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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1418 - (show annotations)
Fri Dec 27 12:23:25 2013 UTC (5 years, 11 months ago) by ph10
File size: 258872 byte(s)
Error occurred while calculating annotation data.
Fix pcretest's handling of patterns when \K in an assertion sets the start of a
match past the end of the match.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.35-RC1 xx-xxxx-201x
5 -----------------------------
6
7 1. A new flag is set, when property checks are present in an XCLASS.
8 When this flag is not set, PCRE can perform certain optimizations
9 such as studying these XCLASS-es.
10
11 2. The auto-possessification of character sets were improved: a normal
12 and an extended character set can be compared now. Furthermore
13 the JIT compiler optimizes more character set checks.
14
15 3. Got rid of some compiler warnings for potentially uninitialized variables
16 that show up only when compiled with -O2.
17
18 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
19 of a match later then the end of the match. The pcretest program was not
20 handling the case sensibly - it was outputting from the start to the next
21 binary zero. It now reports this situation in a message, and outputs the
22 text from the end to the start.
23
24
25 Version 8.34 15-December-2013
26 -----------------------------
27
28 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
29 executable memory. Patch inspired by Carsten Klein.
30
31 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
32 this macro is never tested and has no effect, because the work to support
33 coverage involves only compiling and linking options and special targets in
34 the Makefile. The comment in config.h implied that defining the macro would
35 enable coverage support, which is totally false. There was also support for
36 setting this macro in the CMake files (my fault, I just copied it from
37 configure). SUPPORT_GCOV has now been removed.
38
39 3. Make a small performance improvement in strlen16() and strlen32() in
40 pcretest.
41
42 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
43 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
44 these cases). There was also one in pcretest.c.
45
46 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
47
48 6. In UTF mode, the code for checking whether a group could match an empty
49 string (which is used for indefinitely repeated groups to allow for
50 breaking an infinite loop) was broken when the group contained a repeated
51 negated single-character class with a character that occupied more than one
52 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
53 UTF-8 mode). The effect was undefined: the group might or might not be
54 deemed as matching an empty string, or the program might have crashed.
55
56 7. The code for checking whether a group could match an empty string was not
57 recognizing that \h, \H, \v, \V, and \R must match a character.
58
59 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
60 an empty string. If it can, pcretest shows this in its information output.
61
62 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
63 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
64 matched by pcre_exec() without using JIT:
65
66 (a) If the rest of the pattern did not match after a maximal run of
67 grapheme clusters, the code for backing up to try with fewer of them
68 did not always back up over a full grapheme when characters that do not
69 have the modifier quality were involved, e.g. Hangul syllables.
70
71 (b) If the match point in a subject started with modifier character, and
72 there was no match, the code could incorrectly back up beyond the match
73 point, and potentially beyond the first character in the subject,
74 leading to a segfault or an incorrect match result.
75
76 10. A conditional group with an assertion condition could lead to PCRE
77 recording an incorrect first data item for a match if no other first data
78 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
79 first data item, and therefore matched "ca" after "c" instead of at the
80 start.
81
82 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
83 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
84
85 12. The source of pcregrep now includes z/OS-specific code so that it can be
86 compiled for z/OS as part of the special z/OS distribution.
87
88 13. Added the -T and -TM options to pcretest.
89
90 14. The code in pcre_compile.c for creating the table of named capturing groups
91 has been refactored. Instead of creating the table dynamically during the
92 actual compiling pass, the information is remembered during the pre-compile
93 pass (on the stack unless there are more than 20 named groups, in which
94 case malloc() is used) and the whole table is created before the actual
95 compile happens. This has simplified the code (it is now nearly 150 lines
96 shorter) and prepared the way for better handling of references to groups
97 with duplicate names.
98
99 15. A back reference to a named subpattern when there is more than one of the
100 same name now checks them in the order in which they appear in the pattern.
101 The first one that is set is used for the reference. Previously only the
102 first one was inspected. This change makes PCRE more compatible with Perl.
103
104 16. Unicode character properties were updated from Unicode 6.3.0.
105
106 17. The compile-time code for auto-possessification has been refactored, based
107 on a patch by Zoltan Herczeg. It now happens after instead of during
108 compilation. The code is cleaner, and more cases are handled. The option
109 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
110 options in pcretest are provided to set it. It can also be set by
111 (*NO_AUTO_POSSESS) at the start of a pattern.
112
113 18. The character VT has been added to the default ("C" locale) set of
114 characters that match \s and are generally treated as white space,
115 following this same change in Perl 5.18. There is now no difference between
116 "Perl space" and "POSIX space". Whether VT is treated as white space in
117 other locales depends on the locale.
118
119 19. The code for checking named groups as conditions, either for being set or
120 for being recursed, has been refactored (this is related to 14 and 15
121 above). Processing unduplicated named groups should now be as fast at
122 numerical groups, and processing duplicated groups should be faster than
123 before.
124
125 20. Two patches to the CMake build system, by Alexander Barkov:
126
127 (1) Replace the "source" command by "." in CMakeLists.txt because
128 "source" is a bash-ism.
129
130 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
131 without these the CMake build does not work on Solaris.
132
133 21. Perl has changed its handling of \8 and \9. If there is no previously
134 encountered capturing group of those numbers, they are treated as the
135 literal characters 8 and 9 instead of a binary zero followed by the
136 literals. PCRE now does the same.
137
138 22. Following Perl, added \o{} to specify codepoints in octal, making it
139 possible to specify values greater than 0777 and also making them
140 unambiguous.
141
142 23. Perl now gives an error for missing closing braces after \x{... instead of
143 treating the string as literal. PCRE now does the same.
144
145 24. RunTest used to grumble if an inappropriate test was selected explicitly,
146 but just skip it when running all tests. This make it awkward to run ranges
147 of tests when one of them was inappropriate. Now it just skips any
148 inappropriate tests, as it always did when running all tests.
149
150 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
151 character types such as \d or \w, too many callouts were inserted, and the
152 data that they returned was rubbish.
153
154 26. In UCP mode, \s was not matching two of the characters that Perl matches,
155 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
156 were matched by \h. The code has now been refactored so that the lists of
157 the horizontal and vertical whitespace characters used for \h and \v (which
158 are defined only in one place) are now also used for \s.
159
160 27. Add JIT support for the 64 bit TileGX architecture.
161 Patch by Jiong Wang (Tilera Corporation).
162
163 28. Possessive quantifiers for classes (both explicit and automatically
164 generated) now use special opcodes instead of wrapping in ONCE brackets.
165
166 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
167 (because it's meaningless), this was not happening when PCRE_CASELESS was
168 set. Not wrong, but inefficient.
169
170 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
171 properties for \w, \d, etc) is present in a test regex. Otherwise if the
172 test contains no characters greater than 255, Perl doesn't realise it
173 should be using Unicode semantics.
174
175 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
176 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
177 does in Unicode mode.
178
179 32. Added the "forbid" facility to pcretest so that putting tests into the
180 wrong test files can sometimes be quickly detected.
181
182 33. There is now a limit (default 250) on the depth of nesting of parentheses.
183 This limit is imposed to control the amount of system stack used at compile
184 time. It can be changed at build time by --with-parens-nest-limit=xxx or
185 the equivalent in CMake.
186
187 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
188 errors. Perl warns for these when in warning mode, but PCRE has no facility
189 for giving warnings.
190
191 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
192 However, this was not working for (?!) because it is optimized to (*FAIL),
193 for which PCRE does not allow quantifiers. The optimization is now disabled
194 when a quantifier follows (?!). I can't see any use for this, but it makes
195 things uniform.
196
197 36. Perl no longer allows group names to start with digits, so I have made this
198 change also in PCRE. It simplifies the code a bit.
199
200 37. In extended mode, Perl ignores spaces before a + that indicates a
201 possessive quantifier. PCRE allowed a space before the quantifier, but not
202 before the possessive +. It now does.
203
204 38. The use of \K (reset reported match start) within a repeated possessive
205 group such as (a\Kb)*+ was not working.
206
207 40. Document that the same character tables must be used at compile time and
208 run time, and that the facility to pass tables to pcre_exec() and
209 pcre_dfa_exec() is for use only with saved/restored patterns.
210
211 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
212 features for Builds with MSVC:
213
214 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
215 builds). The .rc files can be used to set FileDescription and many other
216 attributes.
217
218 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
219 This allows higher-level build scripts which want .pdb files to avoid
220 hard-coding the exact files needed."
221
222 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
223 mean "start of word" and "end of word", respectively, as a transition aid.
224
225 43. A minimizing repeat of a class containing codepoints greater than 255 in
226 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
227 compiled to use the heap for recursion.
228
229 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
230 is configured.
231
232
233 Version 8.33 28-May-2013
234 ------------------------
235
236 1. Added 'U' to some constants that are compared to unsigned integers, to
237 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
238 variables that are added to signed variables, to ensure the result is
239 signed and can be negated.
240
241 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
242 CMake config files.
243
244 3. Revise the creation of config.h.generic so that all boolean macros are
245 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
246 overriding via -D on the command line possible.
247
248 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
249 to unsigned int is reported to make a quite noticeable speed difference in
250 a specific Windows environment. Testing on Linux did also appear to show
251 some benefit (and it is clearly not harmful). Also fixed the definition of
252 Xop which should be unsigned.
253
254 5. Related to (4), changing the definition of the intermediate variable cc
255 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
256 improvements.
257
258 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
259 unnecessary spaces.
260
261 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
262 and later.
263
264 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
265
266 9. Optimizing fast_forward_start_bits in JIT.
267
268 10. Adding support for callouts in JIT, and fixing some issues revealed
269 during this work. Namely:
270
271 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
272
273 (b) Minimum length was not checked before the matching is started.
274
275 11. The value of capture_last that is passed to callouts was incorrect in some
276 cases when there was a capture on one path that was subsequently abandoned
277 after a backtrack. Also, the capture_last value is now reset after a
278 recursion, since all captures are also reset in this case.
279
280 12. The interpreter no longer returns the "too many substrings" error in the
281 case when an overflowing capture is in a branch that is subsequently
282 abandoned after a backtrack.
283
284 13. In the pathological case when an offset vector of size 2 is used, pcretest
285 now prints out the matched string after a yield of 0 or 1.
286
287 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
288 Only supported by the JIT compiler at the moment.
289
290 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
291
292 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
293 offset of the starting point of the matching process, provided the offsets
294 vector is large enough.
295
296 17. The \A escape now records a lookbehind value of 1, though its execution
297 does not actually inspect the previous character. This is to ensure that,
298 in partial multi-segment matching, at least one character from the old
299 segment is retained when a new segment is processed. Otherwise, if there
300 are no lookbehinds in the pattern, \A might match incorrectly at the start
301 of a new segment.
302
303 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
304
305 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
306 gives some modest performance improvement in 8-bit mode.
307
308 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
309 be expressed in certain programming languages using Universal Character
310 Names.
311
312 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
313 which points out that "non characters" are not "characters that may not
314 appear in Unicode strings" but rather "characters that are reserved for
315 internal use and have only local meaning".
316
317 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
318 there was a conditional group that depended on an assertion, if the
319 assertion was false, the callout that immediately followed the alternation
320 in the condition was skipped when pcre_exec() was used for matching.
321
322 23. Allow an explicit callout to be inserted before an assertion that is the
323 condition for a conditional group, for compatibility with automatic
324 callouts, which always insert a callout at this point.
325
326 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
327 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
328
329 25. (*PRUNE) is now supported by the JIT compiler.
330
331 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
332
333 27. Fix the case where there are two or more SKIPs with arguments that may be
334 ignored.
335
336 28. (*SKIP) is now supported by the JIT compiler.
337
338 29. (*THEN) is now supported by the JIT compiler.
339
340 30. Update RunTest with additional test selector options.
341
342 31. The way PCRE handles backtracking verbs has been changed in two ways.
343
344 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
345 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
346 backtracking. In some cases this makes it more Perl-compatible, but Perl's
347 rather obscure rules do not always do the same thing.
348
349 (2) Previously, backtracking verbs were confined within assertions. This is
350 no longer the case for positive assertions, except for (*ACCEPT). Again,
351 this sometimes improves Perl compatibility, and sometimes does not.
352
353 32. A number of tests that were in test 2 because Perl did things differently
354 have been moved to test 1, because either Perl or PCRE has changed, and
355 these tests are now compatible.
356
357 32. Backtracking control verbs are now handled in the same way in JIT and
358 interpreter.
359
360 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
361 contained a forward subroutine reference caused a compile error.
362
363 34. Auto-detect and optimize limited repetitions in JIT.
364
365 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
366 blocking (*UTF) etc.
367
368 36. In the interpreter, maximizing pattern repetitions for characters and
369 character types now use tail recursion, which reduces stack usage.
370
371 37. The value of the max lookbehind was not correctly preserved if a compiled
372 and saved regex was reloaded on a host of different endianness.
373
374 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
375 of the compiled pattern block, expand the flags field from 16 to 32 bits
376 because it was almost full.
377
378 39. Try madvise first before posix_madvise.
379
380 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
381 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
382 original change locked out all matches of empty strings. This has been
383 changed so that one match of an empty string per line is recognized.
384 Subsequent searches on the same line (for colouring or for --only-matching,
385 for example) do not recognize empty strings.
386
387 41. Applied a user patch to fix a number of spelling mistakes in comments.
388
389 42. Data lines longer than 65536 caused pcretest to crash.
390
391 43. Clarified the data type for length and startoffset arguments for pcre_exec
392 and pcre_dfa_exec in the function-specific man pages, where they were
393 explicitly stated to be in bytes, never having been updated. I also added
394 some clarification to the pcreapi man page.
395
396 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
397 a segmentation fault.
398
399
400 Version 8.32 30-November-2012
401 -----------------------------
402
403 1. Improved JIT compiler optimizations for first character search and single
404 character iterators.
405
406 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
407 Patch by Daniel Richard G.
408
409 3. Single character iterator optimizations in the JIT compiler.
410
411 4. Improved JIT compiler optimizations for character ranges.
412
413 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
414 Reported by Giuseppe D'Angelo.
415
416 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
417 of a line, was being set incorrectly in cases where .* appeared inside
418 atomic brackets at the start of a pattern, or where there was a subsequent
419 *PRUNE or *SKIP.
420
421 7. Improved instruction cache flush for POWER/PowerPC.
422 Patch by Daniel Richard G.
423
424 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
425 grep:
426
427 (a) There is now no limit to the number of patterns to be matched.
428
429 (b) An error is given if a pattern is too long.
430
431 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
432 are now supported.
433
434 (d) --exclude-from and --include-from (multiple use) have been added.
435
436 (e) Exclusions and inclusions now apply to all files and directories, not
437 just to those obtained from scanning a directory recursively.
438
439 (f) Multiple uses of -f and --file-list are now supported.
440
441 (g) In a Windows environment, the default for -d has been changed from
442 "read" (the GNU grep default) to "skip", because otherwise the presence
443 of a directory in the file list provokes an error.
444
445 (h) The documentation has been revised and clarified in places.
446
447 9. Improve the matching speed of capturing brackets.
448
449 10. Changed the meaning of \X so that it now matches a Unicode extended
450 grapheme cluster.
451
452 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
453 out POSIX threads when JIT support is configured.
454
455 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
456
457 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
458 structure could go wrong in environments where size_t is not the same size
459 as int.
460
461 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
462 set.
463
464 15. The EBCDIC support had decayed; later updates to the code had included
465 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
466 general tidy up of EBCDIC-related issues, and the documentation was also
467 not quite right. There is now a test that can be run on ASCII systems to
468 check some of the EBCDIC-related things (but is it not a full test).
469
470 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
471 in a small tidy to the code.
472
473 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
474
475 18. If the --only-matching (-o) option in pcregrep is specified multiple
476 times, each one causes appropriate output. For example, -o1 -o2 outputs the
477 substrings matched by the 1st and 2nd capturing parentheses. A separating
478 string can be specified by --om-separator (default empty).
479
480 19. Improving the first n character searches.
481
482 20. Turn case lists for horizontal and vertical white space into macros so that
483 they are defined only once.
484
485 21. This set of changes together give more compatible Unicode case-folding
486 behaviour for characters that have more than one other case when UCP
487 support is available.
488
489 (a) The Unicode property table now has offsets into a new table of sets of
490 three or more characters that are case-equivalent. The MultiStage2.py
491 script that generates these tables (the pcre_ucd.c file) now scans
492 CaseFolding.txt instead of UnicodeData.txt for character case
493 information.
494
495 (b) The code for adding characters or ranges of characters to a character
496 class has been abstracted into a generalized function that also handles
497 case-independence. In UTF-mode with UCP support, this uses the new data
498 to handle characters with more than one other case.
499
500 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
501 whose other case is greater than 256 are now correctly matched
502 caselessly. Previously, the high codepoint matched the low one, but not
503 vice versa.
504
505 (d) The processing of \h, \H, \v, and \ in character classes now makes use
506 of the new class addition function, using character lists defined as
507 macros alongside the case definitions of 20 above.
508
509 (e) Caseless back references now work with characters that have more than
510 one other case.
511
512 (f) General caseless matching of characters with more than one other case
513 is supported.
514
515 22. Unicode character properties were updated from Unicode 6.2.0
516
517 23. Improved CMake support under Windows. Patch by Daniel Richard G.
518
519 24. Add support for 32-bit character strings, and UTF-32
520
521 25. Major JIT compiler update (code refactoring and bugfixing).
522 Experimental Sparc 32 support is added.
523
524 26. Applied a modified version of Daniel Richard G's patch to create
525 pcre.h.generic and config.h.generic by "make" instead of in the
526 PrepareRelease script.
527
528 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
529 pcre_compile.c when checking for a zero character.
530
531 28. Introducing a native interface for JIT. Through this interface, the compiled
532 machine code can be directly executed. The purpose of this interface is to
533 provide fast pattern matching, so several sanity checks are not performed.
534 However, feature tests are still performed. The new interface provides
535 1.4x speedup compared to the old one.
536
537 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
538 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
539 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
540
541 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
542 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
543 the "old" RFC 2279). Instead, it ended up passing a negative length to
544 pcre_exec().
545
546 31. Add support for GCC's visibility feature to hide internal functions.
547
548 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
549 "unknown -C option" after outputting 0 or 1.
550
551 33. There is now support for generating a code coverage report for the test
552 suite in environments where gcc is the compiler and lcov is installed. This
553 is mainly for the benefit of the developers.
554
555 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
556 unaddressable using valgrind annotations, allowing valgrind to detect
557 invalid memory accesses. This is mainly for the benefit of the developers.
558
559 25. (*UTF) can now be used to start a pattern in any of the three libraries.
560
561 26. Give configure error if --enable-cpp but no C++ compiler found.
562
563
564 Version 8.31 06-July-2012
565 -------------------------
566
567 1. Fixing a wrong JIT test case and some compiler warnings.
568
569 2. Removed a bashism from the RunTest script.
570
571 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
572 to unsigned type, result still unsigned" that was given by an MS compiler
573 on encountering the code "-sizeof(xxx)".
574
575 4. Partial matching support is added to the JIT compiler.
576
577 5. Fixed several bugs concerned with partial matching of items that consist
578 of more than one character:
579
580 (a) /^(..)\1/ did not partially match "aba" because checking references was
581 done on an "all or nothing" basis. This also applied to repeated
582 references.
583
584 (b) \R did not give a hard partial match if \r was found at the end of the
585 subject.
586
587 (c) \X did not give a hard partial match after matching one or more
588 characters at the end of the subject.
589
590 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
591 a partial match for the string "\r".
592
593 (e) When newline was set to CRLF, the metacharacter "." did not recognize
594 a partial match for a CR character at the end of the subject string.
595
596 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
597 running pcretest, the text "(JIT)" added to the output whenever JIT is
598 actually used to run the match.
599
600 7. Individual JIT compile options can be set in pcretest by following -s+[+]
601 or /S+[+] with a digit between 1 and 7.
602
603 8. OP_NOT now supports any UTF character not just single-byte ones.
604
605 9. (*MARK) control verb is now supported by the JIT compiler.
606
607 10. The command "./RunTest list" lists the available tests without actually
608 running any of them. (Because I keep forgetting what they all are.)
609
610 11. Add PCRE_INFO_MAXLOOKBEHIND.
611
612 12. Applied a (slightly modified) user-supplied patch that improves performance
613 when the heap is used for recursion (compiled with --disable-stack-for-
614 recursion). Instead of malloc and free for each heap frame each time a
615 logical recursion happens, frames are retained on a chain and re-used where
616 possible. This sometimes gives as much as 30% improvement.
617
618 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
619 call.
620
621 14. As documented, (*COMMIT) is now confined to within a positive assertion.
622
623 15. It is now possible to link pcretest with libedit as an alternative to
624 libreadline.
625
626 16. (*COMMIT) control verb is now supported by the JIT compiler.
627
628 17. The Unicode data tables have been updated to Unicode 6.1.0.
629
630 18. Added --file-list option to pcregrep.
631
632 19. Added binary file support to pcregrep, including the -a, --binary-files,
633 -I, and --text options.
634
635 20. The madvise function is renamed for posix_madvise for QNX compatibility
636 reasons. Fixed by Giuseppe D'Angelo.
637
638 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
639
640 22. Changed the option for creating symbolic links for 16-bit man pages from
641 -s to -sf so that re-installing does not cause issues.
642
643 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
644
645 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
646 matches in certain environments (the workspace was not being correctly
647 retained). Also added to pcre_dfa_exec() a simple plausibility check on
648 some of the workspace data at the beginning of a restart.
649
650 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
651 was not doing so when it should - probably a typo introduced by SVN 528
652 (change 8.10/14).
653
654 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
655 \w+ when the character tables indicated that \x{c4} was a word character.
656 There were several related cases, all because the tests for doing a table
657 lookup were testing for characters less than 127 instead of 255.
658
659 27. If a pattern contains capturing parentheses that are not used in a match,
660 their slots in the ovector are set to -1. For those that are higher than
661 any matched groups, this happens at the end of processing. In the case when
662 there were back references that the ovector was too small to contain
663 (causing temporary malloc'd memory to be used during matching), and the
664 highest capturing number was not used, memory off the end of the ovector
665 was incorrectly being set to -1. (It was using the size of the temporary
666 memory instead of the true size.)
667
668 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
669 ovector size, it uses memory at the end of the block that it has got.
670
671 29. Check for an overlong MARK name and give an error at compile time. The
672 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
673
674 30. JIT compiler update.
675
676 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
677 Rill for the patch.
678
679 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
680
681 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
682
683 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
684 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
685 was enabled.
686
687 35. Improve JIT code generation for greedy plus quantifier.
688
689 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
690 1 to "aa" instead of to an empty string. The bug affected repeated groups
691 that could potentially match an empty string.
