchored patterns, unless it
4193 follows something in the pattern that is of non-fixed length. Thus, it still
4194 fulfils its original purpose of finding quick non-matches in cases of nested
4195 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4196
4197
4198 Version 4.0 17-Feb-03
4199 ---------------------
4200
4201 1. If a comment in an extended regex that started immediately after a meta-item
4202 extended to the end of string, PCRE compiled incorrect data. This could lead to
4203 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4204
4205 2. Moved to autoconf 2.53 and libtool 1.4.2.
4206
4207 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4208 the special perltest8 script is no longer needed - all the tests can be run
4209 from a single perltest script.
4210
4211 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4212 by \s. It has now been removed in PCRE. This means it isn't recognized as
4213 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4214 class [:space:] *does* include VT, thereby creating a mess.
4215
4216 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4217 space and tab.
4218
4219 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4220 its new features into the main test script, reducing the number of scripts.
4221
4222 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4223 were backward compatible, and made the (?i) apply to the whole pattern, as if
4224 /i were given. Now it behaves more logically, and applies the option setting
4225 only to what follows. PCRE has been changed to follow suit. However, if it
4226 finds options settings right at the start of the pattern, it extracts them into
4227 the global options, as before. Thus, they show up in the info data.
4228
4229 8. Added support for the \Q...\E escape sequence. Characters in between are
4230 treated as literals. This is slightly different from Perl in that $ and @ are
4231 also handled as literals inside the quotes. In Perl, they will cause variable
4232 interpolation. Note the following examples:
4233
4234 Pattern PCRE matches Perl matches
4235
4236 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4237 \Qabc\$xyz\E abc\$xyz abc\$xyz
4238 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4239
4240 For compatibility with Perl, \Q...\E sequences are recognized inside character
4241 classes as well as outside them.
4242
4243 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4244 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4245 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4246 signed/unsigned warnings.
4247
4248 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4249 option for pcretest, so I've replaced it by a simple function that does just
4250 that job.
4251
4252 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4253 "pcregrep -".
4254
4255 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4256 Java package. This provides some syntactic sugar for simple cases of what my
4257 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4258 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4259 item, you can use this simplified notation. Note that only makes sense with
4260 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4261 greediness, whatever the setting of the PCRE_UNGREEDY option.
4262
4263 13. A change of greediness default within a pattern was not taking effect at
4264 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4265 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4266 was abstracted outside.
4267
4268 14. PCRE now supports the \G assertion. It is true when the current matching
4269 position is at the start point of the match. This differs from \A when the
4270 starting offset is non-zero. Used with the /g option of pcretest (or similar
4271 code), it works in the same way as it does for Perl's /g option. If all
4272 alternatives of a regex begin with \G, the expression is anchored to the start
4273 match position, and the "anchored" flag is set in the compiled expression.
4274
4275 15. Some bugs concerning the handling of certain option changes within patterns
4276 have been fixed. These applied to options other than (?ims). For example,
4277 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4278 been the other way round. Some of this was related to change 7 above.
4279
4280 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4281 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4282 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4283 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4284
4285 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4286 mode. Unlike ".", it always matches newline, whatever the setting of
4287 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4288 assertions. Perl allows it, but it doesn't (in general) work because it can't
4289 calculate the length of the lookbehind. At least, that's the case for Perl
4290 5.8.0 - I've been told they are going to document that it doesn't work in
4291 future.
4292
4293 18. Added an error diagnosis for escapes that PCRE does not support: these are
4294 \L, \l, \N, \P, \p, \U, \u, and \X.
4295
4296 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4297 reading past the end of the pattern in cases such as /[abcd/.
4298
4299 20. PCRE was getting more memory than necessary for patterns with classes that
4300 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4301
4302 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4303 compiling PCRE for use with Virtual Pascal.
4304
4305 22. Small fix to the Makefile to make it work properly if the build is done
4306 outside the source tree.
4307
4308 23. Added a new extension: a condition to go with recursion. If a conditional
4309 subpattern starts with (?(R) the "true" branch is used if recursion has
4310 happened, whereas the "false" branch is used only at the top level.
4311
4312 24. When there was a very long string of literal characters (over 255 bytes
4313 without UTF support, over 250 bytes with UTF support), the computation of how
4314 much memory was required could be incorrect, leading to segfaults or other
4315 strange effects.
4316
4317 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4318 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4319 there was a subsequent back reference to those brackets. This meant that, for
4320 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4321 possible to check for precisely this case. All we can do is abandon the
4322 optimization if .* occurs inside capturing brackets when there are any back
4323 references whatsoever. (See below for a better fix that came later.)
4324
4325 26. The handling of the optimization for finding the first character of a
4326 non-anchored pattern, and for finding a character that is required later in the
4327 match were failing in some cases. This didn't break the matching; it just
4328 failed to optimize when it could. The way this is done has been re-implemented.
