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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1408 - (show annotations)
Tue Dec 3 16:27:00 2013 UTC (5 years, 11 months ago) by ph10
File size: 257665 byte(s)
Add support for [[:<:]] and [[:>:]] as a transition aid.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.34 19-November-2013
5 -----------------------------
6
7 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
8 executable memory. Patch inspired by Carsten Klein.
9
10 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
11 this macro is never tested and has no effect, because the work to support
12 coverage involves only compiling and linking options and special targets in
13 the Makefile. The comment in config.h implied that defining the macro would
14 enable coverage support, which is totally false. There was also support for
15 setting this macro in the CMake files (my fault, I just copied it from
16 configure). SUPPORT_GCOV has now been removed.
17
18 3. Make a small performance improvement in strlen16() and strlen32() in
19 pcretest.
20
21 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
22 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
23 these cases). There was also one in pcretest.c.
24
25 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
26
27 6. In UTF mode, the code for checking whether a group could match an empty
28 string (which is used for indefinitely repeated groups to allow for
29 breaking an infinite loop) was broken when the group contained a repeated
30 negated single-character class with a character that occupied more than one
31 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
32 UTF-8 mode). The effect was undefined: the group might or might not be
33 deemed as matching an empty string, or the program might have crashed.
34
35 7. The code for checking whether a group could match an empty string was not
36 recognizing that \h, \H, \v, \V, and \R must match a character.
37
38 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
39 an empty string. If it can, pcretest shows this in its information output.
40
41 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
42 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
43 matched by pcre_exec() without using JIT:
44
45 (a) If the rest of the pattern did not match after a maximal run of
46 grapheme clusters, the code for backing up to try with fewer of them
47 did not always back up over a full grapheme when characters that do not
48 have the modifier quality were involved, e.g. Hangul syllables.
49
50 (b) If the match point in a subject started with modifier character, and
51 there was no match, the code could incorrectly back up beyond the match
52 point, and potentially beyond the first character in the subject,
53 leading to a segfault or an incorrect match result.
54
55 10. A conditional group with an assertion condition could lead to PCRE
56 recording an incorrect first data item for a match if no other first data
57 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
58 first data item, and therefore matched "ca" after "c" instead of at the
59 start.
60
61 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
62 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
63
64 12. The source of pcregrep now includes z/OS-specific code so that it can be
65 compiled for z/OS as part of the special z/OS distribution.
66
67 13. Added the -T and -TM options to pcretest.
68
69 14. The code in pcre_compile.c for creating the table of named capturing groups
70 has been refactored. Instead of creating the table dynamically during the
71 actual compiling pass, the information is remembered during the pre-compile
72 pass (on the stack unless there are more than 20 named groups, in which
73 case malloc() is used) and the whole table is created before the actual
74 compile happens. This has simplified the code (it is now nearly 150 lines
75 shorter) and prepared the way for better handling of references to groups
76 with duplicate names.
77
78 15. A back reference to a named subpattern when there is more than one of the
79 same name now checks them in the order in which they appear in the pattern.
80 The first one that is set is used for the reference. Previously only the
81 first one was inspected. This change makes PCRE more compatible with Perl.
82
83 16. Unicode character properties were updated from Unicode 6.3.0.
84
85 17. The compile-time code for auto-possessification has been refactored, based
86 on a patch by Zoltan Herczeg. It now happens after instead of during
87 compilation. The code is cleaner, and more cases are handled. The option
88 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
89 options in pcretest are provided to set it. It can also be set by
90 (*NO_AUTO_POSSESS) at the start of a pattern.
91
92 18. The character VT has been added to the default ("C" locale) set of
93 characters that match \s and are generally treated as white space,
94 following this same change in Perl 5.18. There is now no difference between
95 "Perl space" and "POSIX space". Whether VT is treated as white space in
96 other locales depends on the locale.
97
98 19. The code for checking named groups as conditions, either for being set or
99 for being recursed, has been refactored (this is related to 14 and 15
100 above). Processing unduplicated named groups should now be as fast at
101 numerical groups, and processing duplicated groups should be faster than
102 before.
103
104 20. Two patches to the CMake build system, by Alexander Barkov:
105
106 (1) Replace the "source" command by "." in CMakeLists.txt because
107 "source" is a bash-ism.
108
109 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
110 without these the CMake build does not work on Solaris.
111
112 21. Perl has changed its handling of \8 and \9. If there is no previously
113 encountered capturing group of those numbers, they are treated as the
114 literal characters 8 and 9 instead of a binary zero followed by the
115 literals. PCRE now does the same.
116
117 22. Following Perl, added \o{} to specify codepoints in octal, making it
118 possible to specify values greater than 0777 and also making them
119 unambiguous.
120
121 23. Perl now gives an error for missing closing braces after \x{... instead of
122 treating the string as literal. PCRE now does the same.
123
124 24. RunTest used to grumble if an inappropriate test was selected explicitly,
125 but just skip it when running all tests. This make it awkward to run ranges
126 of tests when one of them was inappropriate. Now it just skips any
127 inappropriate tests, as it always did when running all tests.
128
129 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
130 character types such as \d or \w, too many callouts were inserted, and the
131 data that they returned was rubbish.
132
133 26. In UCP mode, \s was not matching two of the characters that Perl matches,
134 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
135 were matched by \h. The code has now been refactored so that the lists of
136 the horizontal and vertical whitespace characters used for \h and \v (which
137 are defined only in one place) are now also used for \s.
138
139 27. Add JIT support for the 64 bit TileGX architecture.
140 Patch by Jiong Wang (Tilera Corporation).
141
142 28. Possessive quantifiers for classes (both explicit and automatically
143 generated) now use special opcodes instead of wrapping in ONCE brackets.
144
145 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
146 (because it's meaningless), this was not happening when PCRE_CASELESS was
147 set. Not wrong, but inefficient.
148
149 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
150 properties for \w, \d, etc) is present in a test regex. Otherwise if the
151 test contains no characters greater than 255, Perl doesn't realise it
152 should be using Unicode semantics.
153
154 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
155 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
156 does in Unicode mode.