692
693 37. Optimizing single character iterators in JIT.
694
695 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
696 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
697 codepoints that are too big for the mode are faulted, and in a UTF mode,
698 disallowed codepoints are also faulted.
699
700 39. If PCRE was compiled with UTF support, in three places in the DFA
701 matcher there was code that should only have been obeyed in UTF mode, but
702 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
703 processing when bytes with values greater than 127 were present. In 16-bit
704 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
705 both cases the values are those that cannot be the first data item in a UTF
706 character. The three items that might have provoked this were recursions,
707 possessively repeated groups, and atomic groups.
708
709 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
710 and pcregrep, because some OS require shared objects to be explicitly
711 passed to ld, causing the link step to fail if they are not.
712
713 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
714 mode, patterns that started with \h* or \R* might be incorrectly matched.
715
716
717 Version 8.30 04-February-2012
718 -----------------------------
719
720 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
721 name is defined in ctype.h.
722
723 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
724 only in quite long subpatterns.
725
726 3. Removed the function pcre_info(), which has been obsolete and deprecated
727 since it was replaced by pcre_fullinfo() in February 2000.
728
729 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
730 match a (*MARK), and the match failed at the start of the subject, a
731 reference to memory before the start of the subject could occur. This bug
732 was introduced by fix 17 of release 8.21.
733
734 5. A reference to an unset group with zero minimum repetition was giving
735 totally wrong answers (in non-JavaScript-compatibility mode). For example,
736 /(another)?(\1?)test/ matched against "hello world test". This bug was
737 introduced in release 8.13.
738
739 6. Add support for 16-bit character strings (a large amount of work involving
740 many changes and refactorings).
741
742 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
743 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
744 from a file.
745
746 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
747 rounding is not applied in this particular case).
748
749 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
750 if they appear, or are escaped, in patterns.
751
752 10. Get rid of a number of -Wunused-but-set-variable warnings.
753
754 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
755 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
756 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
757 also returns the mark "x". This bug applied to capturing parentheses,
758 non-capturing parentheses, and atomic parentheses. It also applied to some
759 assertions.
760
761 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
762 information out of configure.ac instead of relying on pcre.h.generic, which
763 is not stored in the repository.
764
765 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
766 -lreadline.
767
768 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
769
770 15. Applied Graycode's patch to put the top-level frame on the stack rather
771 than the heap when not using the stack for recursion. This gives a
772 performance improvement in many cases when recursion is not deep.
773
774 16. Experimental code added to "pcretest -C" to output the stack frame size.
775
776
777 Version 8.21 12-Dec-2011
778 ------------------------
779
780 1. Updating the JIT compiler.
781
782 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
783 are added as well.
784
785 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
786 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
787 calling _pcre_jit_exec. Some extra comments are added.
788
789 4. (*MARK) settings inside atomic groups that do not contain any capturing
790 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
791 was introduced by change 18 for 8.20.
792
793 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
794 ECMA-262 standard.
795
796 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
797 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
798 This bug was probably introduced by change 9 of 8.13.
799
800 7. While fixing 6 above, I noticed that a number of other items were being
801 incorrectly rejected as "not fixed length". This arose partly because newer
802 opcodes had not been added to the fixed-length checking code. I have (a)
803 corrected the bug and added tests for these items, and (b) arranged for an
804 error to occur if an unknown opcode is encountered while checking for fixed
805 length instead of just assuming "not fixed length". The items that were
806 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
807 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
808 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
809
810 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
811 being incorrectly compiled and would have given unpredicatble results.
812
813 9. A possessively repeated subpattern with minimum repeat count greater than
814 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
815 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
816 the first (A) could occur when it should not.
817
818 10. Add a cast and remove a redundant test from the code.
819
820 11. JIT should use pcre_malloc/pcre_free for allocation.
821
822 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
823 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
824 is anything other than /usr, -L is still shown).
825
826 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
827
828 14. Perl does not support \N without a following name in a [] class; PCRE now
829 also gives an error.
830
831 15. If a forward reference was repeated with an upper limit of around 2000,
832 it caused the error "internal error: overran compiling workspace". The
833 maximum number of forward references (including repeats) was limited by the
834 internal workspace, and dependent on the LINK_SIZE. The code has been
835 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
836 the default depends on LINK_SIZE. There is a new upper limit (for safety)
837 of around 200,000 forward references. While doing this, I also speeded up
838 the filling in of repeated forward references.
839
840 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
841 incorrectly expecting the subject to contain another "a" after the start.
842
843 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
844 in the match, the SKIP should be ignored. This was not happening; instead
845 the SKIP was being treated as NOMATCH. For patterns such as
846 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
847 tested.
848
849 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
850 now much more compatible with Perl, in particular in cases where the result
851 is a non-match for a non-anchored pattern. For example, if
852 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
853 "m", where previously it did not return a name. A side effect of this
854 change is that for partial matches, the last encountered mark name is
855 returned, as for non matches. A number of tests that were previously not
856 Perl-compatible have been moved into the Perl-compatible test files. The
857 refactoring has had the pleasing side effect of removing one argument from
858 the match() function, thus reducing its stack requirements.
859
860 19. If the /S+ option was used in pcretest to study a pattern using JIT,
861 subsequent uses of /S (without +) incorrectly behaved like /S+.
862
863 21. Retrieve executable code size support for the JIT compiler and fixing
864 some warnings.
865
866 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
867 not work when the shorter character appeared right at the end of the
868 subject string.
869
870 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
871 systems.
872
873 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
874 output it when the /M option is used in pcretest.
875
876 25. The CheckMan script was not being included in the distribution. Also, added
877 an explicit "perl" to run Perl scripts from the PrepareRelease script
878 because this is reportedly needed in Windows.
879
880 26. If study data was being save in a file and studying had not found a set of
881 "starts with" bytes for the pattern, the data written to the file (though
882 never used) was taken from uninitialized memory and so caused valgrind to
883 complain.
884
885 27. Updated RunTest.bat as provided by Sheri Pierce.
886
887 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
888
889 29. Computation of memory usage for the table of capturing group names was
890 giving an unnecessarily large value.
891
892
893 Version 8.20 21-Oct-2011
894 ------------------------
895
896 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
897 a POSIX class. After further experiments with Perl, which convinced me that
898 Perl has bugs and confusions, a closing square bracket is no longer allowed
899 in a POSIX name. This bug also affected patterns with classes that started
900 with full stops.
901
902 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
903 captured substring, but while checking the failing first alternative,
904 substring 1 is temporarily captured. If the output vector supplied to
905 pcre_exec() was not big enough for this capture, the yield of the function
906 was still zero ("insufficient space for captured substrings"). This cannot
907 be totally fixed without adding another stack variable, which seems a lot
908 of expense for a edge case. However, I have improved the situation in cases
909 such as /(a)(b)x|abc/ matched against "abc", where the return code
910 indicates that fewer than the maximum number of slots in the ovector have
911 been set.
912
913 3. Related to (2) above: when there are more back references in a pattern than
914 slots in the output vector, pcre_exec() uses temporary memory during
915 matching, and copies in the captures as far as possible afterwards. It was
916 using the entire output vector, but this conflicts with the specification
917 that only 2/3 is used for passing back captured substrings. Now it uses
918 only the first 2/3, for compatibility. This is, of course, another edge
919 case.
920
921 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
922 main code base, and can be used by building with --enable-jit. When this is
923 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
924 runtime --no-jit option is given.
925
926 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
927 ovector, the return from the function was zero, implying that there were
928 other matches that did not fit. The correct "exactly full" value is now
929 returned.
930
931 6. If a subpattern that was called recursively or as a subroutine contained
932 (*PRUNE) or any other control that caused it to give a non-standard return,
933 invalid errors such as "Error -26 (nested recursion at the same subject
934 position)" or even infinite loops could occur.
935
936 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
937 computing the minimum length on reaching *ACCEPT, and so ended up with the
938 wrong value of 1 rather than 0. Further investigation indicates that
939 computing a minimum subject length in the presence of *ACCEPT is difficult
940 (think back references, subroutine calls), and so I have changed the code
941 so that no minimum is registered for a pattern that contains *ACCEPT.
942
943 8. If (*THEN) was present in the first (true) branch of a conditional group,
944 it was not handled as intended. [But see 16 below.]
945
946 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
947 Sheri Pierce.
948
949 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
950 the first byte in a match must be "a".
951
952 11. Change 17 for 8.13 increased the recursion depth for patterns like
953 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
954 pattern contains any instances of (*THEN). If it does not, the old
955 optimizations are restored. It would be nice to do this on a per-group
956 basis, but at the moment that is not feasible.
957
958 12. In some environments, the output of pcretest -C is CRLF terminated. This
959 broke RunTest's code that checks for the link size. A single white space
960 character after the value is now allowed for.
961
962 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
963 For "fr", it uses the Windows-specific input and output files.
964
965 14. If (*THEN) appeared in a group that was called recursively or as a
966 subroutine, it did not work as intended. [But see next item.]
967
968 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
969 pattern fragments (but not containing any | characters). If A and B are
970 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
971 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
972 D. In other words, Perl considers parentheses that do not contain any |
973 characters to be part of a surrounding alternative, whereas PCRE was
974 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
975 differently. PCRE now behaves in the same way as Perl, except in the case
976 of subroutine/recursion calls such as (?1) which have in any case always
977 been different (but PCRE had them first :-).
978
979 16. Related to 15 above: Perl does not treat the | in a conditional group as
980 creating alternatives. Such a group is treated in the same way as an
981 ordinary group without any | characters when processing (*THEN). PCRE has
982 been changed to match Perl's behaviour.
983
984 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
985 RunGrepTest script failed.
986
987 18. Change 22 for version 13 caused atomic groups to use more stack. This is
988 inevitable for groups that contain captures, but it can lead to a lot of
989 stack use in large patterns. The old behaviour has been restored for atomic
990 groups that do not contain any capturing parentheses.
991
992 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
993 suppress the check for a minimum subject length at run time. (If it was
994 given to pcre_exec() or pcre_dfa_exec() it did work.)
995
996 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
997 fail to work when decoding hex characters in data strings in EBCDIC
998 environments.
999
1000 21. It appears that in at least one Mac OS environment, the isxdigit() function
1001 is implemented as a macro that evaluates to its argument more than once,
1002 contravening the C 90 Standard (I haven't checked a later standard). There
1003 was an instance in pcretest which caused it to go wrong when processing
1004 \x{...} escapes in subject strings. The has been rewritten to avoid using
1005 things like p++ in the argument of isxdigit().
1006
1007
1008 Version 8.13 16-Aug-2011
1009 ------------------------
1010
1011 1. The Unicode data tables have been updated to Unicode 6.0.0.
1012
1013 2. Two minor typos in pcre_internal.h have been fixed.
1014
1015 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1016 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1017 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1018
1019 4. There were a number of related bugs in the code for matching backrefences
1020 caselessly in UTF-8 mode when codes for the characters concerned were
1021 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1022 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1023 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1024 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1025 2-byte code at the end of the subject (it thought there wasn't enough data
1026 left).
1027
1028 5. Comprehensive information about what went wrong is now returned by
1029 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1030 as the output vector has at least 2 elements. The offset of the start of
1031 the failing character and a reason code are placed in the vector.
1032
1033 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1034 now returned is for the first byte of the failing character, instead of the
1035 last byte inspected. This is an incompatible change, but I hope it is small
1036 enough not to be a problem. It makes the returned offset consistent with
1037 pcre_exec() and pcre_dfa_exec().
1038
1039 7. pcretest now gives a text phrase as well as the error number when
1040 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1041 failure, the offset and reason code are output.
1042
1043 8. When \R was used with a maximizing quantifier it failed to skip backwards
1044 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1045 back over a single character (\n). This seems wrong (because it treated the
1046 two characters as a single entity when going forwards), conflicts with the
1047 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1048 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1049 has been changed.
1050
1051 9. Some internal refactoring has changed the processing so that the handling
1052 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1053 time (the PCRE_DOTALL option was changed this way some time ago: version
1054 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1055 which was always a bit of a fudge. It also means that there is one less
1056 argument for the match() function, which reduces its stack requirements
1057 slightly. This change also fixes an incompatibility with Perl: the pattern
1058 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1059
1060 10. More internal refactoring has drastically reduced the number of recursive
1061 calls to match() for possessively repeated groups such as (abc)++ when
1062 using pcre_exec().
1063
1064 11. While implementing 10, a number of bugs in the handling of groups were
1065 discovered and fixed:
1066
1067 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1068 (a|)*(?1) gave a compile-time internal error.
1069 ((a|)+)+ did not notice that the outer group could match an empty string.
1070 (^a|^)+ was not marked as anchored.
1071 (.*a|.*)+ was not marked as matching at start or after a newline.
1072
1073 12. Yet more internal refactoring has removed another argument from the match()
1074 function. Special calls to this function are now indicated by setting a
1075 value in a variable in the "match data" data block.
1076
1077 13. Be more explicit in pcre_study() instead of relying on "default" for
1078 opcodes that mean there is no starting character; this means that when new
1079 ones are added and accidentally left out of pcre_study(), testing should
1080 pick them up.
1081
1082 14. The -s option of pcretest has been documented for ages as being an old
1083 synonym of -m (show memory usage). I have changed it to mean "force study
1084 for every regex", that is, assume /S for every regex. This is similar to -i
1085 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1086 using it. It makes it easier to run collections of tests with and without
1087 study enabled, and thereby test pcre_study() more easily. All the standard
1088 tests are now run with and without -s (but some patterns can be marked as
1089 "never study" - see 20 below).
1090
1091 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1092 restoration of the capturing data to the outer values was not happening
1093 correctly.
1094
1095 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1096 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1097 pattern had matched an empty string, and so incorrectly returned a no
1098 match.
1099
1100 17. There was optimizing code for the last branch of non-capturing parentheses,
1101 and also for the obeyed branch of a conditional subexpression, which used
1102 tail recursion to cut down on stack usage. Unfortunately, now that there is
1103 the possibility of (*THEN) occurring in these branches, tail recursion is
1104 no longer possible because the return has to be checked for (*THEN). These
1105 two optimizations have therefore been removed. [But see 8.20/11 above.]
1106
1107 18. If a pattern containing \R was studied, it was assumed that \R always
1108 matched two bytes, thus causing the minimum subject length to be
1109 incorrectly computed because \R can also match just one byte.
1110
1111 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1112 was incorrectly computed.
1113
1114 20. If /S is present twice on a test pattern in pcretest input, it now
1115 *disables* studying, thereby overriding the use of -s on the command line
1116 (see 14 above). This is necessary for one or two tests to keep the output
1117 identical in both cases.
1118
1119 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1120 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1121
1122 22. When an atomic group that contained a capturing parenthesis was
1123 successfully matched, but the branch in which it appeared failed, the
1124 capturing was not being forgotten if a higher numbered group was later
1125 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1126 group 1 to "a", when in fact it should be unset. This applied to multi-
1127 branched capturing and non-capturing groups, repeated or not, and also to
1128 positive assertions (capturing in negative assertions does not happen
1129 in PCRE) and also to nested atomic groups.
1130
1131 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1132 subject after a captured substring, to make it easier to tell which of a
1133 number of identical substrings has been captured.
1134
1135 24. The way atomic groups are processed by pcre_exec() has been changed so that
1136 if they are repeated, backtracking one repetition now resets captured
1137 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1138 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1139 "aaa". Previously, it would have been "a". As part of this code
1140 refactoring, the way recursive calls are handled has also been changed.
1141
1142 25. If an assertion condition captured any substrings, they were not passed
1143 back unless some other capturing happened later. For example, if
1144 (?(?=(a))a) was matched against "a", no capturing was returned.
1145
1146 26. When studying a pattern that contained subroutine calls or assertions,
1147 the code for finding the minimum length of a possible match was handling
1148 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1149 group 1 called group 2 while simultaneously a separate group 2 called group
1150 1). A stack overflow occurred in this case. I have fixed this by limiting
1151 the recursion depth to 10.
1152
1153 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1154 Fortmann. This supports explicit test numbers on the command line, and has
1155 argument validation and error reporting.
1156
1157 28. An instance of \X with an unlimited repeat could fail if at any point the
1158 first character it looked at was a mark character.
1159
1160 29. Some minor code refactoring concerning Unicode properties and scripts
1161 should reduce the stack requirement of match() slightly.
1162
1163 30. Added the '=' option to pcretest to check the setting of unused capturing
1164 slots at the end of the pattern, which are documented as being -1, but are
1165 not included in the return count.
1166
1167 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1168 compiled something random. Now it gives a compile-time error (as does
1169 Perl).
1170
1171 32. A *MARK encountered during the processing of a positive assertion is now
1172 recorded and passed back (compatible with Perl).
1173
1174 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1175 had alternative anchored branches, the search for a second match in a line
1176 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1177 matched the line "0102" twice. The same bug affected patterns that started
1178 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1179 twice.
1180
1181 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1182 does, and because of capturing effects, quantifying parenthesized
1183 assertions may at times be useful. Quantifiers are now allowed for
1184 parenthesized assertions.
1185
1186 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1187
1188 36. \g was being checked for fancy things in a character class, when it should
1189 just be a literal "g".
1190
1191 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1192 appearance of a nested POSIX class supersedes an apparent external class.
1193 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1194 unescaped square brackets may also appear as part of class names. For
1195 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1196 more like Perl. (But see 8.20/1 above.)
1197
1198 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1199 was because it thought it was \N{name}, which is not supported).
1200
1201 39. Add minix to OS list not supporting the -S option in pcretest.
1202
1203 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1204 cannot analyze patterns in sufficient detail to catch mutual recursions
1205 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1206 subgroup is called recursively as a subpattern for a second time at the
1207 same position in the subject string. In previous releases this might have
1208 been caught by the recursion limit, or it might have run out of stack.
1209
1210 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1211 happen only once. PCRE was, however incorrectly giving a compile time error
1212 "recursive call could loop indefinitely" because it cannot analyze the
1213 pattern in sufficient detail. The compile time test no longer happens when
1214 PCRE is compiling a conditional subpattern, but actual runaway loops are
1215 now caught at runtime (see 40 above).
1216
1217 42. It seems that Perl allows any characters other than a closing parenthesis
1218 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1219 has been changed to be the same.
1220
1221 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1222 as not to get warnings when autogen.sh is called. Also changed
1223 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1224
1225 44. To help people who use pcregrep to scan files containing exceedingly long
1226 lines, the following changes have been made:
1227
1228 (a) The default value of the buffer size parameter has been increased from
1229 8K to 20K. (The actual buffer used is three times this size.)
1230
1231 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1232 PCRE is built.
1233
1234 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1235 to be set at run time.
1236
1237 (d) Numerical values in pcregrep options can be followed by K or M, for
1238 example --buffer-size=50K.
1239
1240 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1241 given and the return code is set to 2.
1242
1243 45. Add a pointer to the latest mark to the callout data block.
1244
1245 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1246 partial match of an empty string instead of no match. This was specific to
1247 the use of ".".
1248
1249 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1250 complete match instead of a partial match. This bug was dependent on both
1251 the PCRE_UTF8 and PCRE_DOTALL options being set.
1252
1253 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1254 starting byte set, because \b was not being ignored.
1255
1256
1257 Version 8.12 15-Jan-2011
1258 ------------------------
1259
1260 1. Fixed some typos in the markup of the man pages, and wrote a script that
1261 checks for such things as part of the documentation building process.
1262
1263 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1264 --match-limit and --recursion-limit options (added for 8.11). In
1265 particular, this made one of the standard tests fail. (The integer value
1266 went into the wrong half of a long int.)
1267
1268 3. If the --colour option was given to pcregrep with -v (invert match), it
1269 did strange things, either producing crazy output, or crashing. It should,
1270 of course, ignore a request for colour when reporting lines that do not
1271 match.
1272
1273 4. Another pcregrep bug caused similar problems if --colour was specified with
1274 -M (multiline) and the pattern match finished with a line ending.
1275
1276 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1277 matched in multiline mode, the following line was shown as part of the
1278 match. This seems wrong, so I have changed it.
1279
1280 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1281 the check for further matches in the same line (so they could be coloured)
1282 to overrun the end of the current line. If another match was found, it was
1283 incorrectly shown (and then shown again when found in the next line).
1284
1285 7. If pcregrep was compiled under Windows, there was a reference to the
1286 function pcregrep_exit() before it was defined. I am assuming this was
1287 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1288 reported by a user. I've moved the definition above the reference.
1289
1290
1291 Version 8.11 10-Dec-2010
1292 ------------------------
1293
1294 1. (*THEN) was not working properly if there were untried alternatives prior
1295 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1296 backtracked to try for "b" instead of moving to the next alternative branch
1297 at the same level (in this case, to look for "c"). The Perl documentation
1298 is clear that when (*THEN) is backtracked onto, it goes to the "next
1299 alternative in the innermost enclosing group".
1300
1301 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1302 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1303 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1304 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1305 (*THEN).
1306
1307 3. If \s appeared in a character class, it removed the VT character from
1308 the class, even if it had been included by some previous item, for example
1309 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1310 of \s, but is part of the POSIX "space" class.)
1311
1312 4. A partial match never returns an empty string (because you can always
1313 match an empty string at the end of the subject); however the checking for
1314 an empty string was starting at the "start of match" point. This has been
1315 changed to the "earliest inspected character" point, because the returned
1316 data for a partial match starts at this character. This means that, for
1317 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1318 (previously it gave "no match").
1319
1320 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1321 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1322 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1323 has an implication that the given string is incomplete (because a partial
1324 match is preferred over a full match). For this reason, these items now
1325 give a partial match in this situation. [Aside: previously, the one case
1326 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1327 match rather than a full match, which was wrong by the old rules, but is
1328 now correct.]
1329
1330 6. There was a bug in the handling of #-introduced comments, recognized when
1331 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1332 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1333 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1334 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1335 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1336 places in pcre_compile().
1337
1338 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1339 comments when looking ahead for named forward references to subpatterns,
1340 the only newline sequence it recognized was NL. It now handles newlines
1341 according to the set newline convention.
1342
1343 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1344 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1345 cater for a lack of strerror(). These oversights have been fixed.
1346
1347 9. Added --match-limit and --recursion-limit to pcregrep.
1348
1349 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1350
1351 11. When the -o option was used, pcregrep was setting a return code of 1, even
1352 when matches were found, and --line-buffered was not being honoured.
1353
1354 12. Added an optional parentheses number to the -o and --only-matching options
1355 of pcregrep.
1356
1357 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1358 can match an empty string. The code to do it in pcretest and pcredemo
1359 needed fixing:
1360
1361 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1362 only one byte after an empty string match just before CRLF (this case
1363 just got forgotten; "any" and "anycrlf" were OK).