4329
4330 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4331
4332 28. Added a new feature that provides some of the functionality that Perl
4333 provides with (?{...}). The facility is termed a "callout". The way it is done
4334 in PCRE is for the caller to provide an optional function, by setting
4335 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4336 global variable. By default it is unset, which disables all calling out. To get
4337 the function called, the regex must include (?C) at appropriate points. This
4338 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4339 This provides a means of identifying different callout points. When PCRE
4340 reaches such a point in the regex, if pcre_callout has been set, the external
4341 function is called. It is provided with data in a structure called
4342 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4343 matching continues; if it returns a non-zero value, the match at the current
4344 point fails. However, backtracking will occur if possible. [This was changed
4345 later and other features added - see item 49 below.]
4346
4347 29. pcretest is upgraded to test the callout functionality. It provides a
4348 callout function that displays information. By default, it shows the start of
4349 the match and the current position in the text. There are some new data escapes
4350 to vary what happens:
4351
4352 \C+ in addition, show current contents of captured substrings
4353 \C- do not supply a callout function
4354 \C!n return 1 when callout number n is reached
4355 \C!n!m return 1 when callout number n is reached for the mth time
4356
4357 30. If pcregrep was called with the -l option and just a single file name, it
4358 output "<stdin>" if a match was found, instead of the file name.
4359
4360 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4361 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4362 pcre_exec(). This saves a malloc/free per call. The default value of
4363 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4364 when configuring.
4365
4366 32. The default maximum size of a compiled pattern is 64K. There have been a
4367 few cases of people hitting this limit. The code now uses macros to handle the
4368 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4369 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4370 configuring. Tests 2 and 5 work only with 2-byte links because they output
4371 debugging information about compiled patterns.
4372
4373 33. Internal code re-arrangements:
4374
4375 (a) Moved the debugging function for printing out a compiled regex into
4376 its own source file (printint.c) and used #include to pull it into
4377 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4378 separate copies.
4379
4380 (b) Defined the list of op-code names for debugging as a macro in
4381 internal.h so that it is next to the definition of the opcodes.
4382
4383 (c) Defined a table of op-code lengths for simpler skipping along compiled
4384 code. This is again a macro in internal.h so that it is next to the
4385 definition of the opcodes.
4386
4387 34. Added support for recursive calls to individual subpatterns, along the
4388 lines of Robin Houston's patch (but implemented somewhat differently).
4389
4390 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4391 allow it to read and process whole directories in Win32. This code was
4392 contributed by Lionel Fourquaux; it has not been tested by me.
4393
4394 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4395 used to name a group. Names consist of alphanumerics and underscores, and must
4396 be unique. Back references use the syntax (?P=name) and recursive calls use
4397 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4398 numbers. The function pcre_fullinfo() can be used after compilation to extract
4399 a name/number map. There are three relevant calls:
4400
4401 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4402 PCRE_INFO_NAMECOUNT yields the number of entries
4403 PCRE_INFO_NAMETABLE yields a pointer to the map.
4404
4405 The map is a vector of fixed-size entries. The size of each entry depends on
4406 the length of the longest name used. The first two bytes of each entry are the
4407 group number, most significant byte first. There follows the corresponding
4408 name, zero terminated. The names are in alphabetical order.
4409
4410 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4411 case instead of 255. Making it the same both with and without UTF-8 support
4412 means that the same test output works with both.
4413
4414 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4415 calling malloc() with a zero argument.
4416
4417 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4418 optimization. I've improved things by keeping a bitmap of backreferences with
4419 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4420 fact referenced, the optimization can be applied. It is unlikely that a
4421 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4422 the match to follow \n) will appear inside brackets with a number greater than
4423 31, but if it does, any back reference > 31 suppresses the optimization.
4424
4425 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4426 of disabling numbered capturing parentheses. Any opening parenthesis that is
4427 not followed by ? behaves as if it were followed by ?: but named parentheses
4428 can still be used for capturing (and they will acquire numbers in the usual
4429 way).
4430
4431 41. Redesigned the return codes from the match() function into yes/no/error so
4432 that errors can be passed back from deep inside the nested calls. A malloc
4433 failure while inside a recursive subpattern call now causes the
4434 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4435
4436 42. It is now possible to set a limit on the number of times the match()
4437 function is called in a call to pcre_exec(). This facility makes it possible to
4438 limit the amount of recursion and backtracking, though not in a directly
4439 obvious way, because the match() function is used in a number of different
4440 circumstances. The count starts from zero for each position in the subject
4441 string (for non-anchored patterns). The default limit is, for compatibility, a
4442 large number, namely 10 000 000. You can change this in two ways:
4443
4444 (a) When configuring PCRE before making, you can use --with-match-limit=n
4445 to set a default value for the compiled library.