157
158 32. Added the "forbid" facility to pcretest so that putting tests into the
159 wrong test files can sometimes be quickly detected.
160
161 33. There is now a limit (default 250) on the depth of nesting of parentheses.
162 This limit is imposed to control the amount of system stack used at compile
163 time. It can be changed at build time by --with-parens-nest-limit=xxx or
164 the equivalent in CMake.
165
166 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
167 errors. Perl warns for these when in warning mode, but PCRE has no facility
168 for giving warnings.
169
170 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
171 However, this was not working for (?!) because it is optimized to (*FAIL),
172 for which PCRE does not allow quantifiers. The optimization is now disabled
173 when a quantifier follows (?!). I can't see any use for this, but it makes
174 things uniform.
175
176 36. Perl no longer allows group names to start with digits, so I have made this
177 change also in PCRE. It simplifies the code a bit.
178
179 37. In extended mode, Perl ignores spaces before a + that indicates a
180 possessive quantifier. PCRE allowed a space before the quantifier, but not
181 before the possessive +. It now does.
182
183 38. The use of \K (reset reported match start) within a repeated possessive
184 group such as (a\Kb)*+ was not working.
185
186 40. Document that the same character tables must be used at compile time and
187 run time, and that the facility to pass tables to pcre_exec() and
188 pcre_dfa_exec() is for use only with saved/restored patterns.
189
190 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
191 features for Builds with MSVC:
192
193 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
194 builds). The .rc files can be used to set FileDescription and many other
195 attributes.
196
197 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
198 This allows higher-level build scripts which want .pdb files to avoid
199 hard-coding the exact files needed."
200
201 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
202 mean "start of word" and "end of word", respectively, as a transition aid.
203
204
205 Version 8.33 28-May-2013
206 ------------------------
207
208 1. Added 'U' to some constants that are compared to unsigned integers, to
209 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
210 variables that are added to signed variables, to ensure the result is
211 signed and can be negated.
212
213 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
214 CMake config files.
215
216 3. Revise the creation of config.h.generic so that all boolean macros are
217 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
218 overriding via -D on the command line possible.
219
220 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
221 to unsigned int is reported to make a quite noticeable speed difference in
222 a specific Windows environment. Testing on Linux did also appear to show
223 some benefit (and it is clearly not harmful). Also fixed the definition of
224 Xop which should be unsigned.
225
226 5. Related to (4), changing the definition of the intermediate variable cc
227 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
228 improvements.
229
230 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
231 unnecessary spaces.
232
233 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
234 and later.
235
236 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
237
238 9. Optimizing fast_forward_start_bits in JIT.
239
240 10. Adding support for callouts in JIT, and fixing some issues revealed
241 during this work. Namely:
242
243 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
244
245 (b) Minimum length was not checked before the matching is started.
246
247 11. The value of capture_last that is passed to callouts was incorrect in some
248 cases when there was a capture on one path that was subsequently abandoned
249 after a backtrack. Also, the capture_last value is now reset after a
250 recursion, since all captures are also reset in this case.
251
252 12. The interpreter no longer returns the "too many substrings" error in the
253 case when an overflowing capture is in a branch that is subsequently
254 abandoned after a backtrack.
255
256 13. In the pathological case when an offset vector of size 2 is used, pcretest
257 now prints out the matched string after a yield of 0 or 1.
258
259 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
260 Only supported by the JIT compiler at the moment.
261
262 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
263
264 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
265 offset of the starting point of the matching process, provided the offsets
266 vector is large enough.
267
268 17. The \A escape now records a lookbehind value of 1, though its execution
269 does not actually inspect the previous character. This is to ensure that,
270 in partial multi-segment matching, at least one character from the old
271 segment is retained when a new segment is processed. Otherwise, if there
272 are no lookbehinds in the pattern, \A might match incorrectly at the start
273 of a new segment.
274
275 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
276
277 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
278 gives some modest performance improvement in 8-bit mode.
279
280 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
281 be expressed in certain programming languages using Universal Character
282 Names.
283
284 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
285 which points out that "non characters" are not "characters that may not
286 appear in Unicode strings" but rather "characters that are reserved for
287 internal use and have only local meaning".
288
289 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
290 there was a conditional group that depended on an assertion, if the
291 assertion was false, the callout that immediately followed the alternation
292 in the condition was skipped when pcre_exec() was used for matching.
293
294 23. Allow an explicit callout to be inserted before an assertion that is the
295 condition for a conditional group, for compatibility with automatic
296 callouts, which always insert a callout at this point.
297
298 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
299 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
300
301 25. (*PRUNE) is now supported by the JIT compiler.
302
303 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
304
305 27. Fix the case where there are two or more SKIPs with arguments that may be
306 ignored.
307
308 28. (*SKIP) is now supported by the JIT compiler.
309
310 29. (*THEN) is now supported by the JIT compiler.
311
312 30. Update RunTest with additional test selector options.
313
314 31. The way PCRE handles backtracking verbs has been changed in two ways.
315
316 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
317 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
318 backtracking. In some cases this makes it more Perl-compatible, but Perl's
319 rather obscure rules do not always do the same thing.
320
321 (2) Previously, backtracking verbs were confined within assertions. This is
322 no longer the case for positive assertions, except for (*ACCEPT). Again,
323 this sometimes improves Perl compatibility, and sometimes does not.
324
325 32. A number of tests that were in test 2 because Perl did things differently
326 have been moved to test 1, because either Perl or PCRE has changed, and
327 these tests are now compatible.
328
329 32. Backtracking control verbs are now handled in the same way in JIT and
330 interpreter.
331
332 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
333 contained a forward subroutine reference caused a compile error.
334
335 34. Auto-detect and optimize limited repetitions in JIT.
336
337 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
338 blocking (*UTF) etc.
339
340 36. In the interpreter, maximizing pattern repetitions for characters and
341 character types now use tail recursion, which reduces stack usage.
342
343 37. The value of the max lookbehind was not correctly preserved if a compiled
344 and saved regex was reloaded on a host of different endianness.
345
346 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
347 of the compiled pattern block, expand the flags field from 16 to 32 bits
348 because it was almost full.
349
350 39. Try madvise first before posix_madvise.