1364
1365 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1366 mode when an empty string match preceded an ASCII character followed by
1367 a non-ASCII character. (The code for advancing by one character rather
1368 than one byte was nonsense.)
1369
1370 (c) The pcredemo.c sample program did not have any code at all to handle
1371 the cases when CRLF is a valid newline sequence.
1372
1373 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1374 as a starting offset was within the subject string. There is now a new
1375 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1376 negative or greater than the length of the string. In order to test this,
1377 pcretest is extended to allow the setting of negative starting offsets.
1378
1379 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1380 starting offset points to the beginning of a UTF-8 character was
1381 unnecessarily clumsy. I tidied it up.
1382
1383 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1384 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1385
1386 17. Nobody had reported that the --include_dir option, which was added in
1387 release 7.7 should have been called --include-dir (hyphen, not underscore)
1388 for compatibility with GNU grep. I have changed it to --include-dir, but
1389 left --include_dir as an undocumented synonym, and the same for
1390 --exclude-dir, though that is not available in GNU grep, at least as of
1391 release 2.5.4.
1392
1393 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1394 characters from a string of bytes) have been redefined so as not to use
1395 loops, in order to improve performance in some environments. At the same
1396 time, I abstracted some of the common code into auxiliary macros to save
1397 repetition (this should not affect the compiled code).
1398
1399 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1400 compile-time error is now given if \c is not followed by an ASCII
1401 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1402 different, and any byte value is allowed.)
1403
1404 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1405 START_OPTIMIZE option, which is now allowed at compile time - but just
1406 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1407 to pcregrep and other applications that have no direct access to PCRE
1408 options. The new /Y option in pcretest sets this option when calling
1409 pcre_compile().
1410
1411 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1412 back references. Groups containing recursive back references were forced to
1413 be atomic by that change, but in the case of named groups, the amount of
1414 memory required was incorrectly computed, leading to "Failed: internal
1415 error: code overflow". This has been fixed.
1416
1417 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1418 pcretest.c, to avoid build problems in some Borland environments.
1419
1420
1421 Version 8.10 25-Jun-2010
1422 ------------------------
1423
1424 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1425 THEN.
1426
1427 2. (*ACCEPT) was not working when inside an atomic group.
1428
1429 3. Inside a character class, \B is treated as a literal by default, but
1430 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1431 causes the error). The code is unchanged, but I tidied the documentation.
1432
1433 4. Inside a character class, PCRE always treated \R and \X as literals,
1434 whereas Perl faults them if its -w option is set. I have changed PCRE so
1435 that it faults them when PCRE_EXTRA is set.
1436
1437 5. Added support for \N, which always matches any character other than
1438 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1439
1440 6. When compiling pcregrep with newer versions of gcc which may have
1441 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1442 declared with attribute warn_unused_result" were given. Just casting the
1443 result to (void) does not stop the warnings; a more elaborate fudge is
1444 needed. I've used a macro to implement this.
1445
1446 7. Minor change to pcretest.c to avoid a compiler warning.
1447
1448 8. Added four artifical Unicode properties to help with an option to make
1449 \s etc use properties (see next item). The new properties are: Xan
1450 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1451
1452 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1453 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1454 this option. Modified pcretest to add /W to test this facility. Added
1455 REG_UCP to make it available via the POSIX interface.
1456
1457 10. Added --line-buffered to pcregrep.
1458
1459 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1460 studied, and the match started with a letter with a code point greater than
1461 127 whose first byte was different to the first byte of the other case of
1462 the letter, the other case of this starting letter was not recognized
1463 (#976).
1464
1465 12. If a pattern that was studied started with a repeated Unicode property
1466 test, for example, \p{Nd}+, there was the theoretical possibility of
1467 setting up an incorrect bitmap of starting bytes, but fortunately it could
1468 not have actually happened in practice until change 8 above was made (it
1469 added property types that matched character-matching opcodes).
1470
1471 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1472 possible starting bytes for non-anchored patterns.
1473
1474 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1475 \R, and also a number of cases that involve Unicode properties, both
1476 explicit and implicit when PCRE_UCP is set.
1477
1478 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1479 input, it could crash or give wrong results if characters with values
1480 greater than 0xc0 were present in the subject string. (Detail: it assumed
1481 UTF-8 input when processing these items.)
1482
1483 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1484 size_t is 64-bit (#991).
1485
1486 17. Added a check for running out of memory when PCRE is compiled with
1487 --disable-stack-for-recursion (#990).
1488
1489 18. If the last data line in a file for pcretest does not have a newline on
1490 the end, a newline was missing in the output.
1491
1492 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1493 less than 128) in its various bitmaps. However, there is a facility for
1494 generating tables according to the current locale when PCRE is compiled. It
1495 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1496 characters, are recognized by isspace() and therefore were getting set in
1497 these tables, and indeed these tables seem to approximate to ISO 8859. This
1498 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1499 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1500 which of course cannot start UTF-8 characters. I have changed the code so
1501 that only real ASCII characters (less than 128) and the correct starting
1502 bytes for UTF-8 encodings are set for characters greater than 127 when in
1503 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1504 altogether.)
1505
1506 20. Added the /T option to pcretest so as to be able to run tests with non-
1507 standard character tables, thus making it possible to include the tests
1508 used for 19 above in the standard set of tests.
1509
1510 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1511 reference to a subpattern the other side of a comment that contains an
1512 opening parenthesis caused either an internal compiling error, or a
1513 reference to the wrong subpattern.
1514
1515
1516 Version 8.02 19-Mar-2010
1517 ------------------------
1518
1519 1. The Unicode data tables have been updated to Unicode 5.2.0.
1520
1521 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1522 configured.
1523
1524 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1525 original author of that file, following a query about its status.
1526
1527 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1528 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1529
1530 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1531 quantifier applied to a forward-referencing subroutine call, could compile
1532 incorrect code or give the error "internal error: previously-checked
1533 referenced subpattern not found".
1534
1535 6. Both MS Visual Studio and Symbian OS have problems with initializing
1536 variables to point to external functions. For these systems, therefore,
1537 pcre_malloc etc. are now initialized to local functions that call the
1538 relevant global functions.
1539
1540 7. There were two entries missing in the vectors called coptable and poptable
1541 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1542 I've fixed the data, and added a kludgy way of testing at compile time that
1543 the lengths are correct (equal to the number of opcodes).
1544
1545 8. Following on from 7, I added a similar kludge to check the length of the
1546 eint vector in pcreposix.c.
1547
1548 9. Error texts for pcre_compile() are held as one long string to avoid too
1549 much relocation at load time. To find a text, the string is searched,
1550 counting zeros. There was no check for running off the end of the string,
1551 which could happen if a new error number was added without updating the
1552 string.
1553
1554 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1555
1556 11. \K was not working if it appeared in an atomic group or in a group that
1557 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1558 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1559 the assertion is positive, but not if it is negative.
1560
1561 12. Change 11 fortuitously reduced the size of the stack frame used in the
1562 "match()" function of pcre_exec.c by one pointer. Forthcoming
1563 implementation of support for (*MARK) will need an extra pointer on the
1564 stack; I have reserved it now, so that the stack frame size does not
1565 decrease.
1566
1567 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1568 item in branch that calls a recursion is a subroutine call - as in the
1569 second branch in the above example - was incorrectly given the compile-
1570 time error "recursive call could loop indefinitely" because pcre_compile()
1571 was not correctly checking the subroutine for matching a non-empty string.
1572
1573 14. The checks for overrunning compiling workspace could trigger after an
1574 overrun had occurred. This is a "should never occur" error, but it can be
1575 triggered by pathological patterns such as hundreds of nested parentheses.
1576 The checks now trigger 100 bytes before the end of the workspace.
1577
1578 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1579
1580
1581 Version 8.01 19-Jan-2010
1582 ------------------------
1583
1584 1. If a pattern contained a conditional subpattern with only one branch (in
1585 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1586 computed the wrong minimum data length (which is of course zero for such
1587 subpatterns). This could cause incorrect "no match" results.
1588
1589 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1590 the pattern is reset in the first branch, pcre_compile() failed with
1591 "internal error: code overflow at offset...". This happened only when
1592 the reset was to the original external option setting. (An optimization
1593 abstracts leading options settings into an external setting, which was the
1594 cause of this.)
1595
1596 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1597 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1598 assertion pattern did not match (meaning that the assertion was true), it
1599 was incorrectly treated as false if the SKIP had been reached during the
1600 matching. This also applied to assertions used as conditions.
1601
1602 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1603 assertion subpattern, including such a pattern used as a condition,
1604 unpredictable results occurred, instead of the error return
1605 PCRE_ERROR_DFA_UITEM.
1606
1607 5. The C++ GlobalReplace function was not working like Perl for the special
1608 situation when an empty string is matched. It now does the fancy magic
1609 stuff that is necessary.
1610
1611 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1612 removed. (These were left over from very, very early versions of PCRE.)
1613
1614 7. Some cosmetic changes to the code to make life easier when compiling it
1615 as part of something else:
1616
1617 (a) Change DEBUG to PCRE_DEBUG.
1618
1619 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1620 called "current" as "current_branch", to prevent a collision with the
1621 Linux macro when compiled as a kernel module.
1622
1623 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1624 prevent a collision with the Linux macro when compiled as a kernel
1625 module.
1626
1627 8. In pcre_compile() there are some checks for integer overflows that used to
1628 cast potentially large values to (double). This has been changed to that
1629 when building, a check for int64_t is made, and if it is found, it is used
1630 instead, thus avoiding the use of floating point arithmetic. (There is no
1631 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1632 double.
1633
1634 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1635 2005 (difference between two addresses compared to an unsigned value).
1636
1637 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1638 custom one, because of the following reported problem in Windows:
1639
1640 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1641 under Win32.
1642 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1643 therefore missing the function definition.
1644 - The compiler thus generates a "C" signature for the test function.
1645 - The linker fails to find the "C" function.
1646 - PCRE fails to configure if asked to do so against libbz2.
1647
1648 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1649 messages were output:
1650
1651 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1652 rerunning libtoolize, to keep the correct libtool macros in-tree.
1653 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1654
1655 I have done both of these things.
1656
1657 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1658 most of the time, it *can* run out if it is given a pattern that contains a
1659 runaway infinite recursion. I updated the discussion in the pcrestack man
1660 page.
1661
1662 13. Now that we have gone to the x.xx style of version numbers, the minor
1663 version may start with zero. Using 08 or 09 is a bad idea because users
1664 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1665 interpreted as invalid octal numbers. I've updated the previous comment in
1666 configure.ac, and also added a check that gives an error if 08 or 09 are
1667 used.
1668
1669 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1670 causing partial matching to fail when the end of the subject matched \W
1671 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1672
1673 15. There were some discrepancies between the declarations in pcre_internal.h
1674 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1675 their definitions. The declarations used "const uschar *" and the
1676 definitions used USPTR. Even though USPTR is normally defined as "const
1677 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1678 reported that: "This difference in casting confuses some C++ compilers, for
1679 example, SunCC recognizes above declarations as different functions and
1680 generates broken code for hbpcre." I have changed the declarations to use
1681 USPTR.
1682
1683 16. GNU libtool is named differently on some systems. The autogen.sh script now
1684 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1685 (FreeBSD).
1686
1687 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1688 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1689 comment: "Figure out how to create a longlong from a string: strtoll and
1690 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1691 instance, but it only takes 2 args instead of 3!"
1692
1693 18. A subtle bug concerned with back references has been fixed by a change of
1694 specification, with a corresponding code fix. A pattern such as
1695 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1696 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1697 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1698 same bug. Such groups have to be quantified to be useful, or contained
1699 inside another quantified group. (If there's no repetition, the reference
1700 can never match.) The problem arises because, having left the group and
1701 moved on to the rest of the pattern, a later failure that backtracks into
1702 the group uses the captured value from the final iteration of the group
1703 rather than the correct earlier one. I have fixed this in PCRE by forcing
1704 any group that contains a reference to itself to be an atomic group; that
1705 is, there cannot be any backtracking into it once it has completed. This is
1706 similar to recursive and subroutine calls.
1707
1708
1709 Version 8.00 19-Oct-09
1710 ----------------------
1711
1712 1. The table for translating pcre_compile() error codes into POSIX error codes
1713 was out-of-date, and there was no check on the pcre_compile() error code
1714 being within the table. This could lead to an OK return being given in
1715 error.
1716
1717 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1718 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1719 in a Windows environment.
1720
1721 3. The pcregrep --count option prints the count for each file even when it is
1722 zero, as does GNU grep. However, pcregrep was also printing all files when
1723 --files-with-matches was added. Now, when both options are given, it prints
1724 counts only for those files that have at least one match. (GNU grep just
1725 prints the file name in this circumstance, but including the count seems
1726 more useful - otherwise, why use --count?) Also ensured that the
1727 combination -clh just lists non-zero counts, with no names.
1728
1729 4. The long form of the pcregrep -F option was incorrectly implemented as
1730 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1731 but it seems right to fix it, and I didn't think it was worth preserving
1732 the old behaviour.
1733
1734 5. The command line items --regex=pattern and --regexp=pattern were not
1735 recognized by pcregrep, which required --regex pattern or --regexp pattern
1736 (with a space rather than an '='). The man page documented the '=' forms,
1737 which are compatible with GNU grep; these now work.
1738
1739 6. No libpcreposix.pc file was created for pkg-config; there was just
1740 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1741
1742 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1743 when UCP support is not needed, by modifying the Python script that
1744 generates it from Unicode data files. This should not matter if the module
1745 is correctly used as a library, but I received one complaint about 50K of
1746 unwanted data. My guess is that the person linked everything into his
1747 program rather than using a library. Anyway, it does no harm.
1748
1749 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1750 was a minimum greater than 1 for a wide character in a possessive
1751 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1752 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1753 character. Chaos in the form of incorrect output or a compiling loop could
1754 result.
1755
1756 9. The restrictions on what a pattern can contain when partial matching is
1757 requested for pcre_exec() have been removed. All patterns can now be
1758 partially matched by this function. In addition, if there are at least two
1759 slots in the offset vector, the offset of the earliest inspected character
1760 for the match and the offset of the end of the subject are set in them when
1761 PCRE_ERROR_PARTIAL is returned.
1762
1763 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1764 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1765 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1766 and may be more useful for multi-segment matching.
1767
1768 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1769 used to be given if ever the end of the subject was reached; now it is
1770 given only if matching could not proceed because another character was
1771 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1772 string "Z", which now yields "no match" instead of "partial match". In the
1773 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1774 final character ended with (*FAIL).
1775
1776 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1777 if the pattern had a "must contain" character that was already found in the
1778 earlier partial match, unless partial matching was again requested. For
1779 example, with the pattern /dog.(body)?/, the "must contain" character is
1780 "g". If the first part-match was for the string "dog", restarting with
1781 "sbody" failed. This bug has been fixed.
1782
1783 13. The string returned by pcre_dfa_exec() after a partial match has been
1784 changed so that it starts at the first inspected character rather than the
1785 first character of the match. This makes a difference only if the pattern
1786 starts with a lookbehind assertion or \b or \B (\K is not supported by
1787 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1788 matching functions compatible, and I think it's the right thing to do.
1789
1790 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1791 so that the demonstration program is easily available in environments where
1792 PCRE has not been installed from source.
1793
1794 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1795 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1796 library.
1797
1798 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1799 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1800 is not the first non-POSIX option to be added. Clearly some people find
1801 these options useful.
1802
1803 17. If a caller to the POSIX matching function regexec() passes a non-zero
1804 value for nmatch with a NULL value for pmatch, the value of
1805 nmatch is forced to zero.
1806
1807 18. RunGrepTest did not have a test for the availability of the -u option of
1808 the diff command, as RunTest does. It now checks in the same way as
1809 RunTest, and also checks for the -b option.
1810
1811 19. If an odd number of negated classes containing just a single character
1812 interposed, within parentheses, between a forward reference to a named
1813 subpattern and the definition of the subpattern, compilation crashed with
1814 an internal error, complaining that it could not find the referenced
1815 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1816 [The bug was that it was starting one character too far in when skipping
1817 over the character class, thus treating the ] as data rather than
1818 terminating the class. This meant it could skip too much.]
1819
1820 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1821 /g option in pcretest when the pattern contains \K, which makes it possible
1822 to have an empty string match not at the start, even when the pattern is
1823 anchored. Updated pcretest and pcredemo to use this option.
1824
1825 21. If the maximum number of capturing subpatterns in a recursion was greater
1826 than the maximum at the outer level, the higher number was returned, but
1827 with unset values at the outer level. The correct (outer level) value is
1828 now given.
1829
1830 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1831 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1832 make it do so. The string so far is captured, making this feature
1833 compatible with Perl.
1834
1835 23. The tests have been re-organized, adding tests 11 and 12, to make it
1836 possible to check the Perl 5.10 features against Perl 5.10.
1837
1838 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1839 pattern matches a fixed length string. PCRE did not allow this; now it
1840 does. Neither allows recursion.
1841
1842 25. I finally figured out how to implement a request to provide the minimum
1843 length of subject string that was needed in order to match a given pattern.
1844 (It was back references and recursion that I had previously got hung up
1845 on.) This code has now been added to pcre_study(); it finds a lower bound
1846 to the length of subject needed. It is not necessarily the greatest lower
1847 bound, but using it to avoid searching strings that are too short does give
1848 some useful speed-ups. The value is available to calling programs via
1849 pcre_fullinfo().
1850
1851 26. While implementing 25, I discovered to my embarrassment that pcretest had
1852 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1853 study optimizations had never been tested with that matching function.
1854 Oops. What is worse, even when it was passed study data, there was a bug in
1855 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1856 were also very few tests of studied patterns with pcre_dfa_exec().
1857
1858 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1859 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1860 on the other side of the coin, they are no longer allowed to have different
1861 names, because these cannot be distinguished in PCRE, and this has caused
1862 confusion. (This is a difference from Perl.)
1863
1864 28. When duplicate subpattern names are present (necessarily with different
1865 numbers, as required by 27 above), and a test is made by name in a
1866 conditional pattern, either for a subpattern having been matched, or for
1867 recursion in such a pattern, all the associated numbered subpatterns are
1868 tested, and the overall condition is true if the condition is true for any
1869 one of them. This is the way Perl works, and is also more like the way
1870 testing by number works.
1871
1872
1873 Version 7.9 11-Apr-09
1874 ---------------------
1875
1876 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1877 (pcretest), all targets were linked against these libraries. This included
1878 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1879 libraries. This caused unwanted dependencies to be created. This problem
1880 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1881 pcretest is linked with readline.
1882
1883 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1884 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1885 moved outside it again, because FALSE and TRUE are already defined in AIX,
1886 but BOOL is not.
1887
1888 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1889 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1890
1891 4. The pcregrep documentation said spaces were inserted as well as colons (or
1892 hyphens) following file names and line numbers when outputting matching
1893 lines. This is not true; no spaces are inserted. I have also clarified the
1894 wording for the --colour (or --color) option.
1895
1896 5. In pcregrep, when --colour was used with -o, the list of matching strings
1897 was not coloured; this is different to GNU grep, so I have changed it to be
1898 the same.
1899
1900 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1901 each matching line was coloured. Now it goes on to look for further matches
1902 of any of the test patterns, which is the same behaviour as GNU grep.
1903
1904 7. A pattern that could match an empty string could cause pcregrep to loop; it
1905 doesn't make sense to accept an empty string match in pcregrep, so I have
1906 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1907 seems to be how GNU grep behaves. [But see later change 40 for release
1908 8.33.]
1909
1910 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1911 start or after a newline", because the conditional assertion was not being
1912 correctly handled. The rule now is that both the assertion and what follows
1913 in the first alternative must satisfy the test.
1914
1915 9. If auto-callout was enabled in a pattern with a conditional group whose
1916 condition was an assertion, PCRE could crash during matching, both with
1917 pcre_exec() and pcre_dfa_exec().
1918
1919 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1920 used for matching.
1921
1922 11. Unicode property support in character classes was not working for
1923 characters (bytes) greater than 127 when not in UTF-8 mode.
1924
1925 12. Added the -M command line option to pcretest.
1926
1927 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1928
1929 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1930
1931 16. Added comments and documentation about mis-use of no_arg in the C++
1932 wrapper.
1933
1934 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1935 from Martin Jerabek that uses macro names for all relevant character and
1936 string constants.
1937
1938 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1939 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1940 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1941 these, but not everybody uses configure.
1942
1943 19. A conditional group that had only one branch was not being correctly
1944 recognized as an item that could match an empty string. This meant that an
1945 enclosing group might also not be so recognized, causing infinite looping
1946 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1947 with the subject "ab", where knowledge that the repeated group can match
1948 nothing is needed in order to break the loop.
1949
1950 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1951 exec(), but without supplying a callout function, matching went wrong.
1952
1953 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1954 leak if the size of the offset vector was greater than 30. When the vector
1955 is smaller, the saved offsets during recursion go onto a local stack
1956 vector, but for larger vectors malloc() is used. It was failing to free
1957 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1958 error, in fact).
1959
1960 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1961 heapframe that is used only when UTF-8 support is enabled. This caused no
1962 problem, but was untidy.
1963
1964 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1965 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1966 included within another project.
1967
1968 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1969 slightly modified by me:
1970
1971 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1972 not building pcregrep.
1973
1974 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1975 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1976
1977 25. Forward references, both numeric and by name, in patterns that made use of
1978 duplicate group numbers, could behave incorrectly or give incorrect errors,
1979 because when scanning forward to find the reference group, PCRE was not
1980 taking into account the duplicate group numbers. A pattern such as
1981 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1982
1983 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1984 the feature of a custom pointer more persuasive (as requested by a user).
1985
1986 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1987 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1988 pre-defined.
1989
1990 28. Added support for (*UTF8) at the start of a pattern.
1991
1992 29. Arrange for flags added by the "release type" setting in CMake to be shown
1993 in the configuration summary.
1994
1995
1996 Version 7.8 05-Sep-08
1997 ---------------------
1998
1999 1. Replaced UCP searching code with optimized version as implemented for Ad
2000 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2001 stage table and inline lookup instead of a function, giving speed ups of 2
2002 to 5 times on some simple patterns that I tested. Permission was given to
2003 distribute the MultiStage2.py script that generates the tables (it's not in
2004 the tarball, but is in the Subversion repository).
2005
2006 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2007 scripts.
2008
2009 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2010 a group with a zero qualifier. The result of the study could be incorrect,
2011 or the function might crash, depending on the pattern.
2012
2013 4. Caseless matching was not working for non-ASCII characters in back
2014 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2015 It now works when Unicode Property Support is available.
2016
2017 5. In pcretest, an escape such as \x{de} in the data was always generating
2018 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2019 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2020 truncation.