4446
4447 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4448 a different value is set. See 45 below.
4449
4450 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4451
4452 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4453 of things that can be changed at compile time. The first argument specifies
4454 what is wanted and the second points to where the information is to be placed.
4455 The current list of available information is:
4456
4457 PCRE_CONFIG_UTF8
4458
4459 The output is an integer that is set to one if UTF-8 support is available;
4460 otherwise it is set to zero.
4461
4462 PCRE_CONFIG_NEWLINE
4463
4464 The output is an integer that it set to the value of the code that is used for
4465 newline. It is either LF (10) or CR (13).
4466
4467 PCRE_CONFIG_LINK_SIZE
4468
4469 The output is an integer that contains the number of bytes used for internal
4470 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4471
4472 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4473
4474 The output is an integer that contains the threshold above which the POSIX
4475 interface uses malloc() for output vectors. See item 31 above.
4476
4477 PCRE_CONFIG_MATCH_LIMIT
4478
4479 The output is an unsigned integer that contains the default limit of the number
4480 of match() calls in a pcre_exec() execution. See 42 above.
4481
4482 44. pcretest has been upgraded by the addition of the -C option. This causes it
4483 to extract all the available output from the new pcre_config() function, and to
4484 output it. The program then exits immediately.
4485
4486 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4487 order to support additional features. One way would have been to define
4488 pcre_exec2() (for example) with extra arguments, but this would not have been
4489 extensible, and would also have required all calls to the original function to
4490 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4491 is used for passing in "extra" data from pcre_study().
4492
4493 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4494 contains the following fields:
4495
4496 flags a bitmap indicating which of the following fields are set
4497 study_data opaque data from pcre_study()
4498 match_limit a way of specifying a limit on match() calls for a specific
4499 call to pcre_exec()
4500 callout_data data for callouts (see 49 below)
4501
4502 The flag bits are also defined in pcre.h, and are
4503
4504 PCRE_EXTRA_STUDY_DATA
4505 PCRE_EXTRA_MATCH_LIMIT
4506 PCRE_EXTRA_CALLOUT_DATA
4507
4508 The pcre_study() function now returns one of these new pcre_extra blocks, with
4509 the actual study data pointed to by the study_data field, and the
4510 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4511 before. That is, this change is entirely upwards-compatible and requires no
4512 change to existing code.
4513
4514 If you want to pass in additional data to pcre_exec(), you can either place it
4515 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4516 block.
4517
4518 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4519 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4520 times with different match limits, until it finds the minimum value needed for
4521 pcre_exec() to complete. The value is then output. This can be instructive; for
4522 most simple matches the number is quite small, but for pathological cases it
4523 gets very large very quickly.
4524
4525 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4526 returns the size of the data block pointed to by the study_data field in a
4527 pcre_extra block, that is, the value that was passed as the argument to
4528 pcre_malloc() when PCRE was getting memory in which to place the information
4529 created by pcre_study(). The fourth argument should point to a size_t variable.
4530 pcretest has been extended so that this information is shown after a successful
4531 pcre_study() call when information about the compiled regex is being displayed.
4532
4533 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4534 because what follows is always an absolute path. (Later: it turns out that this
4535 is more than cosmetic for MinGW, because it doesn't like empty path
4536 components.)
4537
4538 49. Some changes have been made to the callout feature (see 28 above):
4539
4540 (i) A callout function now has three choices for what it returns:
4541
4542 0 => success, carry on matching
4543 > 0 => failure at this point, but backtrack if possible
4544 < 0 => serious error, return this value from pcre_exec()
4545
4546 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4547 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4548 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4549 use by callout functions. It will never be used by PCRE itself.
4550
4551 (ii) The pcre_extra structure (see 45 above) has a void * field called
4552 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4553 pcre_callout_block structure has a field of the same name. The contents of
4554 the field passed in the pcre_extra structure are passed to the callout
4555 function in the corresponding field in the callout block. This makes it
4556 easier to use the same callout-containing regex from multiple threads. For
4557 testing, the pcretest program has a new data escape
4558
4559 \C*n pass the number n (may be negative) as callout_data
4560
4561 If the callout function in pcretest receives a non-zero value as
4562 callout_data, it returns that value.
4563
4564 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4565 there were some redundant $(CFLAGS) in commands that are now specified as
4566 $(LINK), which already includes $(CFLAGS).
4567
4568 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4569 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4570 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4571 one-byte characters throughout. Note that case-insensitive matching applies
4572 only to characters whose values are less than 256. PCRE doesn't support the
4573 notion of cases for higher-valued characters.
4574
4575 (i) A character class whose characters are all within 0-255 is handled as
4576 a bit map, and the map is inverted for negative classes. Previously, a
4577 character > 255 always failed to match such a class; however it should
4578 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4579
4580 (ii) A negated character class with a single character < 255 is coded as
4581 "not this character" (OP_NOT). This wasn't working properly when the test
4582 character was multibyte, either singly or repeated.