351
352 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
353 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
354 original change locked out all matches of empty strings. This has been
355 changed so that one match of an empty string per line is recognized.
356 Subsequent searches on the same line (for colouring or for --only-matching,
357 for example) do not recognize empty strings.
358
359 41. Applied a user patch to fix a number of spelling mistakes in comments.
360
361 42. Data lines longer than 65536 caused pcretest to crash.
362
363 43. Clarified the data type for length and startoffset arguments for pcre_exec
364 and pcre_dfa_exec in the function-specific man pages, where they were
365 explicitly stated to be in bytes, never having been updated. I also added
366 some clarification to the pcreapi man page.
367
368 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
369 a segmentation fault.
370
371
372 Version 8.32 30-November-2012
373 -----------------------------
374
375 1. Improved JIT compiler optimizations for first character search and single
376 character iterators.
377
378 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
379 Patch by Daniel Richard G.
380
381 3. Single character iterator optimizations in the JIT compiler.
382
383 4. Improved JIT compiler optimizations for character ranges.
384
385 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
386 Reported by Giuseppe D'Angelo.
387
388 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
389 of a line, was being set incorrectly in cases where .* appeared inside
390 atomic brackets at the start of a pattern, or where there was a subsequent
391 *PRUNE or *SKIP.
392
393 7. Improved instruction cache flush for POWER/PowerPC.
394 Patch by Daniel Richard G.
395
396 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
397 grep:
398
399 (a) There is now no limit to the number of patterns to be matched.
400
401 (b) An error is given if a pattern is too long.
402
403 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
404 are now supported.
405
406 (d) --exclude-from and --include-from (multiple use) have been added.
407
408 (e) Exclusions and inclusions now apply to all files and directories, not
409 just to those obtained from scanning a directory recursively.
410
411 (f) Multiple uses of -f and --file-list are now supported.
412
413 (g) In a Windows environment, the default for -d has been changed from
414 "read" (the GNU grep default) to "skip", because otherwise the presence
415 of a directory in the file list provokes an error.
416
417 (h) The documentation has been revised and clarified in places.
418
419 9. Improve the matching speed of capturing brackets.
420
421 10. Changed the meaning of \X so that it now matches a Unicode extended
422 grapheme cluster.
423
424 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
425 out POSIX threads when JIT support is configured.
426
427 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
428
429 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
430 structure could go wrong in environments where size_t is not the same size
431 as int.
432
433 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
434 set.
435
436 15. The EBCDIC support had decayed; later updates to the code had included
437 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
438 general tidy up of EBCDIC-related issues, and the documentation was also
439 not quite right. There is now a test that can be run on ASCII systems to
440 check some of the EBCDIC-related things (but is it not a full test).
441
442 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
443 in a small tidy to the code.
444
445 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
446
447 18. If the --only-matching (-o) option in pcregrep is specified multiple
448 times, each one causes appropriate output. For example, -o1 -o2 outputs the
449 substrings matched by the 1st and 2nd capturing parentheses. A separating
450 string can be specified by --om-separator (default empty).
451
452 19. Improving the first n character searches.
453
454 20. Turn case lists for horizontal and vertical white space into macros so that
455 they are defined only once.
456
457 21. This set of changes together give more compatible Unicode case-folding
458 behaviour for characters that have more than one other case when UCP
459 support is available.
460
461 (a) The Unicode property table now has offsets into a new table of sets of
462 three or more characters that are case-equivalent. The MultiStage2.py
463 script that generates these tables (the pcre_ucd.c file) now scans
464 CaseFolding.txt instead of UnicodeData.txt for character case
465 information.
466
467 (b) The code for adding characters or ranges of characters to a character
468 class has been abstracted into a generalized function that also handles
469 case-independence. In UTF-mode with UCP support, this uses the new data
470 to handle characters with more than one other case.
471
472 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
473 whose other case is greater than 256 are now correctly matched
474 caselessly. Previously, the high codepoint matched the low one, but not
475 vice versa.
476
477 (d) The processing of \h, \H, \v, and \ in character classes now makes use
478 of the new class addition function, using character lists defined as
479 macros alongside the case definitions of 20 above.
480
481 (e) Caseless back references now work with characters that have more than
482 one other case.
483
484 (f) General caseless matching of characters with more than one other case
485 is supported.
486
487 22. Unicode character properties were updated from Unicode 6.2.0
488
489 23. Improved CMake support under Windows. Patch by Daniel Richard G.
490
491 24. Add support for 32-bit character strings, and UTF-32
492
493 25. Major JIT compiler update (code refactoring and bugfixing).
494 Experimental Sparc 32 support is added.
495
496 26. Applied a modified version of Daniel Richard G's patch to create
497 pcre.h.generic and config.h.generic by "make" instead of in the
498 PrepareRelease script.
499
500 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
501 pcre_compile.c when checking for a zero character.
502
503 28. Introducing a native interface for JIT. Through this interface, the compiled
504 machine code can be directly executed. The purpose of this interface is to
505 provide fast pattern matching, so several sanity checks are not performed.
506 However, feature tests are still performed. The new interface provides
507 1.4x speedup compared to the old one.
508
509 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
510 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
511 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
512
513 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
514 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
515 the "old" RFC 2279). Instead, it ended up passing a negative length to
516 pcre_exec().
517
518 31. Add support for GCC's visibility feature to hide internal functions.
519
520 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
521 "unknown -C option" after outputting 0 or 1.
522
523 33. There is now support for generating a code coverage report for the test
524 suite in environments where gcc is the compiler and lcov is installed. This
525 is mainly for the benefit of the developers.
526
527 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
528 unaddressable using valgrind annotations, allowing valgrind to detect
529 invalid memory accesses. This is mainly for the benefit of the developers.
530
531 25. (*UTF) can now be used to start a pattern in any of the three libraries.
532
533 26. Give configure error if --enable-cpp but no C++ compiler found.
534
535
536 Version 8.31 06-July-2012
537 -------------------------
538
539 1. Fixing a wrong JIT test case and some compiler warnings.
540
541 2. Removed a bashism from the RunTest script.