2021
2022 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2023
2024 7. Added two (int) casts to pcregrep when printing the difference of two
2025 pointers, in case they are 64-bit values.
2026
2027 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2028 test 2 if it fails.
2029
2030 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2031 and a #define of that name to empty if it is not externally set. This is to
2032 allow users of MSVC to set it if necessary.
2033
2034 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2035 the convenience functions in the pcre_get.c source file.
2036
2037 11. An option change at the start of a pattern that had top-level alternatives
2038 could cause overwriting and/or a crash. This command provoked a crash in
2039 some environments:
2040
2041 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2042
2043 This potential security problem was recorded as CVE-2008-2371.
2044
2045 12. For a pattern where the match had to start at the beginning or immediately
2046 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2047 pcre_dfa_exec() could read past the end of the passed subject if there was
2048 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2049 pcretest so that it places the subject at the end of its malloc-ed buffer.
2050
2051 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2052 exec() might read past the end of the data buffer in UTF-8 mode.
2053
2054 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2055 the data contained the byte 0x85 as part of a UTF-8 character within its
2056 first line. This applied both to normal and DFA matching.
2057
2058 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2059 /^[^d]*?$/8 failed to match "abc".
2060
2061 16. Added a missing copyright notice to pcrecpp_internal.h.
2062
2063 17. Make it more clear in the documentation that values returned from
2064 pcre_exec() in ovector are byte offsets, not character counts.
2065
2066 18. Tidied a few places to stop certain compilers from issuing warnings.
2067
2068 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2069 supplied by Stefan Weber. I made a further small update for 7.8 because
2070 there is a change of source arrangements: the pcre_searchfuncs.c module is
2071 replaced by pcre_ucd.c.
2072
2073
2074 Version 7.7 07-May-08
2075 ---------------------
2076
2077 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2078 a string to a long long, pretend we don't even have a long long." This is
2079 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2080
2081 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2082 pre-7.6 versions, which defined a global no_arg variable instead of putting
2083 it in the RE class. (See also #8 below.)
2084
2085 3. Remove a line of dead code, identified by coverity and reported by Nuno
2086 Lopes.
2087
2088 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2089
2090 (1) The include/exclude patterns were being applied to the whole pathnames
2091 of files, instead of just to the final components.
2092
2093 (2) If there was more than one level of directory, the subdirectories were
2094 skipped unless they satisfied the include/exclude conditions. This is
2095 inconsistent with GNU grep (and could even be seen as contrary to the
2096 pcregrep specification - which I improved to make it absolutely clear).
2097 The action now is always to scan all levels of directory, and just
2098 apply the include/exclude patterns to regular files.
2099
2100 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2101 --exclude_dir in the tests to avoid scanning .svn directories.
2102
2103 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2104 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2105 doesn't support NULs in patterns.
2106
2107 7. Added some missing "const"s to declarations of static tables in
2108 pcre_compile.c and pcre_dfa_exec.c.
2109
2110 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2111 caused by fix #2 above. (Subsequently also a second patch to fix the
2112 first patch. And a third patch - this was a messy problem.)
2113
2114 9. Applied Craig's patch to remove the use of push_back().
2115
2116 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2117 matching function regexec().
2118
2119 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2120 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2121 references. PCRE supports relative numbers with this syntax (I don't think
2122 Oniguruma does).
2123
2124 12. Previously, a group with a zero repeat such as (...){0} was completely
2125 omitted from the compiled regex. However, this means that if the group
2126 was called as a subroutine from elsewhere in the pattern, things went wrong
2127 (an internal error was given). Such groups are now left in the compiled
2128 pattern, with a new opcode that causes them to be skipped at execution
2129 time.
2130
2131 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2132 to the way PCRE behaves:
2133
2134 (a) A lone ] character is dis-allowed (Perl treats it as data).
2135
2136 (b) A back reference to an unmatched subpattern matches an empty string
2137 (Perl fails the current match path).
2138
2139 (c) A data ] in a character class must be notated as \] because if the
2140 first data character in a class is ], it defines an empty class. (In
2141 Perl it is not possible to have an empty class.) The empty class []
2142 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2143 The negative empty class [^] matches any one character, independently
2144 of the DOTALL setting.
2145
2146 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2147 non-existent subpattern following a character class starting with ']' and
2148 containing () gave an internal compiling error instead of "reference to
2149 non-existent subpattern". Fortunately, when the pattern did exist, the
2150 compiled code was correct. (When scanning forwards to check for the
2151 existence of the subpattern, it was treating the data ']' as terminating
2152 the class, so got the count wrong. When actually compiling, the reference
2153 was subsequently set up correctly.)
2154
2155 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2156 it was being rejected as not supported by pcre_dfa_exec(), even though
2157 other assertions are supported. I have made pcre_dfa_exec() support
2158 (*FAIL).
2159
2160 16. The implementation of 13c above involved the invention of a new opcode,
2161 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2162 cannot be changed at match time, I realized I could make a small
2163 improvement to matching performance by compiling OP_ALLANY instead of
2164 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2165 on the OP_ANY path.
2166
2167 17. Compiling pcretest on Windows with readline support failed without the
2168 following two fixes: (1) Make the unistd.h include conditional on
2169 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2170
2171 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2172 ncurses library to be included for pcretest when ReadLine support is
2173 requested, but also to allow for it to be overridden. This patch came from
2174 Daniel Bergström.
2175
2176 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2177 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2178 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2179 spotting this.
2180
2181
2182 Version 7.6 28-Jan-08
2183 ---------------------
2184
2185 1. A character class containing a very large number of characters with
2186 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2187 overflow.
2188
2189 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2190 HAVE_LONG_LONG is not defined.
2191
2192 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2193 bring them up to date and include new features. This patch includes:
2194
2195 - Fixed PH's badly added libz and libbz2 support.
2196 - Fixed a problem with static linking.
2197 - Added pcredemo. [But later removed - see 7 below.]
2198 - Fixed dftables problem and added an option.
2199 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2200 HAVE_LONG_LONG.
2201 - Added readline support for pcretest.
2202 - Added an listing of the option settings after cmake has run.
2203
2204 4. A user submitted a patch to Makefile that makes it easy to create
2205 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2206 Makefile.am that cause it to include this special target, without
2207 affecting anything else. Note that the same mingw target plus all
2208 the other distribution libraries and programs are now supported
2209 when configuring with CMake (see 6 below) instead of with
2210 Configure/Make.
2211
2212 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2213 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2214 exported in the Windows port". It has not yet been confirmed that the patch
2215 solves the problem, but it does no harm.
2216
2217 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2218 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2219 with CMake, and also correct the comment about stack recursion.
2220
2221 7. Remove the automatic building of pcredemo from the ./configure system and
2222 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2223 of a program that users should build themselves after PCRE is installed, so
2224 building it automatically is not really right. What is more, it gave
2225 trouble in some build environments.
2226
2227 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2228
2229
2230 Version 7.5 10-Jan-08
2231 ---------------------
2232
2233 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2234 values in parens when parsing an RE using the C++ wrapper."
2235
2236 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2237 Characters greater than 255 were excluded from the class instead of being
2238 included.
2239
2240 3. The same bug as (2) above applied to negated POSIX classes such as
2241 [:^space:].
2242
2243 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2244 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2245 I have changed it.
2246
2247 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2248 first named subpattern) and a construct such as (?&a) would reference the
2249 first named subpattern whose name started with "a" (in other words, the
2250 length check was missing). Both these problems are fixed. "Subpattern name
2251 expected" is now given for (?&) (a zero-length name), and this patch also
2252 makes it give the same error for \k'' (previously it complained that that
2253 was a reference to a non-existent subpattern).
2254
2255 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2256 this is right because (?- can be followed by option settings as well as by
2257 digits. I have, however, made the messages clearer.
2258
2259 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2260 than the number used in the conditional) now cause a compile-time error.
2261 This is actually not compatible with Perl, which accepts such patterns, but
2262 treats the conditional as always being FALSE (as PCRE used to), but it
2263 seems to me that giving a diagnostic is better.
2264
2265 8. Change "alphameric" to the more common word "alphanumeric" in comments
2266 and messages.
2267
2268 9. Fix two occurrences of "backslash" in comments that should have been
2269 "backspace".
2270
2271 10. Remove two redundant lines of code that can never be obeyed (their function
2272 was moved elsewhere).
2273
2274 11. The program that makes PCRE's Unicode character property table had a bug
2275 which caused it to generate incorrect table entries for sequences of
2276 characters that have the same character type, but are in different scripts.
2277 It amalgamated them into a single range, with the script of the first of
2278 them. In other words, some characters were in the wrong script. There were
2279 thirteen such cases, affecting characters in the following ranges:
2280
2281 U+002b0 - U+002c1
2282 U+0060c - U+0060d
2283 U+0061e - U+00612
2284 U+0064b - U+0065e
2285 U+0074d - U+0076d
2286 U+01800 - U+01805
2287 U+01d00 - U+01d77
2288 U+01d9b - U+01dbf
2289 U+0200b - U+0200f
2290 U+030fc - U+030fe
2291 U+03260 - U+0327f
2292 U+0fb46 - U+0fbb1
2293 U+10450 - U+1049d
2294
2295 12. The -o option (show only the matching part of a line) for pcregrep was not
2296 compatible with GNU grep in that, if there was more than one match in a
2297 line, it showed only the first of them. It now behaves in the same way as
2298 GNU grep.
2299
2300 13. If the -o and -v options were combined for pcregrep, it printed a blank
2301 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2302 does the same. The return code can be used to tell if there were any
2303 non-matching lines.
2304
2305 14. Added --file-offsets and --line-offsets to pcregrep.
2306
2307 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2308 infinitely looping recursion. The bug was that positive lookaheads were not
2309 being skipped when checking for a possible empty match (negative lookaheads
2310 and both kinds of lookbehind were skipped).
2311
2312 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2313 inclusion of <windows.h> to before rather than after the definition of
2314 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2315
2316 17. Specifying a possessive quantifier with a specific limit for a Unicode
2317 character property caused pcre_compile() to compile bad code, which led at
2318 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2319 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2320 caused the error; without that there was no problem.
2321
2322 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2323
2324 19. Added --enable-pcretest-libreadline.
2325
2326 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2327 RE::GlobalReplace(). As a result, the number of replacements returned was
2328 double what it should be. I removed one of the increments, but Craig sent a
2329 later patch that removed the other one (the right fix) and added unit tests
2330 that check the return values (which was not done before).
2331
2332 21. Several CMake things:
2333
2334 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2335 the names libpcre and libpcreposix, not just pcre and pcreposix.
2336
2337 (2) The above change means that pcretest and pcregrep are now correctly
2338 linked with the newly-built libraries, not previously installed ones.
2339
2340 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2341
2342 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2343 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2344 UTF-8 newline character). The key issue is that the pattern starts .*;
2345 this means that the match must be either at the beginning, or after a
2346 newline. The bug was in the code for advancing after a failed match and
2347 checking that the new position followed a newline. It was not taking
2348 account of UTF-8 characters correctly.
2349
2350 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2351 character classes. PCRE was not treating the sequence [:...:] as a
2352 character class unless the ... were all letters. Perl, however, seems to
2353 allow any characters between [: and :], though of course it rejects as
2354 unknown any "names" that contain non-letters, because all the known class
2355 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2356 for example, whereas PCRE did not - it did not recognize a POSIX character
2357 class. This seemed a bit dangerous, so the code has been changed to be
2358 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2359 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2360 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2361 Perl does, and where it didn't before.
2362
2363 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2364 Windows environments %n is disabled by default.
2365
2366
2367 Version 7.4 21-Sep-07
2368 ---------------------
2369
2370 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2371 means that a class such as [\s] counted as "explicit reference to CR or
2372 LF". That isn't really right - the whole point of the change was to try to
2373 help when there was an actual mention of one of the two characters. So now
2374 the change happens only if \r or \n (or a literal CR or LF) character is
2375 encountered.
2376
2377 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2378 of both had grown to the point where there were only 3 bits left.
2379 Fortunately, there was spare space in the data structure, and so I have
2380 moved the internal flags into a new 16-bit field to free up more option
2381 bits.
2382
2383 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2384 but did not set the internal JCHANGED flag - either of these is enough to
2385 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2386 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2387 start sets both bits.
2388
2389 4. Added options (at build time, compile time, exec time) to change \R from
2390 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2391
2392 5. doc/pcresyntax.html was missing from the distribution.
2393
2394 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2395 compatibility, even though it is no longer used.
2396
2397 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2398 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2399 windows.h file is present (where different names are used). [This was
2400 reversed later after testing - see 16 below.]
2401
2402 8. Changed all #include <config.h> to #include "config.h". There were also
2403 some further <pcre.h> cases that I changed to "pcre.h".
2404
2405 9. When pcregrep was used with the --colour option, it missed the line ending
2406 sequence off the lines that it output.
2407
2408 10. It was pointed out to me that arrays of string pointers cause lots of
2409 relocations when a shared library is dynamically loaded. A technique of
2410 using a single long string with a table of offsets can drastically reduce
2411 these. I have refactored PCRE in four places to do this. The result is
2412 dramatic:
2413
2414 Originally: 290
2415 After changing UCP table: 187
2416 After changing error message table: 43
2417 After changing table of "verbs" 36
2418 After changing table of Posix names 22
2419
2420 Thanks to the folks working on Gregex for glib for this insight.
2421
2422 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2423 unicode-properties was also set.
2424
2425 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2426
2427 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2428 checked only for CRLF.
2429
2430 14. Added casts to pcretest.c to avoid compiler warnings.
2431
2432 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2433
2434 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2435 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2436 entirely. This removes changes made in 7 above.
2437
2438 17. The CMake files have been updated, and there is now more information about
2439 building with CMake in the NON-UNIX-USE document.
2440
2441
2442 Version 7.3 28-Aug-07
2443 ---------------------
2444
2445 1. In the rejigging of the build system that eventually resulted in 7.1, the
2446 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2447 brackets there is not right, since it causes compilers to look for an
2448 installed pcre.h, not the version that is in the source that is being
2449 compiled (which of course may be different). I have changed it back to:
2450
2451 #include "pcre.h"
2452
2453 I have a vague recollection that the change was concerned with compiling in
2454 different directories, but in the new build system, that is taken care of
2455 by the VPATH setting the Makefile.
2456
2457 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2458 when the subject happened to end in the byte 0x85 (e.g. if the last
2459 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2460 characters but of course it shouldn't be taken as a newline when it is part
2461 of another character. The bug was that, for an unlimited repeat of . in
2462 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2463 characters when looking for a newline.
2464
2465 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2466
2467 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2468 in debug output.
2469
2470 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2471 long printing in the pcrecpp unittest when running under MinGW.
2472
2473 6. ESC_K was left out of the EBCDIC table.
2474
2475 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2476 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2477 limit also applies to "virtual nesting" when a pattern is recursive, and in
2478 this case 1000 isn't so big. I have been able to remove this limit at the
2479 expense of backing off one optimization in certain circumstances. Normally,
2480 when pcre_exec() would call its internal match() function recursively and
2481 immediately return the result unconditionally, it uses a "tail recursion"
2482 feature to save stack. However, when a subpattern that can match an empty
2483 string has an unlimited repetition quantifier, it no longer makes this
2484 optimization. That gives it a stack frame in which to save the data for
2485 checking that an empty string has been matched. Previously this was taken
2486 from the 1000-entry workspace that had been reserved. So now there is no
2487 explicit limit, but more stack is used.
2488
2489 8. Applied Daniel's patches to solve problems with the import/export magic
2490 syntax that is required for Windows, and which was going wrong for the
2491 pcreposix and pcrecpp parts of the library. These were overlooked when this
2492 problem was solved for the main library.
2493
2494 9. There were some crude static tests to avoid integer overflow when computing
2495 the size of patterns that contain repeated groups with explicit upper
2496 limits. As the maximum quantifier is 65535, the maximum group length was
2497 set at 30,000 so that the product of these two numbers did not overflow a
2498 32-bit integer. However, it turns out that people want to use groups that
2499 are longer than 30,000 bytes (though not repeat them that many times).
2500 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2501 made it possible to implement the integer overflow checks in a much more
2502 dynamic way, which I have now done. The artificial limitation on group
2503 length has been removed - we now have only the limit on the total length of
2504 the compiled pattern, which depends on the LINK_SIZE setting.
2505
2506 10. Fixed a bug in the documentation for get/copy named substring when
2507 duplicate names are permitted. If none of the named substrings are set, the
2508 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2509 empty string.
2510
2511 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2512 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2513 because the ] is interpreted as the first data character and the
2514 terminating ] is not found. PCRE has been made compatible with Perl in this
2515 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2516 cause memory overwriting.
2517
2518 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2519 string has been matched (to stop an infinite loop). It was not recognizing
2520 a conditional subpattern that could match an empty string if that
2521 subpattern was within another subpattern. For example, it looped when
2522 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2523 condition was not nested. This bug has been fixed.
2524
2525 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2526 past the start of the subject in the presence of bytes with the top bit
2527 set, for example "\x8aBCD".
2528
2529 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2530 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2531
2532 14. Optimized (?!) to (*FAIL).
2533
2534 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2535 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2536 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2537 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2538 does: it's just the validity check that is more restrictive.
2539
2540 16. Inserted checks for integer overflows during escape sequence (backslash)
2541 processing, and also fixed erroneous offset values for syntax errors during
2542 backslash processing.
2543
2544 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2545 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2546
2547 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2548 caused an overrun.
2549
2550 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2551 something other than just ASCII characters) inside a group that had an
2552 unlimited repeat caused a loop at compile time (while checking to see
2553 whether the group could match an empty string).
2554
2555 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2556 [\P{Any}] did so. (Error in the code for printing property names.)
2557
2558 21. An orphan \E inside a character class could cause a crash.
2559
2560 22. A repeated capturing bracket such as (A)? could cause a wild memory
2561 reference during compilation.
2562
2563 23. There are several functions in pcre_compile() that scan along a compiled
2564 expression for various reasons (e.g. to see if it's fixed length for look
2565 behind). There were bugs in these functions when a repeated \p or \P was
2566 present in the pattern. These operators have additional parameters compared
2567 with \d, etc, and these were not being taken into account when moving along
2568 the compiled data. Specifically:
2569
2570 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2571 length.
2572
2573 (b) An item such as \pL+ within a repeated group could cause crashes or
2574 loops.
2575
2576 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2577 "reference to non-existent subpattern" error.
2578
2579 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2580
2581 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2582 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2583
2584 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2585
2586 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2587 character were causing crashes (broken optimization).
2588
2589 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2590 \p or \P) caused a compile-time loop.
2591
2592 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2593 break. For example, the unstudied pattern [\r\n]A does not match the string
2594 "\r\nA" because change 7.0/46 below moves the current point on by two
2595 characters after failing to match at the start. However, the pattern \nA
2596 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2597 the same is true. There doesn't seem any very clean way out of this, but
2598 what I have chosen to do makes the common cases work: PCRE now takes note
2599 of whether there can be an explicit match for \r or \n anywhere in the
2600 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2601 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2602 pattern has explicit CR or LF references.
2603
2604 29. Added (*CR) etc for changing newline setting at start of pattern.
2605
2606
2607 Version 7.2 19-Jun-07
2608 ---------------------
2609
2610 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2611 which is apparently normally available under Windows.
2612
2613 2. Re-jig the pcregrep tests with different newline settings in an attempt
2614 to make them independent of the local environment's newline setting.
2615
2616 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2617
2618 4. Some of the "internals" tests were previously cut out when the link size
2619 was not 2, because the output contained actual offsets. The recent new
2620 "Z" feature of pcretest means that these can be cut out, making the tests
2621 usable with all link sizes.
2622
2623 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2624 stack recursion. This gives a massive performance boost under BSD, but just
2625 a small improvement under Linux. However, it saves one field in the frame
2626 in all cases.
2627
2628 6. Added more features from the forthcoming Perl 5.10:
2629
2630 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2631 recursion call. It refers to the nth most recently opened parentheses.
2632
2633 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2634 to be opened parentheses.
2635
2636 (c) Conditions that refer to capturing parentheses can be specified
2637 relatively, for example, (?(-2)... or (?(+3)...
2638
2639 (d) \K resets the start of the current match so that everything before
2640 is not part of it.
2641
2642 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2643
2644 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2645 reference syntax.
2646
2647 (g) (?| introduces a group in which the numbering of parentheses in each
2648 alternative starts with the same number.
2649
2650 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2651
2652 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2653 PCRE_INFO_JCHANGED.
2654
2655 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2656 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2657 for detecting groups that can match an empty string.
2658
2659 9. A pattern with a very large number of alternatives (more than several
2660 hundred) was running out of internal workspace during the pre-compile
2661 phase, where pcre_compile() figures out how much memory will be needed. A
2662 bit of new cunning has reduced the workspace needed for groups with
2663 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2664 workspace instead of running out of the 4096 that are available.
2665
2666 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2667
2668 11. Applied patch from Google to remove an optimization that didn't quite work.
2669 The report of the bug said:
2670
2671 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2672 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2673 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2674
2675 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2676 it matched the wrong number of bytes.
2677
2678
2679 Version 7.1 24-Apr-07
2680 ---------------------
2681
2682 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2683 that is more "standard", making use of automake and other Autotools. There
2684 is some re-arrangement of the files and adjustment of comments consequent
2685 on this.
2686
2687 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2688 for recursive directory scanning broke on some systems because the files
2689 are not scanned in any specific order and on different systems the order
2690 was different. A call to "sort" has been inserted into RunGrepTest for the
2691 approprate test as a short-term fix. In the longer term there may be an
2692 alternative.
2693
2694 3. I had an email from Eric Raymond about problems translating some of PCRE's
2695 man pages to HTML (despite the fact that I distribute HTML pages, some
2696 people do their own conversions for various reasons). The problems
2697 concerned the use of low-level troff macros .br and .in. I have therefore
2698 removed all such uses from the man pages (some were redundant, some could
2699 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2700 HTML has been updated to handle .nf/.fi and to complain if it encounters
2701 .br or .in.
2702
2703 4. Updated comments in configure.ac that get placed in config.h.in and also
2704 arranged for config.h to be included in the distribution, with the name
2705 config.h.generic, for the benefit of those who have to compile without
2706 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2707
2708 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2709 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2710 makevp.bat for the current PCRE, using the additional files
2711 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2712
2713 6. A Windows user reported a minor discrepancy with test 2, which turned out
2714 to be caused by a trailing space on an input line that had got lost in his
2715 copy. The trailing space was an accident, so I've just removed it.
2716
2717 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2718 that is needed.