4583
4584 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4585 mode, for example: \x{100}{2,3}.
4586
4587 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4588 singly or repeated) now correctly test multibyte characters. However,
4589 PCRE doesn't recognize any characters with values greater than 255 as
4590 digits, spaces, or word characters. Such characters always match \D, \S,
4591 and \W, and never match \d, \s, or \w.
4592
4593 (v) Classes may now contain characters and character ranges with values
4594 greater than 255. For example: [ab\x{100}-\x{400}].
4595
4596 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4597 PCRE in UTF-8 mode.
4598
4599 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4600 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4601 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4602 value.)
4603
4604 53. The single man page has become too large. I have therefore split it up into
4605 a number of separate man pages. These also give rise to individual HTML pages;
4606 these are now put in a separate directory, and there is an index.html page that
4607 lists them all. Some hyperlinking between the pages has been installed.
4608
4609 54. Added convenience functions for handling named capturing parentheses.
4610
4611 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4612 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4613 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4614 are faulted.
4615
4616 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4617 calling configure. These values are used when compiling the dftables.c program
4618 which is run to generate the source of the default character tables. They
4619 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4620 you will need to set these values.
4621
4622 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4623
4624
4625 Version 3.9 02-Jan-02
4626 ---------------------
4627
4628 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4629
4630 2. If --disable-static was given, the building process failed when trying to
4631 build pcretest and pcregrep. (For some reason it was using libtool to compile
4632 them, which is not right, as they aren't part of the library.)
4633
4634
4635 Version 3.8 18-Dec-01
4636 ---------------------
4637
4638 1. The experimental UTF-8 code was completely screwed up. It was packing the
4639 bytes in the wrong order. How dumb can you get?
4640
4641
4642 Version 3.7 29-Oct-01
4643 ---------------------
4644
4645 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4646 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4647 this didn't happen under Solaris 8, where I normally test things.
4648
4649 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4650 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4651 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4652 this fix an hour or so after the initial 3.7 release.)
4653
4654
4655 Version 3.6 23-Oct-01
4656 ---------------------
4657
4658 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4659 offsets passed as NULL with zero offset count.
4660
4661 2. The config.guess and config.sub files had not been updated when I moved to
4662 the latest autoconf.
4663
4664
4665 Version 3.5 15-Aug-01
4666 ---------------------
4667
4668 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4669 had been forgotten.
4670
4671 2. By using declared but undefined structures, we can avoid using "void"
4672 definitions in pcre.h while keeping the internal definitions of the structures
4673 private.
4674
4675 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4676 user point of view, this means that both static and shared libraries are built
4677 by default, but this can be individually controlled. More of the work of
4678 handling this static/shared cases is now inside libtool instead of PCRE's make
4679 file.
4680
4681 4. The pcretest utility is now installed along with pcregrep because it is
4682 useful for users (to test regexs) and by doing this, it automatically gets
4683 relinked by libtool. The documentation has been turned into a man page, so
4684 there are now .1, .txt, and .html versions in /doc.
4685
4686 5. Upgrades to pcregrep:
4687 (i) Added long-form option names like gnu grep.
4688 (ii) Added --help to list all options with an explanatory phrase.
4689 (iii) Added -r, --recursive to recurse into sub-directories.
4690 (iv) Added -f, --file to read patterns from a file.
4691
4692 6. pcre_exec() was referring to its "code" argument before testing that
4693 argument for NULL (and giving an error if it was NULL).
4694
4695 7. Upgraded Makefile.in to allow for compiling in a different directory from
4696 the source directory.
4697
4698 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4699 options bits, the pointer it was passed was to an int instead of to an unsigned
4700 long int. This mattered only on 64-bit systems.
4701
4702 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4703 generated) instead of pcre.in, which it its source. Also made the same change
4704 in several of the .c files.
4705
4706 10. A new release of gcc defines printf() as a macro, which broke pcretest
4707 because it had an ifdef in the middle of a string argument for printf(). Fixed
4708 by using separate calls to printf().
4709
4710 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4711 script, to force use of CR or LF instead of \n in the source. On non-Unix
4712 systems, the value can be set in config.h.
4713
4714 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4715 absolute limit. Changed the text of the error message to make this clear, and
4716 likewise updated the man page.
4717
4718 13. The limit of 99 on the number of capturing subpatterns has been removed.
4719 The new limit is 65535, which I hope will not be a "real" limit.
4720
4721
4722 Version 3.4 22-Aug-00
4723 ---------------------
4724
4725 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4726
4727 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4728
4729
4730 Version 3.3 01-Aug-00
4731 ---------------------
4732
4733 1. If an octal character was given, but the value was greater than \377, it
4734 was not getting masked to the least significant bits, as documented. This could
4735 lead to crashes in some systems.