542
543 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
544 to unsigned type, result still unsigned" that was given by an MS compiler
545 on encountering the code "-sizeof(xxx)".
546
547 4. Partial matching support is added to the JIT compiler.
548
549 5. Fixed several bugs concerned with partial matching of items that consist
550 of more than one character:
551
552 (a) /^(..)\1/ did not partially match "aba" because checking references was
553 done on an "all or nothing" basis. This also applied to repeated
554 references.
555
556 (b) \R did not give a hard partial match if \r was found at the end of the
557 subject.
558
559 (c) \X did not give a hard partial match after matching one or more
560 characters at the end of the subject.
561
562 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
563 a partial match for the string "\r".
564
565 (e) When newline was set to CRLF, the metacharacter "." did not recognize
566 a partial match for a CR character at the end of the subject string.
567
568 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
569 running pcretest, the text "(JIT)" added to the output whenever JIT is
570 actually used to run the match.
571
572 7. Individual JIT compile options can be set in pcretest by following -s+[+]
573 or /S+[+] with a digit between 1 and 7.
574
575 8. OP_NOT now supports any UTF character not just single-byte ones.
576
577 9. (*MARK) control verb is now supported by the JIT compiler.
578
579 10. The command "./RunTest list" lists the available tests without actually
580 running any of them. (Because I keep forgetting what they all are.)
581
582 11. Add PCRE_INFO_MAXLOOKBEHIND.
583
584 12. Applied a (slightly modified) user-supplied patch that improves performance
585 when the heap is used for recursion (compiled with --disable-stack-for-
586 recursion). Instead of malloc and free for each heap frame each time a
587 logical recursion happens, frames are retained on a chain and re-used where
588 possible. This sometimes gives as much as 30% improvement.
589
590 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
591 call.
592
593 14. As documented, (*COMMIT) is now confined to within a positive assertion.
594
595 15. It is now possible to link pcretest with libedit as an alternative to
596 libreadline.
597
598 16. (*COMMIT) control verb is now supported by the JIT compiler.
599
600 17. The Unicode data tables have been updated to Unicode 6.1.0.
601
602 18. Added --file-list option to pcregrep.
603
604 19. Added binary file support to pcregrep, including the -a, --binary-files,
605 -I, and --text options.
606
607 20. The madvise function is renamed for posix_madvise for QNX compatibility
608 reasons. Fixed by Giuseppe D'Angelo.
609
610 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
611
612 22. Changed the option for creating symbolic links for 16-bit man pages from
613 -s to -sf so that re-installing does not cause issues.
614
615 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
616
617 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
618 matches in certain environments (the workspace was not being correctly
619 retained). Also added to pcre_dfa_exec() a simple plausibility check on
620 some of the workspace data at the beginning of a restart.
621
622 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
623 was not doing so when it should - probably a typo introduced by SVN 528
624 (change 8.10/14).
625
626 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
627 \w+ when the character tables indicated that \x{c4} was a word character.
628 There were several related cases, all because the tests for doing a table
629 lookup were testing for characters less than 127 instead of 255.
630
631 27. If a pattern contains capturing parentheses that are not used in a match,
632 their slots in the ovector are set to -1. For those that are higher than
633 any matched groups, this happens at the end of processing. In the case when
634 there were back references that the ovector was too small to contain
635 (causing temporary malloc'd memory to be used during matching), and the
636 highest capturing number was not used, memory off the end of the ovector
637 was incorrectly being set to -1. (It was using the size of the temporary
638 memory instead of the true size.)
639
640 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
641 ovector size, it uses memory at the end of the block that it has got.
642
643 29. Check for an overlong MARK name and give an error at compile time. The
644 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
645
646 30. JIT compiler update.
647
648 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
649 Rill for the patch.
650
651 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
652
653 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
654
655 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
656 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
657 was enabled.
658
659 35. Improve JIT code generation for greedy plus quantifier.
660
661 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
662 1 to "aa" instead of to an empty string. The bug affected repeated groups
663 that could potentially match an empty string.
664
665 37. Optimizing single character iterators in JIT.
666
667 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
668 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
669 codepoints that are too big for the mode are faulted, and in a UTF mode,
670 disallowed codepoints are also faulted.
671
672 39. If PCRE was compiled with UTF support, in three places in the DFA
673 matcher there was code that should only have been obeyed in UTF mode, but
674 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
675 processing when bytes with values greater than 127 were present. In 16-bit
676 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
677 both cases the values are those that cannot be the first data item in a UTF
678 character. The three items that might have provoked this were recursions,
679 possessively repeated groups, and atomic groups.
680
681 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
682 and pcregrep, because some OS require shared objects to be explicitly
683 passed to ld, causing the link step to fail if they are not.
684
685 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
686 mode, patterns that started with \h* or \R* might be incorrectly matched.
687
688
689 Version 8.30 04-February-2012
690 -----------------------------
691
692 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
693 name is defined in ctype.h.
694
695 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
696 only in quite long subpatterns.
697
698 3. Removed the function pcre_info(), which has been obsolete and deprecated
699 since it was replaced by pcre_fullinfo() in February 2000.
700
701 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
702 match a (*MARK), and the match failed at the start of the subject, a
703 reference to memory before the start of the subject could occur. This bug
704 was introduced by fix 17 of release 8.21.
705
706 5. A reference to an unset group with zero minimum repetition was giving
707 totally wrong answers (in non-JavaScript-compatibility mode). For example,
708 /(another)?(\1?)test/ matched against "hello world test". This bug was
709 introduced in release 8.13.
710
711 6. Add support for 16-bit character strings (a large amount of work involving
712 many changes and refactorings).
713
714 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
715 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
716 from a file.
717
718 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
719 rounding is not applied in this particular case).
720
721 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
722 if they appear, or are escaped, in patterns.
723
724 10. Get rid of a number of -Wunused-but-set-variable warnings.
725
726 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
727 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
728 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
729 also returns the mark "x". This bug applied to capturing parentheses,
730 non-capturing parentheses, and atomic parentheses. It also applied to some
731 assertions.
732
733 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
734 information out of configure.ac instead of relying on pcre.h.generic, which
735 is not stored in the repository.
736
737 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
738 -lreadline.