2719
2720 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2721 as "const" (a) because they are and (b) because it helps the PHP
2722 maintainers who have recently made a script to detect big data structures
2723 in the php code that should be moved to the .rodata section. I remembered
2724 to update Builducptable as well, so it won't revert if ucptable.h is ever
2725 re-created.
2726
2727 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2728 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2729 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2730 support is not required. This saves 1.5-2K of code, which is important in
2731 some applications.
2732
2733 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2734 so as not to refer to the tables, even though these functions will never be
2735 called when UTF-8 support is disabled. Otherwise there are problems with a
2736 shared library.
2737
2738 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2739
2740 (a) It was defining its arguments as char * instead of void *.
2741
2742 (b) It was assuming that all moves were upwards in memory; this was true
2743 a long time ago when I wrote it, but is no longer the case.
2744
2745 The emulated memove() is provided for those environments that have neither
2746 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2747 is clearly not the case, as these two bugs were recently reported.
2748
2749 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2750 and Detrail to create the HTML documentation, the .txt form of the man
2751 pages, and it removes trailing spaces from listed files. It also creates
2752 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2753 case, it wraps all the #defines with #ifndefs. This script should be run
2754 before "make dist".
2755
2756 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2757 with Unicode property support.
2758
2759 (a) For a maximizing quantifier, if the two different cases of the
2760 character were of different lengths in their UTF-8 codings (there are
2761 some cases like this - I found 11), and the matching function had to
2762 back up over a mixture of the two cases, it incorrectly assumed they
2763 were both the same length.
2764
2765 (b) When PCRE was configured to use the heap rather than the stack for
2766 recursion during matching, it was not correctly preserving the data for
2767 the other case of a UTF-8 character when checking ahead for a match
2768 while processing a minimizing repeat. If the check also involved
2769 matching a wide character, but failed, corruption could cause an
2770 erroneous result when trying to check for a repeat of the original
2771 character.
2772
2773 13. Some tidying changes to the testing mechanism:
2774
2775 (a) The RunTest script now detects the internal link size and whether there
2776 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2777 values substituted by "configure". (The RunGrepTest script already did
2778 this for UTF-8.) The configure.ac script no longer substitutes the
2779 relevant variables.
2780
2781 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2782 with length and offset values. This means that the output is different
2783 for different internal link sizes. Test 2 is skipped for link sizes
2784 other than 2 because of this, bypassing the problem. Unfortunately,
2785 there was also a test in test 3 (the locale tests) that used /B and
2786 failed for link sizes other than 2. Rather than cut the whole test out,
2787 I have added a new /Z option to pcretest that replaces the length and
2788 offset values with spaces. This is now used to make test 3 independent
2789 of link size. (Test 2 will be tidied up later.)
2790
2791 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2792 segmentation fault instead of returning the appropriate error message.
2793
2794 15. In multiline mode when the newline sequence was set to "any", the pattern
2795 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2796 This doesn't seem right; it now treats the CRLF combination as the line
2797 ending, and so does not match in that case. It's only a pattern such as ^$
2798 that would hit this one: something like ^ABC$ would have failed after \r
2799 and then tried again after \r\n.
2800
2801 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2802 in an attempt to make files that differ only in their line terminators
2803 compare equal. This works on Linux.
2804
2805 17. Under certain error circumstances pcregrep might try to free random memory
2806 as it exited. This is now fixed, thanks to valgrind.
2807
2808 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2809 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2810 was because its rules for how to advance for /g after matching an empty
2811 string at the end of a line did not allow for this case. They now check for
2812 it specially.
2813
2814 20. pcretest is supposed to handle patterns and data of any length, by
2815 extending its buffers when necessary. It was getting this wrong when the
2816 buffer for a data line had to be extended.
2817
2818 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2819 CRLF as a newline sequence.
2820
2821 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2822 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2823 I have nevertheless tidied it up.
2824
2825 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2826
2827 24. Added a man page for pcre-config.
2828
2829
2830 Version 7.0 19-Dec-06
2831 ---------------------
2832
2833 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2834 moving to gcc 4.1.1.
2835
2836 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2837 sys/time.h, which is documented as needed for this function. It doesn't
2838 seem to matter on Linux, but it showed up on some releases of OS X.
2839
2840 3. It seems that there are systems where bytes whose values are greater than
2841 127 match isprint() in the "C" locale. The "C" locale should be the
2842 default when a C program starts up. In most systems, only ASCII printing
2843 characters match isprint(). This difference caused the output from pcretest
2844 to vary, making some of the tests fail. I have changed pcretest so that:
2845
2846 (a) When it is outputting text in the compiled version of a pattern, bytes
2847 other than 32-126 are always shown as hex escapes.
2848
2849 (b) When it is outputting text that is a matched part of a subject string,
2850 it does the same, unless a different locale has been set for the match
2851 (using the /L modifier). In this case, it uses isprint() to decide.
2852
2853 4. Fixed a major bug that caused incorrect computation of the amount of memory
2854 required for a compiled pattern when options that changed within the
2855 pattern affected the logic of the preliminary scan that determines the
2856 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2857 that the computed length was too small. The symptoms of this bug were
2858 either the PCRE error "internal error: code overflow" from pcre_compile(),
2859 or a glibc crash with a message such as "pcretest: free(): invalid next
2860 size (fast)". Examples of patterns that provoked this bug (shown in
2861 pcretest format) are:
2862
2863 /(?-x: )/x
2864 /(?x)(?-x: \s*#\s*)/
2865 /((?i)[\x{c0}])/8
2866 /(?i:[\x{c0}])/8
2867
2868 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2869 is now done differently.
2870
2871 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2872 wrapper classes; (b) implement a new function in the C++ scanner that is
2873 more efficient than the old way of doing things because it avoids levels of
2874 recursion in the regex matching; (c) add a paragraph to the documentation
2875 for the FullMatch() function.
2876
2877 6. The escape sequence \n was being treated as whatever was defined as
2878 "newline". Not only was this contrary to the documentation, which states
2879 that \n is character 10 (hex 0A), but it also went horribly wrong when
2880 "newline" was defined as CRLF. This has been fixed.
2881
2882 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2883 was being set to -1 for the "end of line" case (supposedly a value that no
2884 character can have). Though this value is never used (the check for end of
2885 line is "zero bytes in current character"), it caused compiler complaints.
2886 I've changed it to 0xffffffff.
2887
2888 8. In pcre_version.c, the version string was being built by a sequence of
2889 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2890 string (as it is for production releases) called a macro with an empty
2891 argument. The C standard says the result of this is undefined. The gcc
2892 compiler treats it as an empty string (which was what was wanted) but it is
2893 reported that Visual C gives an error. The source has been hacked around to
2894 avoid this problem.
2895
2896 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2897 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2898 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2899 of them did).
2900
2901 10. Originally, pcretest opened its input and output without "b"; then I was
2902 told that "b" was needed in some environments, so it was added for release
2903 5.0 to both the input and output. (It makes no difference on Unix-like
2904 systems.) Later I was told that it is wrong for the input on Windows. I've
2905 now abstracted the modes into two macros, to make it easier to fiddle with
2906 them, and removed "b" from the input mode under Windows.
2907
2908 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2909
2910 12. Added -help and --help to pcretest as an official way of being reminded
2911 of the options.
2912
2913 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2914 and pcrecpp.cc because they annoy compilers at high warning levels.
2915
2916 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2917
2918 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2919 scripts are not C programs :-) and which was not noticed because it works
2920 on Linux.
2921
2922 16. pcretest is supposed to handle any length of pattern and data line (as one
2923 line or as a continued sequence of lines) by extending its input buffer if
2924 necessary. This feature was broken for very long pattern lines, leading to
2925 a string of junk being passed to pcre_compile() if the pattern was longer
2926 than about 50K.
2927
2928 17. I have done a major re-factoring of the way pcre_compile() computes the
2929 amount of memory needed for a compiled pattern. Previously, there was code
2930 that made a preliminary scan of the pattern in order to do this. That was
2931 OK when PCRE was new, but as the facilities have expanded, it has become
2932 harder and harder to keep it in step with the real compile phase, and there
2933 have been a number of bugs (see for example, 4 above). I have now found a
2934 cunning way of running the real compile function in a "fake" mode that
2935 enables it to compute how much memory it would need, while actually only
2936 ever using a few hundred bytes of working memory and without too many
2937 tests of the mode. This should make future maintenance and development
2938 easier. A side effect of this work is that the limit of 200 on the nesting
2939 depth of parentheses has been removed (though this was never a serious
2940 limitation, I suspect). However, there is a downside: pcre_compile() now
2941 runs more slowly than before (30% or more, depending on the pattern). I
2942 hope this isn't a big issue. There is no effect on runtime performance.
2943
2944 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2945 newline (only possible for the last line of a file) and it was a
2946 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2947
2948 19. Added additional timing features to pcretest. (1) The -tm option now times
2949 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2950 separate command line item, by a number that specifies the number of
2951 repeats to use when timing. The default is 50000; this gives better
2952 precision, but takes uncomfortably long for very large patterns.
2953
2954 20. Extended pcre_study() to be more clever in cases where a branch of a
2955 subpattern has no definite first character. For example, (a*|b*)[cd] would
2956 previously give no result from pcre_study(). Now it recognizes that the
2957 first character must be a, b, c, or d.
2958
2959 21. There was an incorrect error "recursive call could loop indefinitely" if
2960 a subpattern (or the entire pattern) that was being tested for matching an
2961 empty string contained only one non-empty item after a nested subpattern.
2962 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2963 incorrectly, because the \d was being skipped in the check.
2964
2965 22. The pcretest program now has a new pattern option /B and a command line
2966 option -b, which is equivalent to adding /B to every pattern. This causes
2967 it to show the compiled bytecode, without the additional information that
2968 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2969 is the same as /B/I).
2970
2971 23. A new optimization is now able automatically to treat some sequences such
2972 as a*b as a*+b. More specifically, if something simple (such as a character
2973 or a simple class like \d) has an unlimited quantifier, and is followed by
2974 something that cannot possibly match the quantified thing, the quantifier
2975 is automatically "possessified".
2976
2977 24. A recursive reference to a subpattern whose number was greater than 39
2978 went wrong under certain circumstances in UTF-8 mode. This bug could also
2979 have affected the operation of pcre_study().
2980
2981 25. Realized that a little bit of performance could be had by replacing
2982 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2983
2984 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2985
2986 27. Possessive quantifiers such as a++ were previously implemented by turning
2987 them into atomic groups such as ($>a+). Now they have their own opcodes,
2988 which improves performance. This includes the automatically created ones
2989 from 23 above.
2990
2991 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2992 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2993 the first matched character to be a colon. This applied both to named and
2994 numbered groups.
2995
2996 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2997
2998 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2999 building PCRE on Windows easier, so I have included it in the distribution.
3000
3001 31. There is now a check in pcretest against a ridiculously large number being
3002 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3003 loop, the loop is abandoned.
3004
3005 32. Forward references to subpatterns in conditions such as (?(2)...) where
3006 subpattern 2 is defined later cause pcre_compile() to search forwards in
3007 the pattern for the relevant set of parentheses. This search went wrong
3008 when there were unescaped parentheses in a character class, parentheses
3009 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3010
3011 33. "Subroutine" calls and backreferences were previously restricted to
3012 referencing subpatterns earlier in the regex. This restriction has now
3013 been removed.
3014
3015 34. Added a number of extra features that are going to be in Perl 5.10. On the
3016 whole, these are just syntactic alternatives for features that PCRE had
3017 previously implemented using the Python syntax or my own invention. The
3018 other formats are all retained for compatibility.
3019
3020 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3021 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3022 also .NET compatible.
3023
3024 (b) A recursion or subroutine call to a named group can now be defined as
3025 (?&name) as well as (?P>name).
3026
3027 (c) A backreference to a named group can now be defined as \k<name> or
3028 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3029 5.10, are also .NET compatible.
3030
3031 (d) A conditional reference to a named group can now use the syntax
3032 (?(<name>) or (?('name') as well as (?(name).
3033
3034 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3035 groups (named and numbered) that are never evaluated inline, but can be
3036 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3037 is always false. There may be only one alternative in such a group.
3038
3039 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3040 as the simple (?(R). The condition is true only if the most recent
3041 recursion is that of the given number or name. It does not search out
3042 through the entire recursion stack.
3043
3044 (g) The escape \gN or \g{N} has been added, where N is a positive or
3045 negative number, specifying an absolute or relative reference.
3046
3047 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3048 some "unreachable code" warnings.
3049
3050 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3051 things, this adds five new scripts.
3052
3053 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3054 There were also incompatibilities regarding the handling of \Q..\E inside
3055 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3056 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3057
3058 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3059 matches an empty string, and forcibly breaks the loop. There were bugs in
3060 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3061 against aaaa the result was just "a" rather than "aaaa", for example. Two
3062 separate and independent bugs (that affected different cases) have been
3063 fixed.
3064
3065 39. Refactored the code to abolish the use of different opcodes for small
3066 capturing bracket numbers. This is a tidy that I avoided doing when I
3067 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3068 The new approach is not only tidier, it makes it possible to reduce the
3069 memory needed to fix the previous bug (38).
3070
3071 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3072 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3073 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3074 mode.
3075
3076 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3077 report.
3078
3079 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3080 copy construction and assignment in the C++ wrapper.
3081
3082 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3083 couple of bugs that could have given wrong results in the "--newline=crlf"
3084 case.
3085
3086 44. Added a number of casts and did some reorganization of signed/unsigned int
3087 variables following suggestions from Dair Grant. Also renamed the variable
3088 "this" as "item" because it is a C++ keyword.
3089
3090 45. Arranged for dftables to add
3091
3092 #include "pcre_internal.h"
3093
3094 to pcre_chartables.c because without it, gcc 4.x may remove the array
3095 definition from the final binary if PCRE is built into a static library and
3096 dead code stripping is activated.
3097
3098 46. For an unanchored pattern, if a match attempt fails at the start of a
3099 newline sequence, and the newline setting is CRLF or ANY, and the next two
3100 characters are CRLF, advance by two characters instead of one.
3101
3102
3103 Version 6.7 04-Jul-06
3104 ---------------------
3105
3106 1. In order to handle tests when input lines are enormously long, pcretest has
3107 been re-factored so that it automatically extends its buffers when
3108 necessary. The code is crude, but this _is_ just a test program. The
3109 default size has been increased from 32K to 50K.
3110
3111 2. The code in pcre_study() was using the value of the re argument before
3112 testing it for NULL. (Of course, in any sensible call of the function, it
3113 won't be NULL.)
3114
3115 3. The memmove() emulation function in pcre_internal.h, which is used on
3116 systems that lack both memmove() and bcopy() - that is, hardly ever -
3117 was missing a "static" storage class specifier.
3118
3119 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3120 containing an extended class (one that cannot be represented by a bitmap
3121 because it contains high-valued characters or Unicode property items, e.g.
3122 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3123 pattern, but PCRE should not loop if you do not (it no longer does).
3124 [Detail: two cases were found: (a) a repeated subpattern containing an
3125 extended class; (b) a recursive reference to a subpattern that followed a
3126 previous extended class. It wasn't skipping over the extended class
3127 correctly when UTF-8 mode was not set.]
3128
3129 5. A negated single-character class was not being recognized as fixed-length
3130 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3131 compile error "lookbehind assertion is not fixed length".
3132
3133 6. The RunPerlTest auxiliary script was showing an unexpected difference
3134 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3135 write a Perl script that can interpret lines of an input file either as
3136 byte characters or as UTF-8, which is what "perltest" was being required to
3137 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3138 can't do is switch easily at run time between having the "use utf8;" pragma
3139 or not. In the end, I fudged it by using the RunPerlTest script to insert
3140 "use utf8;" explicitly for the UTF-8 tests.
3141
3142 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3143 the end of the subject string, contrary to the documentation and to what
3144 Perl does. This was true of both matching functions. Now it matches only at
3145 the start of the subject and immediately after *internal* newlines.
3146
3147 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3148 a pointer to an int instead of a pointer to an unsigned long int. This
3149 caused problems on 64-bit systems.
3150
3151 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3152 instance of the 'standard' template library not being so standard".
3153
3154 10. There was no check on the number of named subpatterns nor the maximum
3155 length of a subpattern name. The product of these values is used to compute
3156 the size of the memory block for a compiled pattern. By supplying a very
3157 long subpattern name and a large number of named subpatterns, the size
3158 computation could be caused to overflow. This is now prevented by limiting
3159 the length of names to 32 characters, and the number of named subpatterns
3160 to 10,000.
3161
3162 11. Subpatterns that are repeated with specific counts have to be replicated in
3163 the compiled pattern. The size of memory for this was computed from the
3164 length of the subpattern and the repeat count. The latter is limited to
3165 65535, but there was no limit on the former, meaning that integer overflow
3166 could in principle occur. The compiled length of a repeated subpattern is
3167 now limited to 30,000 bytes in order to prevent this.
3168
3169 12. Added the optional facility to have named substrings with the same name.
3170
3171 13. Added the ability to use a named substring as a condition, using the
3172 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3173 are numbers (not recommended). Forward references are permitted.
3174
3175 14. Added forward references in named backreferences (if you see what I mean).
3176
3177 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3178 pattern could run off the end of the subject. For example, the pattern
3179 "(?s)(.{1,5})"8 did this with the subject "ab".
3180
3181 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3182 PCRE_CASELESS was set when matching characters that were quantified with ?
3183 or *.
3184
3185 17. A character class other than a single negated character that had a minimum
3186 but no maximum quantifier - for example [ab]{6,} - was not handled
3187 correctly by pce_dfa_exec(). It would match only one character.
3188
3189 18. A valid (though odd) pattern that looked like a POSIX character
3190 class but used an invalid character after [ (for example [[,abc,]]) caused
3191 pcre_compile() to give the error "Failed: internal error: code overflow" or
3192 in some cases to crash with a glibc free() error. This could even happen if
3193 the pattern terminated after [[ but there just happened to be a sequence of
3194 letters, a binary zero, and a closing ] in the memory that followed.
3195
3196 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3197 over the years. Originally (before any Unicode support), just the bottom 8
3198 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3199 output from "man perlunicode" includes this:
3200
3201 The regular expression compiler produces polymorphic opcodes. That
3202 is, the pattern adapts to the data and automatically switches to
3203 the Unicode character scheme when presented with Unicode data--or
3204 instead uses a traditional byte scheme when presented with byte
3205 data.
3206
3207 Sadly, a wide octal escape does not cause a switch, and in a string with
3208 no other multibyte characters, these octal escapes are treated as before.
3209 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3210 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3211 Unicode string.
3212
3213 I have not perpetrated such confusion in PCRE. Up till now, it took just
3214 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3215 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3216 translate to the appropriate multibyte character.
3217
3218 29. Applied some refactoring to reduce the number of warnings from Microsoft
3219 and Borland compilers. This has included removing the fudge introduced
3220 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3221 a warning about an unused variable.
3222
3223 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3224 characters since release 4.0, because Perl (from release 5.004) does not.
3225 [Or at least, is documented not to: some releases seem to be in conflict
3226 with the documentation.] However, when a pattern was studied with
3227 pcre_study() and all its branches started with \s, PCRE still included VT
3228 as a possible starting character. Of course, this did no harm; it just
3229 caused an unnecessary match attempt.
3230
3231 22. Removed a now-redundant internal flag bit that recorded the fact that case
3232 dependency changed within the pattern. This was once needed for "required
3233 byte" processing, but is no longer used. This recovers a now-scarce options
3234 bit. Also moved the least significant internal flag bit to the most-
3235 significant bit of the word, which was not previously used (hangover from
3236 the days when it was an int rather than a uint) to free up another bit for
3237 the future.
3238
3239 23. Added support for CRLF line endings as well as CR and LF. As well as the
3240 default being selectable at build time, it can now be changed at runtime
3241 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3242 specify that it is scanning data with non-default line endings.
3243
3244 24. Changed the definition of CXXLINK to make it agree with the definition of
3245 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3246
3247 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3248 recursions. This makes a big different to stack usage for some patterns.
3249
3250 26. If a subpattern containing a named recursion or subroutine reference such
3251 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3252 the space required for the compiled pattern went wrong and gave too small a
3253 value. Depending on the environment, this could lead to "Failed: internal
3254 error: code overflow at offset 49" or "glibc detected double free or
3255 corruption" errors.
3256
3257 27. Applied patches from Google (a) to support the new newline modes and (b) to
3258 advance over multibyte UTF-8 characters in GlobalReplace.
3259
3260 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3261 difference for some implementation of PCRE in some Windows version.
3262
3263 29. Added some extra testing facilities to pcretest:
3264
3265 \q<number> in a data line sets the "match limit" value
3266 \Q<number> in a data line sets the "match recursion limt" value
3267 -S <number> sets the stack size, where <number> is in megabytes
3268
3269 The -S option isn't available for Windows.
3270
3271
3272 Version 6.6 06-Feb-06
3273 ---------------------
3274
3275 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3276 in pcreposix.h. I have copied the definition from pcre.h.
3277
3278 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3279 because pcre.h is no longer a built file.
3280
3281 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3282 not normally included in the compiled code.
3283
3284
3285 Version 6.5 01-Feb-06
3286 ---------------------
3287
3288 1. When using the partial match feature with pcre_dfa_exec(), it was not
3289 anchoring the second and subsequent partial matches at the new starting
3290 point. This could lead to incorrect results. For example, with the pattern
3291 /1234/, partially matching against "123" and then "a4" gave a match.
3292
3293 2. Changes to pcregrep:
3294
3295 (a) All non-match returns from pcre_exec() were being treated as failures
3296 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3297 error message is output. Some extra information is given for the
3298 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3299 probably the only errors that are likely to be caused by users (by
3300 specifying a regex that has nested indefinite repeats, for instance).
3301 If there are more than 20 of these errors, pcregrep is abandoned.
3302
3303 (b) A binary zero was treated as data while matching, but terminated the
3304 output line if it was written out. This has been fixed: binary zeroes
3305 are now no different to any other data bytes.
3306
3307 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3308 used to set a locale for matching. The --locale=xxxx long option has
3309 been added (no short equivalent) to specify a locale explicitly on the
3310 pcregrep command, overriding the environment variables.
3311
3312 (d) When -B was used with -n, some line numbers in the output were one less
3313 than they should have been.
3314
3315 (e) Added the -o (--only-matching) option.
3316
3317 (f) If -A or -C was used with -c (count only), some lines of context were
3318 accidentally printed for the final match.
3319
3320 (g) Added the -H (--with-filename) option.
3321
3322 (h) The combination of options -rh failed to suppress file names for files
3323 that were found from directory arguments.
3324
3325 (i) Added the -D (--devices) and -d (--directories) options.
3326
3327 (j) Added the -F (--fixed-strings) option.