4736
4737 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4738 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4739
4740 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4741 These just pass their arguments on to (pcre_free)(), but they are provided
4742 because some uses of PCRE bind it to non-C systems that can call its functions,
4743 but cannot call free() or pcre_free() directly.
4744
4745 4. Add "make test" as a synonym for "make check". Corrected some comments in
4746 the Makefile.
4747
4748 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4749 Makefile.
4750
4751 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4752 command called pgrep for grepping around the active processes.
4753
4754 7. Added the beginnings of support for UTF-8 character strings.
4755
4756 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4757 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4758 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4759 out for the ar command.)
4760
4761
4762 Version 3.2 12-May-00
4763 ---------------------
4764
4765 This is purely a bug fixing release.
4766
4767 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4768 of ZA. This was just one example of several cases that could provoke this bug,
4769 which was introduced by change 9 of version 2.00. The code for breaking
4770 infinite loops after an iteration that matches an empty string was't working
4771 correctly.
4772
4773 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4774 when matched against abbab (for example). After matching an empty string, it
4775 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4776 caused it to match further down the string than it should.
4777
4778 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4779 was there because it doesn't seem to be needed, and it causes trouble on some
4780 systems, as it is not a Standard C header. It has been removed.
4781
4782 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4783 were reported on the Macintosh. The changes were from
4784
4785 while ((c = *(++ptr)) != 0 && c != '\n');
4786 to
4787 while ((c = *(++ptr)) != 0 && c != '\n') ;
4788
4789 Totally extraordinary, but if that's what it takes...
4790
4791 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4792 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4793 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4794 assumes the way PCRE uses memmove() (always moving upwards).
4795
4796 6. PCRE is being used in one environment where strchr() is not available. There
4797 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4798 faster code anyway.
4799
4800
4801 Version 3.1 09-Feb-00
4802 ---------------------
4803
4804 The only change in this release is the fixing of some bugs in Makefile.in for
4805 the "install" target:
4806
4807 (1) It was failing to install pcreposix.h.
4808
4809 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4810
4811
4812 Version 3.0 01-Feb-00
4813 ---------------------
4814
4815 1. Add support for the /+ modifier to perltest (to output $` like it does in
4816 pcretest).
4817
4818 2. Add support for the /g modifier to perltest.
4819
4820 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4821 matches null strings.
4822
4823 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4824 pattern. Perl treats empty patterns specially - it reuses the most recent
4825 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4826 effect.
4827
4828 5. The POSIX interface was broken in that it was just handing over the POSIX
4829 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4830 required a bigger vector, with some working space on the end. This means that
4831 the POSIX wrapper now has to get and free some memory, and copy the results.
4832
4833 6. Added some simple autoconf support, placing the test data and the
4834 documentation in separate directories, re-organizing some of the
4835 information files, and making it build pcre-config (a GNU standard). Also added
4836 libtool support for building PCRE as a shared library, which is now the
4837 default.
4838
4839 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4840 09 are not valid octal constants. Single digits will be used for minor values
4841 less than 10.
4842
4843 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4844 existing programs that set these in the POSIX interface can use PCRE without
4845 modification.
4846
4847 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4848 return all that pcre_info() returns, plus additional data. The pcre_info()
4849 function is retained for compatibility, but is considered to be obsolete.
4850
4851 10. Added experimental recursion feature (?R) to handle one common case that
4852 Perl 5.6 will be able to do with (?p{...}).
4853
4854 11. Added support for POSIX character classes like [:alpha:], which Perl is
4855 adopting.
4856
4857
4858 Version 2.08 31-Aug-99
4859 ----------------------
4860
4861 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4862 trying to match at the startoffset position, but instead was moving forward to
4863 the next newline as if a previous match had failed.
4864
4865 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4866 and could get into a loop if a null string was matched other than at the start
4867 of the subject.
4868
4869 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4870 be distinguished at compile time, and for completeness also added PCRE_DATE.
4871
4872 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4873 in GnuWin32 environments.
4874
4875
4876 Version 2.07 29-Jul-99
4877 ----------------------
4878
4879 1. The documentation is now supplied in plain text form and HTML as well as in
4880 the form of man page sources.
4881
4882 2. C++ compilers don't like assigning (void *) values to other pointer types.
4883 In particular this affects malloc(). Although there is no problem in Standard
4884 C, I've put in casts to keep C++ compilers happy.
4885
4886 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4887 should be (const char *).
4888
4889 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4890 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4891 However, I haven't made this a standard facility. The documentation doesn't
4892 mention it, and the Makefile doesn't support it.
4893
4894 5. The Makefile now contains an "install" target, with editable destinations at
4895 the top of the file. The pcretest program is not installed.