739
740 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
741
742 15. Applied Graycode's patch to put the top-level frame on the stack rather
743 than the heap when not using the stack for recursion. This gives a
744 performance improvement in many cases when recursion is not deep.
745
746 16. Experimental code added to "pcretest -C" to output the stack frame size.
747
748
749 Version 8.21 12-Dec-2011
750 ------------------------
751
752 1. Updating the JIT compiler.
753
754 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
755 are added as well.
756
757 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
758 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
759 calling _pcre_jit_exec. Some extra comments are added.
760
761 4. (*MARK) settings inside atomic groups that do not contain any capturing
762 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
763 was introduced by change 18 for 8.20.
764
765 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
766 ECMA-262 standard.
767
768 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
769 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
770 This bug was probably introduced by change 9 of 8.13.
771
772 7. While fixing 6 above, I noticed that a number of other items were being
773 incorrectly rejected as "not fixed length". This arose partly because newer
774 opcodes had not been added to the fixed-length checking code. I have (a)
775 corrected the bug and added tests for these items, and (b) arranged for an
776 error to occur if an unknown opcode is encountered while checking for fixed
777 length instead of just assuming "not fixed length". The items that were
778 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
779 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
780 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
781
782 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
783 being incorrectly compiled and would have given unpredicatble results.
784
785 9. A possessively repeated subpattern with minimum repeat count greater than
786 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
787 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
788 the first (A) could occur when it should not.
789
790 10. Add a cast and remove a redundant test from the code.
791
792 11. JIT should use pcre_malloc/pcre_free for allocation.
793
794 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
795 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
796 is anything other than /usr, -L is still shown).
797
798 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
799
800 14. Perl does not support \N without a following name in a [] class; PCRE now
801 also gives an error.
802
803 15. If a forward reference was repeated with an upper limit of around 2000,
804 it caused the error "internal error: overran compiling workspace". The
805 maximum number of forward references (including repeats) was limited by the
806 internal workspace, and dependent on the LINK_SIZE. The code has been
807 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
808 the default depends on LINK_SIZE. There is a new upper limit (for safety)
809 of around 200,000 forward references. While doing this, I also speeded up
810 the filling in of repeated forward references.
811
812 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
813 incorrectly expecting the subject to contain another "a" after the start.
814
815 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
816 in the match, the SKIP should be ignored. This was not happening; instead
817 the SKIP was being treated as NOMATCH. For patterns such as
818 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
819 tested.
820
821 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
822 now much more compatible with Perl, in particular in cases where the result
823 is a non-match for a non-anchored pattern. For example, if
824 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
825 "m", where previously it did not return a name. A side effect of this
826 change is that for partial matches, the last encountered mark name is
827 returned, as for non matches. A number of tests that were previously not
828 Perl-compatible have been moved into the Perl-compatible test files. The
829 refactoring has had the pleasing side effect of removing one argument from
830 the match() function, thus reducing its stack requirements.
831
832 19. If the /S+ option was used in pcretest to study a pattern using JIT,
833 subsequent uses of /S (without +) incorrectly behaved like /S+.
834
835 21. Retrieve executable code size support for the JIT compiler and fixing
836 some warnings.
837
838 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
839 not work when the shorter character appeared right at the end of the
840 subject string.
841
842 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
843 systems.
844
845 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
846 output it when the /M option is used in pcretest.
847
848 25. The CheckMan script was not being included in the distribution. Also, added
849 an explicit "perl" to run Perl scripts from the PrepareRelease script
850 because this is reportedly needed in Windows.
851
852 26. If study data was being save in a file and studying had not found a set of
853 "starts with" bytes for the pattern, the data written to the file (though
854 never used) was taken from uninitialized memory and so caused valgrind to
855 complain.
856
857 27. Updated RunTest.bat as provided by Sheri Pierce.
858
859 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
860
861 29. Computation of memory usage for the table of capturing group names was
862 giving an unnecessarily large value.
863
864
865 Version 8.20 21-Oct-2011
866 ------------------------
867
868 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
869 a POSIX class. After further experiments with Perl, which convinced me that
870 Perl has bugs and confusions, a closing square bracket is no longer allowed
871 in a POSIX name. This bug also affected patterns with classes that started
872 with full stops.
873
874 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
875 captured substring, but while checking the failing first alternative,
876 substring 1 is temporarily captured. If the output vector supplied to
877 pcre_exec() was not big enough for this capture, the yield of the function
878 was still zero ("insufficient space for captured substrings"). This cannot
879 be totally fixed without adding another stack variable, which seems a lot
880 of expense for a edge case. However, I have improved the situation in cases
881 such as /(a)(b)x|abc/ matched against "abc", where the return code
882 indicates that fewer than the maximum number of slots in the ovector have
883 been set.
884
885 3. Related to (2) above: when there are more back references in a pattern than
886 slots in the output vector, pcre_exec() uses temporary memory during
887 matching, and copies in the captures as far as possible afterwards. It was
888 using the entire output vector, but this conflicts with the specification
889 that only 2/3 is used for passing back captured substrings. Now it uses
890 only the first 2/3, for compatibility. This is, of course, another edge
891 case.
892
893 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
894 main code base, and can be used by building with --enable-jit. When this is
895 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
896 runtime --no-jit option is given.
897
898 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
899 ovector, the return from the function was zero, implying that there were
900 other matches that did not fit. The correct "exactly full" value is now
901 returned.
902
903 6. If a subpattern that was called recursively or as a subroutine contained
904 (*PRUNE) or any other control that caused it to give a non-standard return,
905 invalid errors such as "Error -26 (nested recursion at the same subject
906 position)" or even infinite loops could occur.
907
908 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
909 computing the minimum length on reaching *ACCEPT, and so ended up with the
910 wrong value of 1 rather than 0. Further investigation indicates that
911 computing a minimum subject length in the presence of *ACCEPT is difficult
912 (think back references, subroutine calls), and so I have changed the code
913 so that no minimum is registered for a pattern that contains *ACCEPT.
914
915 8. If (*THEN) was present in the first (true) branch of a conditional group,
916 it was not handled as intended. [But see 16 below.]