3328
3329 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3330
3331 (l) Added the --colo(u)r option.
3332
3333 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3334 is not present by default.
3335
3336 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3337 items such as (?R) or (?1), when the recursion could match a number of
3338 alternatives. If it matched one of the alternatives, but subsequently,
3339 outside the recursion, there was a failure, the code tried to back up into
3340 the recursion. However, because of the way PCRE is implemented, this is not
3341 possible, and the result was an incorrect result from the match.
3342
3343 In order to prevent this happening, the specification of recursion has
3344 been changed so that all such subpatterns are automatically treated as
3345 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3346
3347 4. I had overlooked the fact that, in some locales, there are characters for
3348 which isalpha() is true but neither isupper() nor islower() are true. In
3349 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3350 and ordfeminine) are like this. This affected the treatment of \w and \W
3351 when they appeared in character classes, but not when they appeared outside
3352 a character class. The bit map for "word" characters is now created
3353 separately from the results of isalnum() instead of just taking it from the
3354 upper, lower, and digit maps. (Plus the underscore character, of course.)
3355
3356 5. The above bug also affected the handling of POSIX character classes such as
3357 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3358 permanent tables. Instead, the bit maps for such a class were previously
3359 created as the appropriate unions of the upper, lower, and digit bitmaps.
3360 Now they are created by subtraction from the [[:word:]] class, which has
3361 its own bitmap.
3362
3363 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3364 It is created by subtracting the vertical space characters (\x09, \x0a,
3365 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3366 subtraction was done in the overall bitmap for a character class, meaning
3367 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3368 be recognized. This bug has been fixed.
3369
3370 7. Patches from the folks at Google:
3371
3372 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3373 real life, but is still worth protecting against".
3374
3375 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3376 regular expressions".
3377
3378 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3379 have it.
3380
3381 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3382 "configure" and the latter not, in order to fix a problem somebody had
3383 with compiling the Arg class on HP-UX.
3384
3385 (e) Improve the error-handling of the C++ wrapper a little bit.
3386
3387 (f) New tests for checking recursion limiting.
3388
3389 8. The pcre_memmove() function, which is used only if the environment does not
3390 have a standard memmove() function (and is therefore rarely compiled),
3391 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3392 returning a result (though PCRE never actually uses the result).
3393
3394 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3395 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3396 returned instead of calling malloc() with an overflowing number that would
3397 most likely cause subsequent chaos.
3398
3399 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3400
3401 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3402 with this option is matched, the nmatch and pmatch options of regexec() are
3403 ignored.
3404
3405 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3406 provided in case anyone wants to the the POSIX interface with UTF-8
3407 strings.
3408
3409 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3410 C++ linking (needed for some HP-UX environments).
3411
3412 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3413 (unused parameter) and in the pcre_printint() function (omitted "default"
3414 switch label when the default is to do nothing).
3415
3416 15. Added some code to make it possible, when PCRE is compiled as a C++
3417 library, to replace subject pointers for pcre_exec() with a smart pointer
3418 class, thus making it possible to process discontinuous strings.
3419
3420 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3421 much the same function. They were added by different people who were trying
3422 to make PCRE easy to compile on non-Unix systems. It has been suggested
3423 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3424 for compiling on Windows systems. I have therefore replaced it with
3425 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3426 defaults to "extern" for C or "extern C" for C++, which works fine on
3427 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3428 SCOPE with something explicit in config.h. In addition:
3429
3430 (a) pcreposix.h still had just "extern" instead of either of these macros;
3431 I have replaced it with PCRE_DATA_SCOPE.
3432
3433 (b) Functions such as _pcre_xclass(), which are internal to the library,
3434 but external in the C sense, all had PCRE_EXPORT in their definitions.
3435 This is apparently wrong for the Windows case, so I have removed it.
3436 (It makes no difference on Unix-like systems.)
3437
3438 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3439 of recursive calls to match(). This is different to MATCH_LIMIT because
3440 that limits the total number of calls to match(), not all of which increase
3441 the depth of recursion. Limiting the recursion depth limits the amount of
3442 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3443 when PCRE is compiled, and changed at run time. A patch from Google adds
3444 this functionality to the C++ interface.
3445
3446 18. Changes to the handling of Unicode character properties:
3447
3448 (a) Updated the table to Unicode 4.1.0.
3449
3450 (b) Recognize characters that are not in the table as "Cn" (undefined).
3451
3452 (c) I revised the way the table is implemented to a much improved format
3453 which includes recognition of ranges. It now supports the ranges that
3454 are defined in UnicodeData.txt, and it also amalgamates other
3455 characters into ranges. This has reduced the number of entries in the
3456 table from around 16,000 to around 3,000, thus reducing its size
3457 considerably. I realized I did not need to use a tree structure after
3458 all - a binary chop search is just as efficient. Having reduced the
3459 number of entries, I extended their size from 6 bytes to 8 bytes to
3460 allow for more data.
3461
3462 (d) Added support for Unicode script names via properties such as \p{Han}.
3463
3464 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3465 matching that character.
3466
3467 20. When matching a repeated Unicode property with a minimum greater than zero,
3468 (for example \pL{2,}), PCRE could look past the end of the subject if it
3469 reached it while seeking the minimum number of characters. This could
3470 happen only if some of the characters were more than one byte long, because
3471 there is a check for at least the minimum number of bytes.
3472
3473 21. Refactored the implementation of \p and \P so as to be more general, to
3474 allow for more different types of property in future. This has changed the
3475 compiled form incompatibly. Anybody with saved compiled patterns that use
3476 \p or \P will have to recompile them.
3477
3478 22. Added "Any" and "L&" to the supported property types.
3479
3480 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3481 but give a compile time error if the value is greater than 0xff.
3482
3483 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3484 accidentally not being installed or uninstalled.
3485
3486 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3487 made were to insert the current release number. This seemed silly, because
3488 it made things harder for people building PCRE on systems that don't run
3489 "configure". I have turned pcre.h into a distributed file, no longer built
3490 by "configure", with the version identification directly included. There is
3491 no longer a pcre.h.in file.
3492
3493 However, this change necessitated a change to the pcre-config script as
3494 well. It is built from pcre-config.in, and one of the substitutions was the
3495 release number. I have updated configure.ac so that ./configure now finds
3496 the release number by grepping pcre.h.
3497
3498 26. Added the ability to run the tests under valgrind.
3499
3500
3501 Version 6.4 05-Sep-05
3502 ---------------------
3503
3504 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3505 "--" to be printed when multiple files were scanned, even when none of the
3506 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3507 consider it to be a bug, and have restored the previous behaviour.
3508
3509 2. A couple of code tidies to get rid of compiler warnings.
3510
3511 3. The pcretest program used to cheat by referring to symbols in the library
3512 whose names begin with _pcre_. These are internal symbols that are not
3513 really supposed to be visible externally, and in some environments it is
3514 possible to suppress them. The cheating is now confined to including
3515 certain files from the library's source, which is a bit cleaner.
3516
3517 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3518 file's purpose clearer.
3519
3520 5. Reorganized pcre_ucp_findchar().
3521
3522
3523 Version 6.3 15-Aug-05
3524 ---------------------
3525
3526 1. The file libpcre.pc.in did not have general read permission in the tarball.
3527
3528 2. There were some problems when building without C++ support:
3529
3530 (a) If C++ support was not built, "make install" and "make test" still
3531 tried to test it.
3532
3533 (b) There were problems when the value of CXX was explicitly set. Some
3534 changes have been made to try to fix these, and ...
3535
3536 (c) --disable-cpp can now be used to explicitly disable C++ support.
3537
3538 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3539 backslash in a target when C++ was disabled. This confuses some
3540 versions of "make", apparently. Using an intermediate variable solves
3541 this. (Same for CPP_LOBJ.)
3542
3543 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3544 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3545 necessary on certain architectures.
3546
3547 4. Added a setting of -export-symbols-regex to the link command to remove
3548 those symbols that are exported in the C sense, but actually are local
3549 within the library, and not documented. Their names all begin with
3550 "_pcre_". This is not a perfect job, because (a) we have to except some
3551 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3552 available (and never for static libraries). I have made a note to try to
3553 find a way round (a) in the future.
3554
3555
3556 Version 6.2 01-Aug-05
3557 ---------------------
3558
3559 1. There was no test for integer overflow of quantifier values. A construction
3560 such as {1111111111111111} would give undefined results. What is worse, if
3561 a minimum quantifier for a parenthesized subpattern overflowed and became
3562 negative, the calculation of the memory size went wrong. This could have
3563 led to memory overwriting.
3564
3565 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3566
3567 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3568 operating environments where this matters.
3569
3570 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3571 PCRE options from within the C++ wrapper.
3572
3573 5. Named capturing subpatterns were not being correctly counted when a pattern
3574 was compiled. This caused two problems: (a) If there were more than 100
3575 such subpatterns, the calculation of the memory needed for the whole
3576 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3577 back references of the form \12, where the number was greater than 9, were
3578 not recognized as back references, even though there were sufficient
3579 previous subpatterns.
3580
3581 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3582 versions of gcc, e.g. 2.95.4.
3583
3584
3585 Version 6.1 21-Jun-05
3586 ---------------------
3587
3588 1. There was one reference to the variable "posix" in pcretest.c that was not
3589 surrounded by "#if !defined NOPOSIX".
3590
3591 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3592 the cross-check on the old pcre_info() function, for the benefit of the
3593 cut-down version of PCRE that is currently imported into Exim.
3594
3595 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3596 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3597 patterns that start (?i) but I don't think that matters. The use of (?i) is
3598 just an example; this all applies to the other options as well.
3599
3600 4. Since libtool seems to echo the compile commands it is issuing, the output
3601 from "make" can be reduced a bit by putting "@" in front of each libtool
3602 compile command.
3603
3604 5. Patch from the folks at Google for configure.in to be a bit more thorough
3605 in checking for a suitable C++ installation before trying to compile the
3606 C++ stuff. This should fix a reported problem when a compiler was present,
3607 but no suitable headers.
3608
3609 6. The man pages all had just "PCRE" as their title. I have changed them to
3610 be the relevant file name. I have also arranged that these names are
3611 retained in the file doc/pcre.txt, which is a concatenation in text format
3612 of all the man pages except the little individual ones for each function.
3613
3614 7. The NON-UNIX-USE file had not been updated for the different set of source
3615 files that come with release 6. I also added a few comments about the C++
3616 wrapper.
3617
3618
3619 Version 6.0 07-Jun-05
3620 ---------------------
3621
3622 1. Some minor internal re-organization to help with my DFA experiments.
3623
3624 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3625 didn't matter for the library itself when fully configured, but did matter
3626 when compiling without UCP support, or within Exim, where the ucp files are
3627 not imported.
3628
3629 3. Refactoring of the library code to split up the various functions into
3630 different source modules. The addition of the new DFA matching code (see
3631 below) to a single monolithic source would have made it really too
3632 unwieldy, quite apart from causing all the code to be include in a
3633 statically linked application, when only some functions are used. This is
3634 relevant even without the DFA addition now that patterns can be compiled in
3635 one application and matched in another.
3636
3637 The downside of splitting up is that there have to be some external
3638 functions and data tables that are used internally in different modules of
3639 the library but which are not part of the API. These have all had their
3640 names changed to start with "_pcre_" so that they are unlikely to clash
3641 with other external names.
3642
3643 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3644 a different (DFA) algorithm. Although it is slower than the original
3645 function, it does have some advantages for certain types of matching
3646 problem.
3647
3648 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3649 including restarting after a partial match.
3650
3651 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3652 defined when compiling for Windows was sent to me. I have put it into the
3653 code, though I have no means of testing or verifying it.
3654
3655 7. Added the pcre_refcount() auxiliary function.
3656
3657 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3658 match before or at the first newline in the subject string. In pcretest,
3659 the /f option on a pattern can be used to set this.
3660
3661 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3662 would behave wrongly. This has been present in PCRE since release 4.0.
3663
3664 10. A number of changes to the pcregrep command:
3665
3666 (a) Refactored how -x works; insert ^(...)$ instead of setting
3667 PCRE_ANCHORED and checking the length, in preparation for adding
3668 something similar for -w.
3669
3670 (b) Added the -w (match as a word) option.
3671
3672 (c) Refactored the way lines are read and buffered so as to have more
3673 than one at a time available.
3674
3675 (d) Implemented a pcregrep test script.
3676
3677 (e) Added the -M (multiline match) option. This allows patterns to match
3678 over several lines of the subject. The buffering ensures that at least
3679 8K, or the rest of the document (whichever is the shorter) is available
3680 for matching (and similarly the previous 8K for lookbehind assertions).
3681
3682 (f) Changed the --help output so that it now says
3683
3684 -w, --word-regex(p)
3685
3686 instead of two lines, one with "regex" and the other with "regexp"
3687 because that confused at least one person since the short forms are the
3688 same. (This required a bit of code, as the output is generated
3689 automatically from a table. It wasn't just a text change.)
3690
3691 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3692 option but starts with a hyphen. Could be a pattern or a path name
3693 starting with a hyphen, for instance.
3694
3695 (h) "-" can be given as a file name to represent stdin.
3696
3697 (i) When file names are being printed, "(standard input)" is used for
3698 the standard input, for compatibility with GNU grep. Previously
3699 "<stdin>" was used.
3700
3701 (j) The option --label=xxx can be used to supply a name to be used for
3702 stdin when file names are being printed. There is no short form.
3703
3704 (k) Re-factored the options decoding logic because we are going to add
3705 two more options that take data. Such options can now be given in four
3706 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3707
3708 (l) Added the -A, -B, and -C options for requesting that lines of context
3709 around matches be printed.
3710
3711 (m) Added the -L option to print the names of files that do not contain
3712 any matching lines, that is, the complement of -l.
3713
3714 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3715 continue to scan other files.
3716
3717 (o) The -s option was incorrectly implemented. For compatibility with other
3718 greps, it now suppresses the error message for a non-existent or non-
3719 accessible file (but not the return code). There is a new option called
3720 -q that suppresses the output of matching lines, which was what -s was
3721 previously doing.
3722
3723 (p) Added --include and --exclude options to specify files for inclusion
3724 and exclusion when recursing.
3725
3726 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3727 Hopefully, it now does.
3728
3729 12. Missing cast in pcre_study().
3730
3731 13. Added an "uninstall" target to the makefile.
3732
3733 14. Replaced "extern" in the function prototypes in Makefile.in with
3734 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3735 world, but is set differently for Windows.
3736
3737 15. Added a second compiling function called pcre_compile2(). The only
3738 difference is that it has an extra argument, which is a pointer to an
3739 integer error code. When there is a compile-time failure, this is set
3740 non-zero, in addition to the error test pointer being set to point to an
3741 error message. The new argument may be NULL if no error number is required
3742 (but then you may as well call pcre_compile(), which is now just a
3743 wrapper). This facility is provided because some applications need a
3744 numeric error indication, but it has also enabled me to tidy up the way
3745 compile-time errors are handled in the POSIX wrapper.
3746
3747 16. Added VPATH=.libs to the makefile; this should help when building with one
3748 prefix path and installing with another. (Or so I'm told by someone who
3749 knows more about this stuff than I do.)
3750
3751 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3752 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3753 match everything, including newlines. This is not POSIX-compatible, but
3754 somebody wanted the feature. From pcretest it can be activated by using
3755 both the P and the s flags.
3756
3757 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3758
3759 19. libpcre.pc was being incorrectly installed as executable.
3760
3761 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3762 it now also looks for '\r' so that it will work unmodified on Windows.
3763
3764 21. Added Google's contributed C++ wrapper to the distribution.
3765
3766 22. Added some untidy missing memory free() calls in pcretest, to keep
3767 Electric Fence happy when testing.
3768
3769
3770
3771 Version 5.0 13-Sep-04
3772 ---------------------
3773
3774 1. Internal change: literal characters are no longer packed up into items
3775 containing multiple characters in a single byte-string. Each character
3776 is now matched using a separate opcode. However, there may be more than one
3777 byte in the character in UTF-8 mode.
3778
3779 2. The pcre_callout_block structure has two new fields: pattern_position and
3780 next_item_length. These contain the offset in the pattern to the next match
3781 item, and its length, respectively.
3782
3783 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3784 insertion of callouts before each pattern item. Added the /C option to
3785 pcretest to make use of this.
3786
3787 4. On the advice of a Windows user, the lines
3788
3789 #if defined(_WIN32) || defined(WIN32)
3790 _setmode( _fileno( stdout ), 0x8000 );
3791 #endif /* defined(_WIN32) || defined(WIN32) */
3792
3793 have been added to the source of pcretest. This apparently does useful
3794 magic in relation to line terminators.
3795
3796 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3797 for the benefit of those environments where the "b" makes a difference.
3798
3799 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3800 to know about it. I have put a hack into configure.in that adds in code
3801 to set GCC=yes if CC=icc. This seems to end up at a point in the
3802 generated configure script that is early enough to affect the setting of
3803 compiler options, which is what is needed, but I have no means of testing
3804 whether it really works. (The user who reported this had patched the
3805 generated configure script, which of course I cannot do.)
3806
3807 LATER: After change 22 below (new libtool files), the configure script
3808 seems to know about icc (and also ecc). Therefore, I have commented out
3809 this hack in configure.in.
3810
3811 7. Added support for pkg-config (2 patches were sent in).
3812
3813 8. Negated POSIX character classes that used a combination of internal tables
3814 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3815 [[:^ascii]]. Typically, they would match almost any characters. The other
3816 POSIX classes were not broken in this way.
3817
3818 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3819 to find the match, as PCRE was deluded into thinking that the match had to
3820 start at the start point or following a newline. The same bug applied to
3821 patterns with negative forward assertions or any backward assertions
3822 preceding ".*" at the start, unless the pattern required a fixed first
3823 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3824
3825 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3826 starting at the last subject character, bytes beyond the end of the subject
3827 string were read.
3828
3829 11. Renamed the variable "class" as "classbits" to make life easier for C++
3830 users. (Previously there was a macro definition, but it apparently wasn't
3831 enough.)
3832
3833 12. Added the new field "tables" to the extra data so that tables can be passed
3834 in at exec time, or the internal tables can be re-selected. This allows
3835 a compiled regex to be saved and re-used at a later time by a different
3836 program that might have everything at different addresses.
3837
3838 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3839 -R library as well as a -L library.
3840
3841 14. The debugging options of pcretest (-d on the command line or D on a
3842 pattern) showed incorrect output for anything following an extended class
3843 that contained multibyte characters and which was followed by a quantifier.
3844
3845 15. Added optional support for general category Unicode character properties
3846 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3847 support. It adds about 90K to the size of the library. The meanings of the
3848 inbuilt class escapes such as \d and \s have NOT been changed.
3849
3850 16. Updated pcredemo.c to include calls to free() to release the memory for the
3851 compiled pattern.
3852
3853 17. The generated file chartables.c was being created in the source directory
3854 instead of in the building directory. This caused the build to fail if the
3855 source directory was different from the building directory, and was
3856 read-only.
3857
3858 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3859 file. No doubt somebody will tell me if they don't make sense... Also added
3860 Dan Mooney's comments about building on OpenVMS.
3861
3862 19. Added support for partial matching via the PCRE_PARTIAL option for
3863 pcre_exec() and the \P data escape in pcretest.
3864
3865 20. Extended pcretest with 3 new pattern features:
3866
3867 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3868 write the compiled pattern to the file whose name is "rest-of-line".
3869 This is a straight binary dump of the data, with the saved pointer to
3870 the character tables forced to be NULL. The study data, if any, is
3871 written too. After writing, pcretest reads a new pattern.
3872
3873 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3874 compiled pattern from the given file. There must not be any
3875 occurrences of "<" in the file name (pretty unlikely); if there are,
3876 pcretest will instead treat the initial "<" as a pattern delimiter.
3877 After reading in the pattern, pcretest goes on to read data lines as
3878 usual.
3879
3880 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3881 and 16-bit fields in a compiled pattern, to simulate a pattern that
3882 was compiled on a host of opposite endianness.
3883
3884 21. The pcre-exec() function can now cope with patterns that were compiled on
3885 hosts of opposite endianness, with this restriction:
3886
3887 As for any compiled expression that is saved and used later, the tables
3888 pointer field cannot be preserved; the extra_data field in the arguments
3889 to pcre_exec() should be used to pass in a tables address if a value
3890 other than the default internal tables were used at compile time.
3891
3892 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3893 now diagnosed as an error. Previously, most of the time, a negative number
3894 would have been treated as zero, but if in addition "ovector" was passed as
3895 NULL, a crash could occur.
3896
3897 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3898 new versions from the libtool 1.5 distribution (the last one is a copy of
3899 a file called libtool.m4). This seems to have fixed the need to patch
3900 "configure" to support Darwin 1.3 (which I used to do). However, I still
3901 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3902 workstation).
3903
3904 24. Changed the PCRE licence to be the more standard "BSD" licence.
3905
3906
3907 Version 4.5 01-Dec-03
3908 ---------------------
3909
3910 1. There has been some re-arrangement of the code for the match() function so
3911 that it can be compiled in a version that does not call itself recursively.
3912 Instead, it keeps those local variables that need separate instances for
3913 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3914 needs to "recurse". Keeping track of where control must go is done by means
3915 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3916 hide most of the details from the main code, and operates only if
3917 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3918 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3919 operating.
3920
3921 To make it easier for callers to provide specially tailored get/free
3922 functions for this usage, two new functions, pcre_stack_malloc, and
3923 pcre_stack_free, are used. They are always called in strict stacking order,
3924 and the size of block requested is always the same.
3925
3926 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3927 PCRE has been compiled to use the stack or the heap for recursion. The
3928 -C option of pcretest uses this to show which version is compiled.
3929
3930 A new data escape \S, is added to pcretest; it causes the amounts of store
3931 obtained and freed by both kinds of malloc/free at match time to be added
3932 to the output.
3933
3934 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3935 what's available on my current Linux desktop machine.
3936
3937 3. When matching a UTF-8 string, the test for a valid string at the start has
3938 been extended. If start_offset is not zero, PCRE now checks that it points
3939 to a byte that is the start of a UTF-8 character. If not, it returns
3940 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3941 this is necessary because there may be backward assertions in the pattern.
3942 When matching the same subject several times, it may save resources to use
3943 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3944
3945 4. The code for checking the validity of UTF-8 strings has been tightened so
3946 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3947 containing "overlong sequences".
3948
3949 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3950 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3951 should have been "&", but it just so happened that all the cases this let
3952 through by mistake were picked up later in the function.
3953
3954 6. I had used a variable called "isblank" - this is a C99 function, causing
3955 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3956
3957 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3958 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3959 so the version is shown; (c) stop "make test" echoing "./RunTest".