4896
4897 6. pgrep -V now gives the PCRE version number and date.
4898
4899 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4900 causing the entire string to be ignored, instead of just the last character.
4901
4902 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4903 non-matching string, it can take a very, very long time, even for strings of
4904 quite modest length, because of the nested recursion. PCRE now does better in
4905 some of these cases. It does this by remembering the last required literal
4906 character in the pattern, and pre-searching the subject to ensure it is present
4907 before running the real match. In other words, it applies a heuristic to detect
4908 some types of certain failure quickly, and in the above example, if presented
4909 with a string that has no trailing " it gives "no match" very quickly.
4910
4911 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4912 other alternatives are tried instead.
4913
4914
4915 Version 2.06 09-Jun-99
4916 ----------------------
4917
4918 1. Change pcretest's output for amount of store used to show just the code
4919 space, because the remainder (the data block) varies in size between 32-bit and
4920 64-bit systems.
4921
4922 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4923 start matching at. This allows lookbehinds to work when searching for multiple
4924 occurrences in a string.
4925
4926 3. Added additional options to pcretest for testing multiple occurrences:
4927
4928 /+ outputs the rest of the string that follows a match
4929 /g loops for multiple occurrences, using the new startoffset argument
4930 /G loops for multiple occurrences by passing an incremented pointer
4931
4932 4. PCRE wasn't doing the "first character" optimization for patterns starting
4933 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4934 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4935 the letter 'x'. On long subject strings, this gives a significant speed-up.
4936
4937
4938 Version 2.05 21-Apr-99
4939 ----------------------
4940
4941 1. Changed the type of magic_number from int to long int so that it works
4942 properly on 16-bit systems.
4943
4944 2. Fixed a bug which caused patterns starting with .* not to work correctly
4945 when the subject string contained newline characters. PCRE was assuming
4946 anchoring for such patterns in all cases, which is not correct because .* will
4947 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4948 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4949 must be retried after every newline in the subject.
4950
4951
4952 Version 2.04 18-Feb-99
4953 ----------------------
4954
4955 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4956 computation of the store needed to hold the pattern was incorrect (too large).
4957 If such patterns were nested a few deep, this could multiply and become a real
4958 problem.
4959
4960 2. Added /M option to pcretest to show the memory requirement of a specific
4961 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4962
4963 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4964 compiled in such a way that the backtracking after subsequent failure was
4965 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4966 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4967
4968
4969 Version 2.03 02-Feb-99
4970 ----------------------
4971
4972 1. Fixed typo and small mistake in man page.
4973
4974 2. Added 4th condition (GPL supersedes if conflict) and created separate
4975 LICENCE file containing the conditions.
4976
4977 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4978 Perl, that is the internal \ allows the delimiter to be included in the
4979 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4980 the final delimiter, add \ to the end of the pattern (to test the error).
4981
4982 4. Added the convenience functions for extracting substrings after a successful
4983 match. Updated pcretest to make it able to test these functions.
4984
4985
4986 Version 2.02 14-Jan-99
4987 ----------------------
4988
4989 1. Initialized the working variables associated with each extraction so that
4990 their saving and restoring doesn't refer to uninitialized store.
4991
4992 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4993 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4994 fix the problem.
4995
4996 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4997 calls, and wasn't printing the correct value for compiling calls. Increased the
4998 default value of LOOPREPEAT, and the number of significant figures in the
4999 times.
5000
5001 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5002
5003 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5004 a building problem on Windows NT with a FAT file system.
5005
5006
5007 Version 2.01 21-Oct-98
5008 ----------------------
5009
5010 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5011 to character tables built by pcre_maketables() in the current locale. If NULL
5012 is passed, the default tables are used.
5013
5014
5015 Version 2.00 24-Sep-98
5016 ----------------------
5017
5018 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5019 it any more.
5020
5021 2. Allow quantification of (?>) groups, and make it work correctly.
5022
5023 3. The first character computation wasn't working for (?>) groups.
5024
5025 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5026 end of the subject) and add 5.005's \z, which really does match only at the
5027 very end of the subject.
5028
5029 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5030
5031 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5032 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5033 localized options. All options to pcre_study() were also removed.
5034
5035 7. Add other new features from 5.005:
5036
5037 $(?<= positive lookbehind
5038 $(?<! negative lookbehind
5039 (?imsx-imsx) added the unsetting capability
5040 such a setting is global if at outer level; local otherwise
5041 (?imsx-imsx:) non-capturing groups with option setting
5042 (?(cond)re|re) conditional pattern matching
5043
5044 A backreference to itself in a repeated group matches the previous
5045 captured string.
5046
5047 8. General tidying up of studying (both automatic and via "study")
5048 consequential on the addition of new assertions.