917
918 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
919 Sheri Pierce.
920
921 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
922 the first byte in a match must be "a".
923
924 11. Change 17 for 8.13 increased the recursion depth for patterns like
925 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
926 pattern contains any instances of (*THEN). If it does not, the old
927 optimizations are restored. It would be nice to do this on a per-group
928 basis, but at the moment that is not feasible.
929
930 12. In some environments, the output of pcretest -C is CRLF terminated. This
931 broke RunTest's code that checks for the link size. A single white space
932 character after the value is now allowed for.
933
934 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
935 For "fr", it uses the Windows-specific input and output files.
936
937 14. If (*THEN) appeared in a group that was called recursively or as a
938 subroutine, it did not work as intended. [But see next item.]
939
940 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
941 pattern fragments (but not containing any | characters). If A and B are
942 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
943 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
944 D. In other words, Perl considers parentheses that do not contain any |
945 characters to be part of a surrounding alternative, whereas PCRE was
946 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
947 differently. PCRE now behaves in the same way as Perl, except in the case
948 of subroutine/recursion calls such as (?1) which have in any case always
949 been different (but PCRE had them first :-).
950
951 16. Related to 15 above: Perl does not treat the | in a conditional group as
952 creating alternatives. Such a group is treated in the same way as an
953 ordinary group without any | characters when processing (*THEN). PCRE has
954 been changed to match Perl's behaviour.
955
956 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
957 RunGrepTest script failed.
958
959 18. Change 22 for version 13 caused atomic groups to use more stack. This is
960 inevitable for groups that contain captures, but it can lead to a lot of
961 stack use in large patterns. The old behaviour has been restored for atomic
962 groups that do not contain any capturing parentheses.
963
964 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
965 suppress the check for a minimum subject length at run time. (If it was
966 given to pcre_exec() or pcre_dfa_exec() it did work.)
967
968 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
969 fail to work when decoding hex characters in data strings in EBCDIC
970 environments.
971
972 21. It appears that in at least one Mac OS environment, the isxdigit() function
973 is implemented as a macro that evaluates to its argument more than once,
974 contravening the C 90 Standard (I haven't checked a later standard). There
975 was an instance in pcretest which caused it to go wrong when processing
976 \x{...} escapes in subject strings. The has been rewritten to avoid using
977 things like p++ in the argument of isxdigit().
978
979
980 Version 8.13 16-Aug-2011
981 ------------------------
982
983 1. The Unicode data tables have been updated to Unicode 6.0.0.
984
985 2. Two minor typos in pcre_internal.h have been fixed.
986
987 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
988 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
989 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
990
991 4. There were a number of related bugs in the code for matching backrefences
992 caselessly in UTF-8 mode when codes for the characters concerned were
993 different numbers of bytes. For example, U+023A and U+2C65 are an upper
994 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
995 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
996 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
997 2-byte code at the end of the subject (it thought there wasn't enough data
998 left).
999
1000 5. Comprehensive information about what went wrong is now returned by
1001 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1002 as the output vector has at least 2 elements. The offset of the start of
1003 the failing character and a reason code are placed in the vector.
1004
1005 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1006 now returned is for the first byte of the failing character, instead of the
1007 last byte inspected. This is an incompatible change, but I hope it is small
1008 enough not to be a problem. It makes the returned offset consistent with
1009 pcre_exec() and pcre_dfa_exec().
1010
1011 7. pcretest now gives a text phrase as well as the error number when
1012 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1013 failure, the offset and reason code are output.
1014
1015 8. When \R was used with a maximizing quantifier it failed to skip backwards
1016 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1017 back over a single character (\n). This seems wrong (because it treated the
1018 two characters as a single entity when going forwards), conflicts with the
1019 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1020 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1021 has been changed.
1022
1023 9. Some internal refactoring has changed the processing so that the handling
1024 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1025 time (the PCRE_DOTALL option was changed this way some time ago: version
1026 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1027 which was always a bit of a fudge. It also means that there is one less
1028 argument for the match() function, which reduces its stack requirements
1029 slightly. This change also fixes an incompatibility with Perl: the pattern
1030 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1031
1032 10. More internal refactoring has drastically reduced the number of recursive
1033 calls to match() for possessively repeated groups such as (abc)++ when
1034 using pcre_exec().
1035
1036 11. While implementing 10, a number of bugs in the handling of groups were
1037 discovered and fixed:
1038
1039 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1040 (a|)*(?1) gave a compile-time internal error.
1041 ((a|)+)+ did not notice that the outer group could match an empty string.
1042 (^a|^)+ was not marked as anchored.
1043 (.*a|.*)+ was not marked as matching at start or after a newline.
1044
1045 12. Yet more internal refactoring has removed another argument from the match()
1046 function. Special calls to this function are now indicated by setting a
1047 value in a variable in the "match data" data block.
1048
1049 13. Be more explicit in pcre_study() instead of relying on "default" for
1050 opcodes that mean there is no starting character; this means that when new
1051 ones are added and accidentally left out of pcre_study(), testing should
1052 pick them up.
1053
1054 14. The -s option of pcretest has been documented for ages as being an old
1055 synonym of -m (show memory usage). I have changed it to mean "force study
1056 for every regex", that is, assume /S for every regex. This is similar to -i
1057 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1058 using it. It makes it easier to run collections of tests with and without
1059 study enabled, and thereby test pcre_study() more easily. All the standard
1060 tests are now run with and without -s (but some patterns can be marked as
1061 "never study" - see 20 below).
1062
1063 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1064 restoration of the capturing data to the outer values was not happening
1065 correctly.
1066
1067 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1068 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1069 pattern had matched an empty string, and so incorrectly returned a no
1070 match.
1071
1072 17. There was optimizing code for the last branch of non-capturing parentheses,
1073 and also for the obeyed branch of a conditional subexpression, which used
1074 tail recursion to cut down on stack usage. Unfortunately, now that there is
1075 the possibility of (*THEN) occurring in these branches, tail recursion is
1076 no longer possible because the return has to be checked for (*THEN). These
1077 two optimizations have therefore been removed. [But see 8.20/11 above.]