3960
3961 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3962
3963 9. The prototype for memmove() for systems that don't have it was using
3964 size_t, but the inclusion of the header that defines size_t was later. I've
3965 moved the #includes for the C headers earlier to avoid this.
3966
3967 10. Added some adjustments to the code to make it easier to compiler on certain
3968 special systems:
3969
3970 (a) Some "const" qualifiers were missing.
3971 (b) Added the macro EXPORT before all exported functions; by default this
3972 is defined to be empty.
3973 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3974 that it reads its output file name as an argument instead of writing
3975 to the standard output and assuming this can be redirected.
3976
3977 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3978 class containing characters with values greater than 255, PCRE compilation
3979 went into a loop.
3980
3981 12. A recursive reference to a subpattern that was within another subpattern
3982 that had a minimum quantifier of zero caused PCRE to crash. For example,
3983 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3984 recursion. If the recursively-called subpattern itself had a zero repeat,
3985 that was OK.
3986
3987 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3988 buffer into which it was copied (for escape processing) was still set at
3989 1024, so long lines caused crashes.
3990
3991 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3992 "internal error: code overflow...". This applied to any character class
3993 that was followed by a possessive quantifier.
3994
3995 15. Modified the Makefile to add libpcre.la as a prerequisite for
3996 libpcreposix.la because I was told this is needed for a parallel build to
3997 work.
3998
3999 16. If a pattern that contained .* following optional items at the start was
4000 studied, the wrong optimizing data was generated, leading to matching
4001 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4002 matching string must start with a or b or c. The correct conclusion for
4003 this pattern is that a match can start with any character.
4004
4005
4006 Version 4.4 13-Aug-03
4007 ---------------------
4008
4009 1. In UTF-8 mode, a character class containing characters with values between
4010 127 and 255 was not handled correctly if the compiled pattern was studied.
4011 In fixing this, I have also improved the studying algorithm for such
4012 classes (slightly).
4013
4014 2. Three internal functions had redundant arguments passed to them. Removal
4015 might give a very teeny performance improvement.
4016
4017 3. Documentation bug: the value of the capture_top field in a callout is *one
4018 more than* the number of the hightest numbered captured substring.
4019
4020 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4021 in incorrectly linking with a previously installed version. They now link
4022 explicitly with libpcre.la.
4023
4024 5. configure.in no longer needs to recognize Cygwin specially.
4025
4026 6. A problem in pcre.in for Windows platforms is fixed.
4027
4028 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4029 pcretest, it used to include the size of the study block as part of its
4030 output. Unfortunately, the structure contains a field that has a different
4031 size on different hardware architectures. This meant that the tests that
4032 showed this size failed. As the block is currently always of a fixed size,
4033 this information isn't actually particularly useful in pcretest output, so
4034 I have just removed it.
4035
4036 8. Three pre-processor statements accidentally did not start in column 1.
4037 Sadly, there are *still* compilers around that complain, even though
4038 standard C has not required this for well over a decade. Sigh.
4039
4040 9. In pcretest, the code for checking callouts passed small integers in the
4041 callout_data field, which is a void * field. However, some picky compilers
4042 complained about the casts involved for this on 64-bit systems. Now
4043 pcretest passes the address of the small integer instead, which should get
4044 rid of the warnings.
4045
4046 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4047 both compile and run time, and gives an error if an invalid UTF-8 sequence
4048 is found. There is a option for disabling this check in cases where the
4049 string is known to be correct and/or the maximum performance is wanted.
4050
4051 11. In response to a bug report, I changed one line in Makefile.in from
4052
4053 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4054 to
4055 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4056
4057 to look similar to other lines, but I have no way of telling whether this
4058 is the right thing to do, as I do not use Windows. No doubt I'll get told
4059 if it's wrong...
4060
4061
4062 Version 4.3 21-May-03
4063 ---------------------
4064
4065 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4066 Makefile.
4067
4068 2. Some refactoring to improve the quality of the code:
4069
4070 (i) The utf8_table... variables are now declared "const".
4071
4072 (ii) The code for \cx, which used the "case flipping" table to upper case
4073 lower case letters, now just substracts 32. This is ASCII-specific,
4074 but the whole concept of \cx is ASCII-specific, so it seems
4075 reasonable.
4076
4077 (iii) PCRE was using its character types table to recognize decimal and
4078 hexadecimal digits in the pattern. This is silly, because it handles
4079 only 0-9, a-f, and A-F, but the character types table is locale-
4080 specific, which means strange things might happen. A private
4081 table is now used for this - though it costs 256 bytes, a table is
4082 much faster than multiple explicit tests. Of course, the standard
4083 character types table is still used for matching digits in subject
4084 strings against \d.
4085
4086 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4087 ending in _t are). So I've renamed it as ESC_tee.
4088
4089 3. The first argument for regexec() in the POSIX wrapper should have been
4090 defined as "const".
4091
4092 4. Changed pcretest to use malloc() for its buffers so that they can be
4093 Electric Fenced for debugging.
4094
4095 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4096 to read one or more bytes before the start of the subject string. Often this
4097 had no effect on PCRE's behaviour, but in some circumstances it could
4098 provoke a segmentation fault.
4099
4100 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4101 to try to read one or more bytes before the start of the subject string.
4102
4103 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4104 UTF-8 support could misbehave in various ways if the subject string
4105 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4106 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4107 back over UTF-8 characters.)
4108
4109
4110 Version 4.2 14-Apr-03
4111 ---------------------
4112
4113 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4114
4115 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4116 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4117 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4118 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4119 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4120 and BUILD_EXEEXT
4121 Note: automatic setting of the BUILD variables is not yet working
4122 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4123 compile-time but not at link-time
4124 [LINK]: use for linking executables only
4125 make different versions for Windows and non-Windows
4126 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4127 libraries
4128 [LINK_FOR_BUILD]: new variable
4129 [OBJEXT]: use throughout
4130 [EXEEXT]: use throughout
4131 <winshared>: new target
4132 <wininstall>: new target
4133 <dftables.o>: use native compiler
4134 <dftables>: use native linker
4135 <install>: handle Windows platform correctly
4136 <clean>: ditto
4137 <check>: ditto
4138 copy DLL to top builddir before testing
4139
4140 As part of these changes, -no-undefined was removed again. This was reported
4141 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4142 in any case.
4143
4144 3. Some tidies to get rid of compiler warnings:
4145
4146 . In the match_data structure, match_limit was an unsigned long int, whereas
4147 match_call_count was an int. I've made them both unsigned long ints.
4148
4149 . In pcretest the fact that a const uschar * doesn't automatically cast to
4150 a void * provoked a warning.
4151
4152 . Turning on some more compiler warnings threw up some "shadow" variables
4153 and a few more missing casts.
4154
4155 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4156 option, a class that contained a single character with a value between 128
4157 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4158
4159 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4160 option, a class that contained several characters, but with at least one
4161 whose value was between 128 and 255 caused PCRE to crash.
4162
4163
4164 Version 4.1 12-Mar-03
4165 ---------------------
4166
4167 1. Compiling with gcc -pedantic found a couple of places where casts were
4168 needed, and a string in dftables.c that was longer than standard compilers are
4169 required to support.
4170
4171 2. Compiling with Sun's compiler found a few more places where the code could
4172 be tidied up in order to avoid warnings.
4173
4174 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4175 first of these names is deprecated in the latest Autoconf in favour of the name
4176 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4177 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4178 analogy I have changed it to CFLAGS_FOR_BUILD.
4179
4180 4. Added -no-undefined to the linking command in the Makefile, because this is
4181 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4182 linking step for the pcreposix library.
4183
4184 5. PCRE was failing to diagnose the case of two named groups with the same
4185 name.
4186
4187 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4188 literal character that is needed in the subject for a match, and scans along to
4189 ensure that it is present before embarking on the full matching process. This
4190 saves time in cases of nested unlimited repeats that are never going to match.
4191 Problem: the scan can take a lot of time if the subject is very long (e.g.
4192 megabytes), thus penalizing straightforward matches. It is now done only if the
4193 amount of subject to be scanned is less than 1000 bytes.
4194
4195 7. A lesser problem with the same optimization is that it was recording the
4196 first character of an anchored pattern as "needed", thus provoking a search
4197 right along the subject, even when the first match of the pattern was going to
4198 fail. The "needed" character is now not set for anchored patterns, unless it
4199 follows something in the pattern that is of non-fixed length. Thus, it still
4200 fulfils its original purpose of finding quick non-matches in cases of nested
4201 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4202
4203
4204 Version 4.0 17-Feb-03
4205 ---------------------
4206
4207 1. If a comment in an extended regex that started immediately after a meta-item
4208 extended to the end of string, PCRE compiled incorrect data. This could lead to
4209 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4210
4211 2. Moved to autoconf 2.53 and libtool 1.4.2.
4212
4213 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4214 the special perltest8 script is no longer needed - all the tests can be run
4215 from a single perltest script.
4216
4217 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4218 by \s. It has now been removed in PCRE. This means it isn't recognized as
4219 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4220 class [:space:] *does* include VT, thereby creating a mess.
4221
4222 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4223 space and tab.
4224
4225 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4226 its new features into the main test script, reducing the number of scripts.
4227
4228 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4229 were backward compatible, and made the (?i) apply to the whole pattern, as if
4230 /i were given. Now it behaves more logically, and applies the option setting
4231 only to what follows. PCRE has been changed to follow suit. However, if it
4232 finds options settings right at the start of the pattern, it extracts them into
4233 the global options, as before. Thus, they show up in the info data.
4234
4235 8. Added support for the \Q...\E escape sequence. Characters in between are
4236 treated as literals. This is slightly different from Perl in that $ and @ are
4237 also handled as literals inside the quotes. In Perl, they will cause variable
4238 interpolation. Note the following examples:
4239
4240 Pattern PCRE matches Perl matches
4241
4242 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4243 \Qabc\$xyz\E abc\$xyz abc\$xyz
4244 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4245
4246 For compatibility with Perl, \Q...\E sequences are recognized inside character
4247 classes as well as outside them.
4248
4249 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4250 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4251 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4252 signed/unsigned warnings.
4253
4254 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4255 option for pcretest, so I've replaced it by a simple function that does just
4256 that job.
4257
4258 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4259 "pcregrep -".
4260
4261 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4262 Java package. This provides some syntactic sugar for simple cases of what my
4263 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4264 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4265 item, you can use this simplified notation. Note that only makes sense with
4266 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4267 greediness, whatever the setting of the PCRE_UNGREEDY option.
4268
4269 13. A change of greediness default within a pattern was not taking effect at
4270 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4271 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4272 was abstracted outside.
4273
4274 14. PCRE now supports the \G assertion. It is true when the current matching
4275 position is at the start point of the match. This differs from \A when the
4276 starting offset is non-zero. Used with the /g option of pcretest (or similar
4277 code), it works in the same way as it does for Perl's /g option. If all
4278 alternatives of a regex begin with \G, the expression is anchored to the start
4279 match position, and the "anchored" flag is set in the compiled expression.
4280
4281 15. Some bugs concerning the handling of certain option changes within patterns
4282 have been fixed. These applied to options other than (?ims). For example,
4283 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4284 been the other way round. Some of this was related to change 7 above.
4285
4286 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4287 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4288 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4289 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4290
4291 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4292 mode. Unlike ".", it always matches newline, whatever the setting of
4293 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4294 assertions. Perl allows it, but it doesn't (in general) work because it can't
4295 calculate the length of the lookbehind. At least, that's the case for Perl
4296 5.8.0 - I've been told they are going to document that it doesn't work in
4297 future.
4298
4299 18. Added an error diagnosis for escapes that PCRE does not support: these are
4300 \L, \l, \N, \P, \p, \U, \u, and \X.
4301
4302 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4303 reading past the end of the pattern in cases such as /[abcd/.
4304
4305 20. PCRE was getting more memory than necessary for patterns with classes that
4306 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4307
4308 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4309 compiling PCRE for use with Virtual Pascal.
4310
4311 22. Small fix to the Makefile to make it work properly if the build is done
4312 outside the source tree.
4313
4314 23. Added a new extension: a condition to go with recursion. If a conditional
4315 subpattern starts with (?(R) the "true" branch is used if recursion has
4316 happened, whereas the "false" branch is used only at the top level.
4317
4318 24. When there was a very long string of literal characters (over 255 bytes
4319 without UTF support, over 250 bytes with UTF support), the computation of how
4320 much memory was required could be incorrect, leading to segfaults or other
4321 strange effects.
4322
4323 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4324 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4325 there was a subsequent back reference to those brackets. This meant that, for
4326 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4327 possible to check for precisely this case. All we can do is abandon the
4328 optimization if .* occurs inside capturing brackets when there are any back
4329 references whatsoever. (See below for a better fix that came later.)
4330
4331 26. The handling of the optimization for finding the first character of a
4332 non-anchored pattern, and for finding a character that is required later in the
4333 match were failing in some cases. This didn't break the matching; it just
4334 failed to optimize when it could. The way this is done has been re-implemented.
4335
4336 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4337
4338 28. Added a new feature that provides some of the functionality that Perl
4339 provides with (?{...}). The facility is termed a "callout". The way it is done
4340 in PCRE is for the caller to provide an optional function, by setting
4341 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4342 global variable. By default it is unset, which disables all calling out. To get
4343 the function called, the regex must include (?C) at appropriate points. This
4344 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4345 This provides a means of identifying different callout points. When PCRE
4346 reaches such a point in the regex, if pcre_callout has been set, the external
4347 function is called. It is provided with data in a structure called
4348 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4349 matching continues; if it returns a non-zero value, the match at the current
4350 point fails. However, backtracking will occur if possible. [This was changed
4351 later and other features added - see item 49 below.]
4352
4353 29. pcretest is upgraded to test the callout functionality. It provides a
4354 callout function that displays information. By default, it shows the start of
4355 the match and the current position in the text. There are some new data escapes
4356 to vary what happens:
4357
4358 \C+ in addition, show current contents of captured substrings
4359 \C- do not supply a callout function
4360 \C!n return 1 when callout number n is reached
4361 \C!n!m return 1 when callout number n is reached for the mth time
4362
4363 30. If pcregrep was called with the -l option and just a single file name, it
4364 output "<stdin>" if a match was found, instead of the file name.
4365
4366 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4367 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4368 pcre_exec(). This saves a malloc/free per call. The default value of
4369 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4370 when configuring.
4371
4372 32. The default maximum size of a compiled pattern is 64K. There have been a
4373 few cases of people hitting this limit. The code now uses macros to handle the
4374 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4375 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4376 configuring. Tests 2 and 5 work only with 2-byte links because they output
4377 debugging information about compiled patterns.
4378
4379 33. Internal code re-arrangements:
4380
4381 (a) Moved the debugging function for printing out a compiled regex into
4382 its own source file (printint.c) and used #include to pull it into
4383 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4384 separate copies.
4385
4386 (b) Defined the list of op-code names for debugging as a macro in
4387 internal.h so that it is next to the definition of the opcodes.
4388
4389 (c) Defined a table of op-code lengths for simpler skipping along compiled
4390 code. This is again a macro in internal.h so that it is next to the
4391 definition of the opcodes.
4392
4393 34. Added support for recursive calls to individual subpatterns, along the
4394 lines of Robin Houston's patch (but implemented somewhat differently).
4395
4396 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4397 allow it to read and process whole directories in Win32. This code was
4398 contributed by Lionel Fourquaux; it has not been tested by me.
4399
4400 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4401 used to name a group. Names consist of alphanumerics and underscores, and must
4402 be unique. Back references use the syntax (?P=name) and recursive calls use
4403 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4404 numbers. The function pcre_fullinfo() can be used after compilation to extract
4405 a name/number map. There are three relevant calls:
4406
4407 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4408 PCRE_INFO_NAMECOUNT yields the number of entries
4409 PCRE_INFO_NAMETABLE yields a pointer to the map.
4410
4411 The map is a vector of fixed-size entries. The size of each entry depends on
4412 the length of the longest name used. The first two bytes of each entry are the
4413 group number, most significant byte first. There follows the corresponding
4414 name, zero terminated. The names are in alphabetical order.
4415
4416 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4417 case instead of 255. Making it the same both with and without UTF-8 support
4418 means that the same test output works with both.
4419
4420 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4421 calling malloc() with a zero argument.
4422
4423 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4424 optimization. I've improved things by keeping a bitmap of backreferences with
4425 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4426 fact referenced, the optimization can be applied. It is unlikely that a
4427 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4428 the match to follow \n) will appear inside brackets with a number greater than
4429 31, but if it does, any back reference > 31 suppresses the optimization.
4430
4431 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4432 of disabling numbered capturing parentheses. Any opening parenthesis that is
4433 not followed by ? behaves as if it were followed by ?: but named parentheses
4434 can still be used for capturing (and they will acquire numbers in the usual
4435 way).
4436
4437 41. Redesigned the return codes from the match() function into yes/no/error so
4438 that errors can be passed back from deep inside the nested calls. A malloc
4439 failure while inside a recursive subpattern call now causes the
4440 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4441
4442 42. It is now possible to set a limit on the number of times the match()
4443 function is called in a call to pcre_exec(). This facility makes it possible to
4444 limit the amount of recursion and backtracking, though not in a directly
4445 obvious way, because the match() function is used in a number of different
4446 circumstances. The count starts from zero for each position in the subject
4447 string (for non-anchored patterns). The default limit is, for compatibility, a
4448 large number, namely 10 000 000. You can change this in two ways:
4449
4450 (a) When configuring PCRE before making, you can use --with-match-limit=n
4451 to set a default value for the compiled library.
4452
4453 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4454 a different value is set. See 45 below.
4455
4456 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4457
4458 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4459 of things that can be changed at compile time. The first argument specifies
4460 what is wanted and the second points to where the information is to be placed.
4461 The current list of available information is:
4462
4463 PCRE_CONFIG_UTF8
4464
4465 The output is an integer that is set to one if UTF-8 support is available;
4466 otherwise it is set to zero.
4467
4468 PCRE_CONFIG_NEWLINE
4469
4470 The output is an integer that it set to the value of the code that is used for
4471 newline. It is either LF (10) or CR (13).
4472
4473 PCRE_CONFIG_LINK_SIZE
4474
4475 The output is an integer that contains the number of bytes used for internal
4476 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4477
4478 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4479
4480 The output is an integer that contains the threshold above which the POSIX
4481 interface uses malloc() for output vectors. See item 31 above.
4482
4483 PCRE_CONFIG_MATCH_LIMIT
4484
4485 The output is an unsigned integer that contains the default limit of the number
4486 of match() calls in a pcre_exec() execution. See 42 above.
4487
4488 44. pcretest has been upgraded by the addition of the -C option. This causes it
4489 to extract all the available output from the new pcre_config() function, and to
4490 output it. The program then exits immediately.
4491
4492 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4493 order to support additional features. One way would have been to define
4494 pcre_exec2() (for example) with extra arguments, but this would not have been
4495 extensible, and would also have required all calls to the original function to
4496 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4497 is used for passing in "extra" data from pcre_study().
4498
4499 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4500 contains the following fields:
4501
4502 flags a bitmap indicating which of the following fields are set
4503 study_data opaque data from pcre_study()
4504 match_limit a way of specifying a limit on match() calls for a specific
4505 call to pcre_exec()
4506 callout_data data for callouts (see 49 below)
4507
4508 The flag bits are also defined in pcre.h, and are
4509
4510 PCRE_EXTRA_STUDY_DATA
4511 PCRE_EXTRA_MATCH_LIMIT
4512 PCRE_EXTRA_CALLOUT_DATA
4513
4514 The pcre_study() function now returns one of these new pcre_extra blocks, with
4515 the actual study data pointed to by the study_data field, and the
4516 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4517 before. That is, this change is entirely upwards-compatible and requires no
4518 change to existing code.
4519
4520 If you want to pass in additional data to pcre_exec(), you can either place it
4521 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4522 block.
4523
4524 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4525 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4526 times with different match limits, until it finds the minimum value needed for
4527 pcre_exec() to complete. The value is then output. This can be instructive; for
4528 most simple matches the number is quite small, but for pathological cases it
4529 gets very large very quickly.
4530
4531 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4532 returns the size of the data block pointed to by the study_data field in a
4533 pcre_extra block, that is, the value that was passed as the argument to
4534 pcre_malloc() when PCRE was getting memory in which to place the information
4535 created by pcre_study(). The fourth argument should point to a size_t variable.
4536 pcretest has been extended so that this information is shown after a successful
4537 pcre_study() call when information about the compiled regex is being displayed.
4538
4539 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4540 because what follows is always an absolute path. (Later: it turns out that this
4541 is more than cosmetic for MinGW, because it doesn't like empty path
4542 components.)
4543
4544 49. Some changes have been made to the callout feature (see 28 above):
4545
4546 (i) A callout function now has three choices for what it returns:
4547
4548 0 => success, carry on matching
4549 > 0 => failure at this point, but backtrack if possible
4550 < 0 => serious error, return this value from pcre_exec()
4551
4552 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4553 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4554 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4555 use by callout functions. It will never be used by PCRE itself.
4556
4557 (ii) The pcre_extra structure (see 45 above) has a void * field called
4558 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4559 pcre_callout_block structure has a field of the same name. The contents of
4560 the field passed in the pcre_extra structure are passed to the callout
4561 function in the corresponding field in the callout block. This makes it
4562 easier to use the same callout-containing regex from multiple threads. For
4563 testing, the pcretest program has a new data escape
4564
4565 \C*n pass the number n (may be negative) as callout_data
4566
4567 If the callout function in pcretest receives a non-zero value as
4568 callout_data, it returns that value.
4569
4570 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4571 there were some redundant $(CFLAGS) in commands that are now specified as
4572 $(LINK), which already includes $(CFLAGS).
4573
4574 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4575 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4576 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4577 one-byte characters throughout. Note that case-insensitive matching applies
4578 only to characters whose values are less than 256. PCRE doesn't support the
4579 notion of cases for higher-valued characters.
4580
4581 (i) A character class whose characters are all within 0-255 is handled as
4582 a bit map, and the map is inverted for negative classes. Previously, a
4583 character > 255 always failed to match such a class; however it should
4584 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4585
4586 (ii) A negated character class with a single character < 255 is coded as
4587 "not this character" (OP_NOT). This wasn't working properly when the test
4588 character was multibyte, either singly or repeated.
4589
4590 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4591 mode, for example: \x{100}{2,3}.
4592
4593 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4594 singly or repeated) now correctly test multibyte characters. However,
4595 PCRE doesn't recognize any characters with values greater than 255 as
4596 digits, spaces, or word characters. Such characters always match \D, \S,
4597 and \W, and never match \d, \s, or \w.
4598
4599 (v) Classes may now contain characters and character ranges with values
4600 greater than 255. For example: [ab\x{100}-\x{400}].