5049
5050 9. As in 5.005, unlimited repeated groups that could match an empty substring
5051 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5052 runtime if any iteration does actually match an empty substring.
5053
5054 10. Include the RunTest script in the distribution.
5055
5056 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5057 discrepancies, some of which were old and were also with respect to 5.004. They
5058 have now been fixed.
5059
5060
5061 Version 1.09 28-Apr-98
5062 ----------------------
5063
5064 1. A negated single character class followed by a quantifier with a minimum
5065 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5066 program crashes, or just wrong answers. This did not apply to negated classes
5067 containing more than one character, or to minima other than one.
5068
5069
5070 Version 1.08 27-Mar-98
5071 ----------------------
5072
5073 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5074
5075 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5076 latter must appear before anything that relies on it in the pattern.
5077
5078
5079 Version 1.07 16-Feb-98
5080 ----------------------
5081
5082 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5083 repeat of a potentially empty string).
5084
5085
5086 Version 1.06 23-Jan-98
5087 ----------------------
5088
5089 1. Added Markus Oberhumer's little patches for C++.
5090
5091 2. Literal strings longer than 255 characters were broken.
5092
5093
5094 Version 1.05 23-Dec-97
5095 ----------------------
5096
5097 1. Negated character classes containing more than one character were failing if
5098 PCRE_CASELESS was set at run time.
5099
5100
5101 Version 1.04 19-Dec-97
5102 ----------------------
5103
5104 1. Corrected the man page, where some "const" qualifiers had been omitted.
5105
5106 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5107 input syntax.
5108
5109 3. Fixed memory leak which occurred when a regex with back references was
5110 matched with an offsets vector that wasn't big enough. The temporary memory
5111 that is used in this case wasn't being freed if the match failed.
5112
5113 4. Tidied pcretest to ensure it frees memory that it gets.
5114
5115 5. Temporary memory was being obtained in the case where the passed offsets
5116 vector was exactly big enough.
5117
5118 6. Corrected definition of offsetof() from change 5 below.
5119
5120 7. I had screwed up change 6 below and broken the rules for the use of
5121 setjmp(). Now fixed.
5122
5123
5124 Version 1.03 18-Dec-97
5125 ----------------------
5126
5127 1. A erroneous regex with a missing opening parenthesis was correctly
5128 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5129 on some systems.
5130
5131 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5132 it was reported that one broken compiler failed on the former because "code" is
5133 also an independent variable.
5134
5135 3. The erroneous regex a[]b caused an array overrun reference.
5136
5137 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5138 fail on data ending with that character. (It was going on too far, and checking
5139 the next character, typically a binary zero.) This was specific to the
5140 optimized code for single-character negative classes.
5141
5142 5. Added a contributed patch from the TIN world which does the following:
5143
5144 + Add an undef for memmove, in case the the system defines a macro for it.
5145
5146 + Add a definition of offsetof(), in case there isn't one. (I don't know
5147 the reason behind this - offsetof() is part of the ANSI standard - but
5148 it does no harm).
5149
5150 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5151 most of the places where whitespace preceded '#'. I have given up and
5152 allowed the remaining 2 cases to be at the margin.
5153
5154 + Rename some variables in pcre to eliminate shadowing. This seems very
5155 pedantic, but does no harm, of course.
5156
5157 6. Moved the call to setjmp() into its own function, to get rid of warnings
5158 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5159
5160 7. Constructs such as \d{8,} were compiling into the equivalent of
5161 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5162 outcome, but in this particular case used more store than had been allocated,
5163 which caused the bug to be discovered because it threw up an internal error.
5164
5165 8. The debugging code in both pcre and pcretest for outputting the compiled
5166 form of a regex was going wrong in the case of back references followed by
5167 curly-bracketed repeats.
5168
5169
5170 Version 1.02 12-Dec-97
5171 ----------------------
5172
5173 1. Typos in pcre.3 and comments in the source fixed.
5174
5175 2. Applied a contributed patch to get rid of places where it used to remove
5176 'const' from variables, and fixed some signed/unsigned and uninitialized
5177 variable warnings.
5178
5179 3. Added the "runtest" target to Makefile.
5180
5181 4. Set default compiler flag to -O2 rather than just -O.
5182
5183
5184 Version 1.01 19-Nov-97
5185 ----------------------
5186
5187 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5188 like /([ab]*)*/, that is, for classes with more than one character in them.
5189
5190 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5191 as /((?>a*))*/ (a PCRE_EXTRA facility).
5192
5193
5194 Version 1.00 18-Nov-97
5195 ----------------------
5196
5197 1. Added compile-time macros to support systems such as SunOS4 which don't have
5198 memmove() or strerror() but have other things that can be used instead.
5199
5200 2. Arranged that "make clean" removes the executables.