1078
1079 18. If a pattern containing \R was studied, it was assumed that \R always
1080 matched two bytes, thus causing the minimum subject length to be
1081 incorrectly computed because \R can also match just one byte.
1082
1083 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1084 was incorrectly computed.
1085
1086 20. If /S is present twice on a test pattern in pcretest input, it now
1087 *disables* studying, thereby overriding the use of -s on the command line
1088 (see 14 above). This is necessary for one or two tests to keep the output
1089 identical in both cases.
1090
1091 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1092 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1093
1094 22. When an atomic group that contained a capturing parenthesis was
1095 successfully matched, but the branch in which it appeared failed, the
1096 capturing was not being forgotten if a higher numbered group was later
1097 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1098 group 1 to "a", when in fact it should be unset. This applied to multi-
1099 branched capturing and non-capturing groups, repeated or not, and also to
1100 positive assertions (capturing in negative assertions does not happen
1101 in PCRE) and also to nested atomic groups.
1102
1103 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1104 subject after a captured substring, to make it easier to tell which of a
1105 number of identical substrings has been captured.
1106
1107 24. The way atomic groups are processed by pcre_exec() has been changed so that
1108 if they are repeated, backtracking one repetition now resets captured
1109 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1110 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1111 "aaa". Previously, it would have been "a". As part of this code
1112 refactoring, the way recursive calls are handled has also been changed.
1113
1114 25. If an assertion condition captured any substrings, they were not passed
1115 back unless some other capturing happened later. For example, if
1116 (?(?=(a))a) was matched against "a", no capturing was returned.
1117
1118 26. When studying a pattern that contained subroutine calls or assertions,
1119 the code for finding the minimum length of a possible match was handling
1120 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1121 group 1 called group 2 while simultaneously a separate group 2 called group
1122 1). A stack overflow occurred in this case. I have fixed this by limiting
1123 the recursion depth to 10.
1124
1125 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1126 Fortmann. This supports explicit test numbers on the command line, and has
1127 argument validation and error reporting.
1128
1129 28. An instance of \X with an unlimited repeat could fail if at any point the
1130 first character it looked at was a mark character.
1131
1132 29. Some minor code refactoring concerning Unicode properties and scripts
1133 should reduce the stack requirement of match() slightly.
1134
1135 30. Added the '=' option to pcretest to check the setting of unused capturing
1136 slots at the end of the pattern, which are documented as being -1, but are
1137 not included in the return count.
1138
1139 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1140 compiled something random. Now it gives a compile-time error (as does
1141 Perl).
1142
1143 32. A *MARK encountered during the processing of a positive assertion is now
1144 recorded and passed back (compatible with Perl).
1145
1146 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1147 had alternative anchored branches, the search for a second match in a line
1148 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1149 matched the line "0102" twice. The same bug affected patterns that started
1150 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1151 twice.
1152
1153 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1154 does, and because of capturing effects, quantifying parenthesized
1155 assertions may at times be useful. Quantifiers are now allowed for
1156 parenthesized assertions.
1157
1158 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1159
1160 36. \g was being checked for fancy things in a character class, when it should
1161 just be a literal "g".
1162
1163 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1164 appearance of a nested POSIX class supersedes an apparent external class.
1165 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1166 unescaped square brackets may also appear as part of class names. For
1167 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1168 more like Perl. (But see 8.20/1 above.)
1169
1170 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1171 was because it thought it was \N{name}, which is not supported).
1172
1173 39. Add minix to OS list not supporting the -S option in pcretest.
1174
1175 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1176 cannot analyze patterns in sufficient detail to catch mutual recursions
1177 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1178 subgroup is called recursively as a subpattern for a second time at the
1179 same position in the subject string. In previous releases this might have
1180 been caught by the recursion limit, or it might have run out of stack.
1181
1182 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1183 happen only once. PCRE was, however incorrectly giving a compile time error
1184 "recursive call could loop indefinitely" because it cannot analyze the
1185 pattern in sufficient detail. The compile time test no longer happens when
1186 PCRE is compiling a conditional subpattern, but actual runaway loops are
1187 now caught at runtime (see 40 above).
1188
1189 42. It seems that Perl allows any characters other than a closing parenthesis
1190 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1191 has been changed to be the same.
1192
1193 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1194 as not to get warnings when autogen.sh is called. Also changed
1195 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1196
1197 44. To help people who use pcregrep to scan files containing exceedingly long
1198 lines, the following changes have been made:
1199
1200 (a) The default value of the buffer size parameter has been increased from
1201 8K to 20K. (The actual buffer used is three times this size.)
1202
1203 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1204 PCRE is built.
1205
1206 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1207 to be set at run time.
1208
1209 (d) Numerical values in pcregrep options can be followed by K or M, for
1210 example --buffer-size=50K.
1211
1212 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1213 given and the return code is set to 2.
1214
1215 45. Add a pointer to the latest mark to the callout data block.
1216
1217 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1218 partial match of an empty string instead of no match. This was specific to
1219 the use of ".".
1220
1221 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1222 complete match instead of a partial match. This bug was dependent on both
1223 the PCRE_UTF8 and PCRE_DOTALL options being set.
1224
1225 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1226 starting byte set, because \b was not being ignored.
1227
1228
1229 Version 8.12 15-Jan-2011
1230 ------------------------
1231
1232 1. Fixed some typos in the markup of the man pages, and wrote a script that
1233 checks for such things as part of the documentation building process.
1234
1235 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1236 --match-limit and --recursion-limit options (added for 8.11). In
1237 particular, this made one of the standard tests fail. (The integer value
1238 went into the wrong half of a long int.)
1239
1240 3. If the --colour option was given to pcregrep with -v (invert match), it
1241 did strange things, either producing crazy output, or crashing. It should,
1242 of course, ignore a request for colour when reporting lines that do not
1243 match.
1244
1245 4. Another pcregrep bug caused similar problems if --colour was specified with
1246 -M (multiline) and the pattern match finished with a line ending.
1247
1248 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1249 matched in multiline mode, the following line was shown as part of the
1250 match. This seems wrong, so I have changed it.