4601
4602 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4603 PCRE in UTF-8 mode.
4604
4605 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4606 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4607 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4608 value.)
4609
4610 53. The single man page has become too large. I have therefore split it up into
4611 a number of separate man pages. These also give rise to individual HTML pages;
4612 these are now put in a separate directory, and there is an index.html page that
4613 lists them all. Some hyperlinking between the pages has been installed.
4614
4615 54. Added convenience functions for handling named capturing parentheses.
4616
4617 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4618 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4619 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4620 are faulted.
4621
4622 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4623 calling configure. These values are used when compiling the dftables.c program
4624 which is run to generate the source of the default character tables. They
4625 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4626 you will need to set these values.
4627
4628 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4629
4630
4631 Version 3.9 02-Jan-02
4632 ---------------------
4633
4634 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4635
4636 2. If --disable-static was given, the building process failed when trying to
4637 build pcretest and pcregrep. (For some reason it was using libtool to compile
4638 them, which is not right, as they aren't part of the library.)
4639
4640
4641 Version 3.8 18-Dec-01
4642 ---------------------
4643
4644 1. The experimental UTF-8 code was completely screwed up. It was packing the
4645 bytes in the wrong order. How dumb can you get?
4646
4647
4648 Version 3.7 29-Oct-01
4649 ---------------------
4650
4651 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4652 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4653 this didn't happen under Solaris 8, where I normally test things.
4654
4655 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4656 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4657 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4658 this fix an hour or so after the initial 3.7 release.)
4659
4660
4661 Version 3.6 23-Oct-01
4662 ---------------------
4663
4664 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4665 offsets passed as NULL with zero offset count.
4666
4667 2. The config.guess and config.sub files had not been updated when I moved to
4668 the latest autoconf.
4669
4670
4671 Version 3.5 15-Aug-01
4672 ---------------------
4673
4674 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4675 had been forgotten.
4676
4677 2. By using declared but undefined structures, we can avoid using "void"
4678 definitions in pcre.h while keeping the internal definitions of the structures
4679 private.
4680
4681 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4682 user point of view, this means that both static and shared libraries are built
4683 by default, but this can be individually controlled. More of the work of
4684 handling this static/shared cases is now inside libtool instead of PCRE's make
4685 file.
4686
4687 4. The pcretest utility is now installed along with pcregrep because it is
4688 useful for users (to test regexs) and by doing this, it automatically gets
4689 relinked by libtool. The documentation has been turned into a man page, so
4690 there are now .1, .txt, and .html versions in /doc.
4691
4692 5. Upgrades to pcregrep:
4693 (i) Added long-form option names like gnu grep.
4694 (ii) Added --help to list all options with an explanatory phrase.
4695 (iii) Added -r, --recursive to recurse into sub-directories.
4696 (iv) Added -f, --file to read patterns from a file.
4697
4698 6. pcre_exec() was referring to its "code" argument before testing that
4699 argument for NULL (and giving an error if it was NULL).
4700
4701 7. Upgraded Makefile.in to allow for compiling in a different directory from
4702 the source directory.
4703
4704 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4705 options bits, the pointer it was passed was to an int instead of to an unsigned
4706 long int. This mattered only on 64-bit systems.
4707
4708 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4709 generated) instead of pcre.in, which it its source. Also made the same change
4710 in several of the .c files.
4711
4712 10. A new release of gcc defines printf() as a macro, which broke pcretest
4713 because it had an ifdef in the middle of a string argument for printf(). Fixed
4714 by using separate calls to printf().
4715
4716 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4717 script, to force use of CR or LF instead of \n in the source. On non-Unix
4718 systems, the value can be set in config.h.
4719
4720 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4721 absolute limit. Changed the text of the error message to make this clear, and
4722 likewise updated the man page.
4723
4724 13. The limit of 99 on the number of capturing subpatterns has been removed.
4725 The new limit is 65535, which I hope will not be a "real" limit.
4726
4727
4728 Version 3.4 22-Aug-00
4729 ---------------------
4730
4731 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4732
4733 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4734
4735
4736 Version 3.3 01-Aug-00
4737 ---------------------
4738
4739 1. If an octal character was given, but the value was greater than \377, it
4740 was not getting masked to the least significant bits, as documented. This could
4741 lead to crashes in some systems.
4742
4743 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4744 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4745
4746 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4747 These just pass their arguments on to (pcre_free)(), but they are provided
4748 because some uses of PCRE bind it to non-C systems that can call its functions,
4749 but cannot call free() or pcre_free() directly.
4750
4751 4. Add "make test" as a synonym for "make check". Corrected some comments in
4752 the Makefile.
4753
4754 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4755 Makefile.
4756
4757 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4758 command called pgrep for grepping around the active processes.
4759
4760 7. Added the beginnings of support for UTF-8 character strings.
4761
4762 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4763 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4764 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4765 out for the ar command.)
4766
4767
4768 Version 3.2 12-May-00
4769 ---------------------
4770
4771 This is purely a bug fixing release.
4772
4773 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4774 of ZA. This was just one example of several cases that could provoke this bug,
4775 which was introduced by change 9 of version 2.00. The code for breaking
4776 infinite loops after an iteration that matches an empty string was't working
4777 correctly.
4778
4779 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4780 when matched against abbab (for example). After matching an empty string, it
4781 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4782 caused it to match further down the string than it should.
4783
4784 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4785 was there because it doesn't seem to be needed, and it causes trouble on some
4786 systems, as it is not a Standard C header. It has been removed.
4787
4788 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4789 were reported on the Macintosh. The changes were from
4790
4791 while ((c = *(++ptr)) != 0 && c != '\n');
4792 to
4793 while ((c = *(++ptr)) != 0 && c != '\n') ;
4794
4795 Totally extraordinary, but if that's what it takes...
4796
4797 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4798 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4799 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4800 assumes the way PCRE uses memmove() (always moving upwards).
4801
4802 6. PCRE is being used in one environment where strchr() is not available. There
4803 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4804 faster code anyway.
4805
4806
4807 Version 3.1 09-Feb-00
4808 ---------------------
4809
4810 The only change in this release is the fixing of some bugs in Makefile.in for
4811 the "install" target:
4812
4813 (1) It was failing to install pcreposix.h.
4814
4815 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4816
4817
4818 Version 3.0 01-Feb-00
4819 ---------------------
4820
4821 1. Add support for the /+ modifier to perltest (to output $` like it does in
4822 pcretest).
4823
4824 2. Add support for the /g modifier to perltest.
4825
4826 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4827 matches null strings.
4828
4829 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4830 pattern. Perl treats empty patterns specially - it reuses the most recent
4831 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4832 effect.
4833
4834 5. The POSIX interface was broken in that it was just handing over the POSIX
4835 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4836 required a bigger vector, with some working space on the end. This means that
4837 the POSIX wrapper now has to get and free some memory, and copy the results.
4838
4839 6. Added some simple autoconf support, placing the test data and the
4840 documentation in separate directories, re-organizing some of the
4841 information files, and making it build pcre-config (a GNU standard). Also added
4842 libtool support for building PCRE as a shared library, which is now the
4843 default.
4844
4845 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4846 09 are not valid octal constants. Single digits will be used for minor values
4847 less than 10.
4848
4849 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4850 existing programs that set these in the POSIX interface can use PCRE without
4851 modification.
4852
4853 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4854 return all that pcre_info() returns, plus additional data. The pcre_info()
4855 function is retained for compatibility, but is considered to be obsolete.
4856
4857 10. Added experimental recursion feature (?R) to handle one common case that
4858 Perl 5.6 will be able to do with (?p{...}).
4859
4860 11. Added support for POSIX character classes like [:alpha:], which Perl is
4861 adopting.
4862
4863
4864 Version 2.08 31-Aug-99
4865 ----------------------
4866
4867 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4868 trying to match at the startoffset position, but instead was moving forward to
4869 the next newline as if a previous match had failed.
4870
4871 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4872 and could get into a loop if a null string was matched other than at the start
4873 of the subject.
4874
4875 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4876 be distinguished at compile time, and for completeness also added PCRE_DATE.
4877
4878 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4879 in GnuWin32 environments.
4880
4881
4882 Version 2.07 29-Jul-99
4883 ----------------------
4884
4885 1. The documentation is now supplied in plain text form and HTML as well as in
4886 the form of man page sources.
4887
4888 2. C++ compilers don't like assigning (void *) values to other pointer types.
4889 In particular this affects malloc(). Although there is no problem in Standard
4890 C, I've put in casts to keep C++ compilers happy.
4891
4892 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4893 should be (const char *).
4894
4895 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4896 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4897 However, I haven't made this a standard facility. The documentation doesn't
4898 mention it, and the Makefile doesn't support it.
4899
4900 5. The Makefile now contains an "install" target, with editable destinations at
4901 the top of the file. The pcretest program is not installed.
4902
4903 6. pgrep -V now gives the PCRE version number and date.
4904
4905 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4906 causing the entire string to be ignored, instead of just the last character.
4907
4908 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4909 non-matching string, it can take a very, very long time, even for strings of
4910 quite modest length, because of the nested recursion. PCRE now does better in
4911 some of these cases. It does this by remembering the last required literal
4912 character in the pattern, and pre-searching the subject to ensure it is present
4913 before running the real match. In other words, it applies a heuristic to detect
4914 some types of certain failure quickly, and in the above example, if presented
4915 with a string that has no trailing " it gives "no match" very quickly.
4916
4917 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4918 other alternatives are tried instead.
4919
4920
4921 Version 2.06 09-Jun-99
4922 ----------------------
4923
4924 1. Change pcretest's output for amount of store used to show just the code
4925 space, because the remainder (the data block) varies in size between 32-bit and
4926 64-bit systems.
4927
4928 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4929 start matching at. This allows lookbehinds to work when searching for multiple
4930 occurrences in a string.
4931
4932 3. Added additional options to pcretest for testing multiple occurrences:
4933
4934 /+ outputs the rest of the string that follows a match
4935 /g loops for multiple occurrences, using the new startoffset argument
4936 /G loops for multiple occurrences by passing an incremented pointer
4937
4938 4. PCRE wasn't doing the "first character" optimization for patterns starting
4939 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4940 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4941 the letter 'x'. On long subject strings, this gives a significant speed-up.
4942
4943
4944 Version 2.05 21-Apr-99
4945 ----------------------
4946
4947 1. Changed the type of magic_number from int to long int so that it works
4948 properly on 16-bit systems.
4949
4950 2. Fixed a bug which caused patterns starting with .* not to work correctly
4951 when the subject string contained newline characters. PCRE was assuming
4952 anchoring for such patterns in all cases, which is not correct because .* will
4953 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4954 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4955 must be retried after every newline in the subject.
4956
4957
4958 Version 2.04 18-Feb-99
4959 ----------------------
4960
4961 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4962 computation of the store needed to hold the pattern was incorrect (too large).
4963 If such patterns were nested a few deep, this could multiply and become a real
4964 problem.
4965
4966 2. Added /M option to pcretest to show the memory requirement of a specific
4967 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4968
4969 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4970 compiled in such a way that the backtracking after subsequent failure was
4971 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4972 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4973
4974
4975 Version 2.03 02-Feb-99
4976 ----------------------
4977
4978 1. Fixed typo and small mistake in man page.
4979
4980 2. Added 4th condition (GPL supersedes if conflict) and created separate
4981 LICENCE file containing the conditions.
4982
4983 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4984 Perl, that is the internal \ allows the delimiter to be included in the
4985 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4986 the final delimiter, add \ to the end of the pattern (to test the error).
4987
4988 4. Added the convenience functions for extracting substrings after a successful
4989 match. Updated pcretest to make it able to test these functions.
4990
4991
4992 Version 2.02 14-Jan-99
4993 ----------------------
4994
4995 1. Initialized the working variables associated with each extraction so that
4996 their saving and restoring doesn't refer to uninitialized store.
4997
4998 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4999 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5000 fix the problem.
5001
5002 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5003 calls, and wasn't printing the correct value for compiling calls. Increased the
5004 default value of LOOPREPEAT, and the number of significant figures in the
5005 times.
5006
5007 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5008
5009 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5010 a building problem on Windows NT with a FAT file system.
5011
5012
5013 Version 2.01 21-Oct-98
5014 ----------------------
5015
5016 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5017 to character tables built by pcre_maketables() in the current locale. If NULL
5018 is passed, the default tables are used.
5019
5020
5021 Version 2.00 24-Sep-98
5022 ----------------------
5023
5024 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5025 it any more.
5026
5027 2. Allow quantification of (?>) groups, and make it work correctly.
5028
5029 3. The first character computation wasn't working for (?>) groups.
5030
5031 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5032 end of the subject) and add 5.005's \z, which really does match only at the
5033 very end of the subject.
5034
5035 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5036
5037 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5038 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5039 localized options. All options to pcre_study() were also removed.
5040
5041 7. Add other new features from 5.005:
5042
5043 $(?<= positive lookbehind
5044 $(?<! negative lookbehind
5045 (?imsx-imsx) added the unsetting capability
5046 such a setting is global if at outer level; local otherwise
5047 (?imsx-imsx:) non-capturing groups with option setting
5048 (?(cond)re|re) conditional pattern matching
5049
5050 A backreference to itself in a repeated group matches the previous
5051 captured string.
5052
5053 8. General tidying up of studying (both automatic and via "study")
5054 consequential on the addition of new assertions.
5055
5056 9. As in 5.005, unlimited repeated groups that could match an empty substring
5057 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5058 runtime if any iteration does actually match an empty substring.
5059
5060 10. Include the RunTest script in the distribution.
5061
5062 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5063 discrepancies, some of which were old and were also with respect to 5.004. They
5064 have now been fixed.
5065
5066
5067 Version 1.09 28-Apr-98
5068 ----------------------
5069
5070 1. A negated single character class followed by a quantifier with a minimum
5071 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5072 program crashes, or just wrong answers. This did not apply to negated classes
5073 containing more than one character, or to minima other than one.
5074
5075
5076 Version 1.08 27-Mar-98
5077 ----------------------
5078
5079 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5080
5081 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5082 latter must appear before anything that relies on it in the pattern.
5083
5084
5085 Version 1.07 16-Feb-98
5086 ----------------------
5087
5088 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5089 repeat of a potentially empty string).
5090
5091
5092 Version 1.06 23-Jan-98
5093 ----------------------
5094
5095 1. Added Markus Oberhumer's little patches for C++.
5096
5097 2. Literal strings longer than 255 characters were broken.
5098
5099
5100 Version 1.05 23-Dec-97
5101 ----------------------
5102
5103 1. Negated character classes containing more than one character were failing if
5104 PCRE_CASELESS was set at run time.
5105
5106
5107 Version 1.04 19-Dec-97
5108 ----------------------
5109
5110 1. Corrected the man page, where some "const" qualifiers had been omitted.
5111
5112 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5113 input syntax.
5114
5115 3. Fixed memory leak which occurred when a regex with back references was
5116 matched with an offsets vector that wasn't big enough. The temporary memory
5117 that is used in this case wasn't being freed if the match failed.
5118
5119 4. Tidied pcretest to ensure it frees memory that it gets.
5120
5121 5. Temporary memory was being obtained in the case where the passed offsets
5122 vector was exactly big enough.
5123
5124 6. Corrected definition of offsetof() from change 5 below.
5125
5126 7. I had screwed up change 6 below and broken the rules for the use of
5127 setjmp(). Now fixed.
5128
5129
5130 Version 1.03 18-Dec-97
5131 ----------------------
5132
5133 1. A erroneous regex with a missing opening parenthesis was correctly
5134 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5135 on some systems.
5136
5137 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5138 it was reported that one broken compiler failed on the former because "code" is
5139 also an independent variable.
5140
5141 3. The erroneous regex a[]b caused an array overrun reference.
5142
5143 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5144 fail on data ending with that character. (It was going on too far, and checking
5145 the next character, typically a binary zero.) This was specific to the
5146 optimized code for single-character negative classes.
5147
5148 5. Added a contributed patch from the TIN world which does the following:
5149
5150 + Add an undef for memmove, in case the the system defines a macro for it.
5151
5152 + Add a definition of offsetof(), in case there isn't one. (I don't know
5153 the reason behind this - offsetof() is part of the ANSI standard - but
5154 it does no harm).
5155
5156 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5157 most of the places where whitespace preceded '#'. I have given up and
5158 allowed the remaining 2 cases to be at the margin.
5159
5160 + Rename some variables in pcre to eliminate shadowing. This seems very
5161 pedantic, but does no harm, of course.
5162
5163 6. Moved the call to setjmp() into its own function, to get rid of warnings
5164 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5165
5166 7. Constructs such as \d{8,} were compiling into the equivalent of
5167 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5168 outcome, but in this particular case used more store than had been allocated,
5169 which caused the bug to be discovered because it threw up an internal error.
5170
5171 8. The debugging code in both pcre and pcretest for outputting the compiled
5172 form of a regex was going wrong in the case of back references followed by
5173 curly-bracketed repeats.
5174
5175
5176 Version 1.02 12-Dec-97
5177 ----------------------
5178
5179 1. Typos in pcre.3 and comments in the source fixed.
5180
5181 2. Applied a contributed patch to get rid of places where it used to remove
5182 'const' from variables, and fixed some signed/unsigned and uninitialized
5183 variable warnings.
5184
5185 3. Added the "runtest" target to Makefile.
5186
5187 4. Set default compiler flag to -O2 rather than just -O.
5188
5189
5190 Version 1.01 19-Nov-97
5191 ----------------------
5192
5193 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5194 like /([ab]*)*/, that is, for classes with more than one character in them.
5195
5196 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5197 as /((?>a*))*/ (a PCRE_EXTRA facility).
5198
5199
5200 Version 1.00 18-Nov-97
5201 ----------------------
5202
5203 1. Added compile-time macros to support systems such as SunOS4 which don't have
5204 memmove() or strerror() but have other things that can be used instead.
5205
5206 2. Arranged that "make clean" removes the executables.
5207
5208
5209 Version 0.99 27-Oct-97
5210 ----------------------
5211
5212 1. Fixed bug in code for optimizing classes with only one character. It was
5213 initializing a 32-byte map regardless, which could cause it to run off the end
5214 of the memory it had got.
5215
5216 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5217
5218
5219 Version 0.98 22-Oct-97
5220 ----------------------
5221
5222 1. Fixed bug in code for handling temporary memory usage when there are more
5223 back references than supplied space in the ovector. This could cause segfaults.
5224
5225
5226 Version 0.97 21-Oct-97
5227 ----------------------
5228
5229 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5230
5231 2. Optimized negated single characters not to use a bit map.
5232
5233 3. Brought error texts together as macro definitions; clarified some of them;
5234 fixed one that was wrong - it said "range out of order" when it meant "invalid
5235 escape sequence".
5236
5237 4. Changed some char * arguments to const char *.
5238
5239 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5240
5241 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5242 pcretest.
5243
5244
5245 Version 0.96 16-Oct-97
5246 ----------------------
5247
5248 1. Added a simple "pgrep" utility to the distribution.
5249
5250 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5251 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5252 where "ddd" means "one or more decimal digits".
5253
5254 3. Fixed serious bug. If a pattern had a back reference, but the call to
5255 pcre_exec() didn't supply a large enough ovector to record the related
5256 identifying subpattern, the match always failed. PCRE now remembers the number
5257 of the largest back reference, and gets some temporary memory in which to save
5258 the offsets during matching if necessary, in order to ensure that
5259 backreferences always work.
5260
5261 4. Increased the compatibility with Perl in a number of ways:
5262
5263 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5264 to request this handling. The option can be set at compile or exec time.
5265
5266 (b) $ matches before a terminating newline by default; an option
5267 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5268 mode). The option can be set at compile or exec time.
5269
5270 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5271 the same as Perl's. If the decimal number it represents is less than 10
5272 or there aren't that many previous left capturing parentheses, an octal
5273 escape is read. Inside a character class, it's always an octal escape,
5274 even if it is a single digit.
5275
5276 (d) An escaped but undefined alphabetic character is taken as a literal,
5277 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5278 escapes.
5279
5280 (e) {0} is now permitted. (The previous item is removed from the compiled
5281 pattern).
5282
5283 5. Changed all the names of code files so that the basic parts are no longer
5284 than 10 characters, and abolished the teeny "globals.c" file.
5285
5286 6. Changed the handling of character classes; they are now done with a 32-byte
5287 bit map always.
5288
5289 7. Added the -d and /D options to pcretest to make it possible to look at the
5290 internals of compilation without having to recompile pcre.
5291
5292
5293 Version 0.95 23-Sep-97
5294 ----------------------
5295
5296 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5297 \x20 at the start of a run of normal characters. These were being treated as
5298 real characters, instead of the source characters being re-checked.
5299
5300
5301 Version 0.94 18-Sep-97
5302 ----------------------
5303
5304 1. The functions are now thread-safe, with the caveat that the global variables
5305 containing pointers to malloc() and free() or alternative functions are the
5306 same for all threads.
5307
5308 2. Get pcre_study() to generate a bitmap of initial characters for non-
5309 anchored patterns when this is possible, and use it if passed to pcre_exec().
5310
5311
5312 Version 0.93 15-Sep-97
5313 ----------------------
5314
5315 1. /(b)|(:+)/ was computing an incorrect first character.
5316
5317 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5318 but not actually doing anything yet.
5319
5320 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5321 as Perl does (e.g. [-az] or [az-]).
5322
5323 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5324 all possible positions.
5325
5326 5. Split up into different modules to avoid including unneeded functions in a
5327 compiled binary. However, compile and exec are still in one module. The "study"
5328 function is split off.
5329
5330 6. The character tables are now in a separate module whose source is generated
5331 by an auxiliary program - but can then be edited by hand if required. There are
5332 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5333 toupper() in the code.
5334
5335 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5336 make them global. Abolish the function for setting them, as the caller can now
5337 set them directly.
5338
5339
5340 Version 0.92 11-Sep-97
5341 ----------------------
5342
5343 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5344 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5345
5346 2. Caseless matching was not working in character classes if the characters in
5347 the pattern were in upper case.
5348
5349 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5350
5351 4. Make PCRE_ANCHORED public and accept as a compile option.
5352
5353 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5354 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5355 pass them.
5356
5357 6. Give an error if bad option bits passed at compile or run time.
5358
5359 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5360 pcretest to cause it to pass that flag.
5361
5362 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5363 options, and the first character, if set.
5364
5365 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5366
5367
5368 Version 0.91 10-Sep-97
5369 ----------------------
5370
5371 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5372 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5373
5374 2. PCRE was looping on encountering an indefinitely repeated back reference to
5375 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5376 Perl does - treats the match as successful.
5377
5378 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5