5201
5202
5203 Version 0.99 27-Oct-97
5204 ----------------------
5205
5206 1. Fixed bug in code for optimizing classes with only one character. It was
5207 initializing a 32-byte map regardless, which could cause it to run off the end
5208 of the memory it had got.
5209
5210 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5211
5212
5213 Version 0.98 22-Oct-97
5214 ----------------------
5215
5216 1. Fixed bug in code for handling temporary memory usage when there are more
5217 back references than supplied space in the ovector. This could cause segfaults.
5218
5219
5220 Version 0.97 21-Oct-97
5221 ----------------------
5222
5223 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5224
5225 2. Optimized negated single characters not to use a bit map.
5226
5227 3. Brought error texts together as macro definitions; clarified some of them;
5228 fixed one that was wrong - it said "range out of order" when it meant "invalid
5229 escape sequence".
5230
5231 4. Changed some char * arguments to const char *.
5232
5233 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5234
5235 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5236 pcretest.
5237
5238
5239 Version 0.96 16-Oct-97
5240 ----------------------
5241
5242 1. Added a simple "pgrep" utility to the distribution.
5243
5244 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5245 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5246 where "ddd" means "one or more decimal digits".
5247
5248 3. Fixed serious bug. If a pattern had a back reference, but the call to
5249 pcre_exec() didn't supply a large enough ovector to record the related
5250 identifying subpattern, the match always failed. PCRE now remembers the number
5251 of the largest back reference, and gets some temporary memory in which to save
5252 the offsets during matching if necessary, in order to ensure that
5253 backreferences always work.
5254
5255 4. Increased the compatibility with Perl in a number of ways:
5256
5257 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5258 to request this handling. The option can be set at compile or exec time.
5259
5260 (b) $ matches before a terminating newline by default; an option
5261 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5262 mode). The option can be set at compile or exec time.
5263
5264 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5265 the same as Perl's. If the decimal number it represents is less than 10
5266 or there aren't that many previous left capturing parentheses, an octal
5267 escape is read. Inside a character class, it's always an octal escape,
5268 even if it is a single digit.
5269
5270 (d) An escaped but undefined alphabetic character is taken as a literal,
5271 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5272 escapes.
5273
5274 (e) {0} is now permitted. (The previous item is removed from the compiled
5275 pattern).
5276
5277 5. Changed all the names of code files so that the basic parts are no longer
5278 than 10 characters, and abolished the teeny "globals.c" file.
5279
5280 6. Changed the handling of character classes; they are now done with a 32-byte
5281 bit map always.
5282
5283 7. Added the -d and /D options to pcretest to make it possible to look at the
5284 internals of compilation without having to recompile pcre.
5285
5286
5287 Version 0.95 23-Sep-97
5288 ----------------------
5289
5290 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5291 \x20 at the start of a run of normal characters. These were being treated as
5292 real characters, instead of the source characters being re-checked.
5293
5294
5295 Version 0.94 18-Sep-97
5296 ----------------------
5297
5298 1. The functions are now thread-safe, with the caveat that the global variables
5299 containing pointers to malloc() and free() or alternative functions are the
5300 same for all threads.
5301
5302 2. Get pcre_study() to generate a bitmap of initial characters for non-
5303 anchored patterns when this is possible, and use it if passed to pcre_exec().
5304
5305
5306 Version 0.93 15-Sep-97
5307 ----------------------
5308
5309 1. /(b)|(:+)/ was computing an incorrect first character.
5310
5311 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5312 but not actually doing anything yet.
5313
5314 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5315 as Perl does (e.g. [-az] or [az-]).
5316
5317 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5318 all possible positions.
5319
5320 5. Split up into different modules to avoid including unneeded functions in a
5321 compiled binary. However, compile and exec are still in one module. The "study"
5322 function is split off.
5323
5324 6. The character tables are now in a separate module whose source is generated
5325 by an auxiliary program - but can then be edited by hand if required. There are
5326 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5327 toupper() in the code.
5328
5329 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5330 make them global. Abolish the function for setting them, as the caller can now
5331 set them directly.
5332
5333
5334 Version 0.92 11-Sep-97
5335 ----------------------
5336
5337 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5338 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5339
5340 2. Caseless matching was not working in character classes if the characters in
5341 the pattern were in upper case.
5342
5343 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5344
5345 4. Make PCRE_ANCHORED public and accept as a compile option.
5346
5347 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5348 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5349 pass them.
5350
5351 6. Give an error if bad option bits passed at compile or run time.
5352
5353 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5354 pcretest to cause it to pass that flag.
5355
5356 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5357 options, and the first character, if set.
5358
5359 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5360
5361
5362 Version 0.91 10-Sep-97
5363 ----------------------
5364
5365 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5366 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5367
5368 2. PCRE was looping on encountering an indefinitely repeated back reference to
5369 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5370 Perl does - treats the match as successful.
5371
5372 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5