1251
1252 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1253 the check for further matches in the same line (so they could be coloured)
1254 to overrun the end of the current line. If another match was found, it was
1255 incorrectly shown (and then shown again when found in the next line).
1256
1257 7. If pcregrep was compiled under Windows, there was a reference to the
1258 function pcregrep_exit() before it was defined. I am assuming this was
1259 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1260 reported by a user. I've moved the definition above the reference.
1261
1262
1263 Version 8.11 10-Dec-2010
1264 ------------------------
1265
1266 1. (*THEN) was not working properly if there were untried alternatives prior
1267 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1268 backtracked to try for "b" instead of moving to the next alternative branch
1269 at the same level (in this case, to look for "c"). The Perl documentation
1270 is clear that when (*THEN) is backtracked onto, it goes to the "next
1271 alternative in the innermost enclosing group".
1272
1273 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1274 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1275 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1276 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1277 (*THEN).
1278
1279 3. If \s appeared in a character class, it removed the VT character from
1280 the class, even if it had been included by some previous item, for example
1281 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1282 of \s, but is part of the POSIX "space" class.)
1283
1284 4. A partial match never returns an empty string (because you can always
1285 match an empty string at the end of the subject); however the checking for
1286 an empty string was starting at the "start of match" point. This has been
1287 changed to the "earliest inspected character" point, because the returned
1288 data for a partial match starts at this character. This means that, for
1289 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1290 (previously it gave "no match").
1291
1292 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1293 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1294 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1295 has an implication that the given string is incomplete (because a partial
1296 match is preferred over a full match). For this reason, these items now
1297 give a partial match in this situation. [Aside: previously, the one case
1298 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1299 match rather than a full match, which was wrong by the old rules, but is
1300 now correct.]
1301
1302 6. There was a bug in the handling of #-introduced comments, recognized when
1303 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1304 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1305 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1306 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1307 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1308 places in pcre_compile().
1309
1310 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1311 comments when looking ahead for named forward references to subpatterns,
1312 the only newline sequence it recognized was NL. It now handles newlines
1313 according to the set newline convention.
1314
1315 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1316 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1317 cater for a lack of strerror(). These oversights have been fixed.
1318
1319 9. Added --match-limit and --recursion-limit to pcregrep.
1320
1321 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1322
1323 11. When the -o option was used, pcregrep was setting a return code of 1, even
1324 when matches were found, and --line-buffered was not being honoured.
1325
1326 12. Added an optional parentheses number to the -o and --only-matching options
1327 of pcregrep.
1328
1329 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1330 can match an empty string. The code to do it in pcretest and pcredemo
1331 needed fixing:
1332
1333 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1334 only one byte after an empty string match just before CRLF (this case
1335 just got forgotten; "any" and "anycrlf" were OK).
1336
1337 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1338 mode when an empty string match preceded an ASCII character followed by
1339 a non-ASCII character. (The code for advancing by one character rather
1340 than one byte was nonsense.)
1341
1342 (c) The pcredemo.c sample program did not have any code at all to handle
1343 the cases when CRLF is a valid newline sequence.
1344
1345 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1346 as a starting offset was within the subject string. There is now a new
1347 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1348 negative or greater than the length of the string. In order to test this,
1349 pcretest is extended to allow the setting of negative starting offsets.
1350
1351 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1352 starting offset points to the beginning of a UTF-8 character was
1353 unnecessarily clumsy. I tidied it up.
1354
1355 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1356 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1357
1358 17. Nobody had reported that the --include_dir option, which was added in
1359 release 7.7 should have been called --include-dir (hyphen, not underscore)
1360 for compatibility with GNU grep. I have changed it to --include-dir, but
1361 left --include_dir as an undocumented synonym, and the same for
1362 --exclude-dir, though that is not available in GNU grep, at least as of
1363 release 2.5.4.
1364
1365 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1366 characters from a string of bytes) have been redefined so as not to use
1367 loops, in order to improve performance in some environments. At the same
1368 time, I abstracted some of the common code into auxiliary macros to save
1369 repetition (this should not affect the compiled code).
1370
1371 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1372 compile-time error is now given if \c is not followed by an ASCII
1373 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1374 different, and any byte value is allowed.)
1375
1376 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1377 START_OPTIMIZE option, which is now allowed at compile time - but just
1378 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1379 to pcregrep and other applications that have no direct access to PCRE
1380 options. The new /Y option in pcretest sets this option when calling
1381 pcre_compile().
1382
1383 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1384 back references. Groups containing recursive back references were forced to
1385 be atomic by that change, but in the case of named groups, the amount of
1386 memory required was incorrectly computed, leading to "Failed: internal
1387 error: code overflow". This has been fixed.
1388
1389 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1390 pcretest.c, to avoid build problems in some Borland environments.
1391
1392
1393 Version 8.10 25-Jun-2010
1394 ------------------------
1395
1396 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1397 THEN.
1398
1399 2. (*ACCEPT) was not working when inside an atomic group.
1400
1401 3. Inside a character class, \B is treated as a literal by default, but
1402 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1403 causes the error). The code is unchanged, but I tidied the documentation.
1404
1405 4. Inside a character class, PCRE always treated \R and \X as literals,
1406 whereas Perl faults them if its -w option is set. I have changed PCRE so
1407 that it faults them when PCRE_EXTRA is set.
1408
1409 5. Added support for \N, which always matches any character other than
1410 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1411
1412 6. When compiling pcregrep with newer versions of gcc which may have
1413 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1414 declared with attribute warn_unused_result" were given. Just casting the
1415 result to (void) does not stop the warnings; a more elaborate fudge is
1416 needed. I've used a macro to implement this.
1417
1418 7. Minor change to pcretest.c to avoid a compiler warning.
1419
1420 8. Added four artifical Unicode properties to help with an option to make
1421 \s etc use properties (see next item). The new properties are: Xan
1422 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1423
1424 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1425 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1426 this option. Modified pcretest to add /W to test this facility. Added
1427 REG_UCP to make it available via the POSIX interface.
1428
1429 10. Added --line-buffered to pcregrep.
1430
1431 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1432 studied, and the match started with a letter with a code point greater than
1433 127 whose first byte was different to the first byte of the other case of
1434 the letter, the other case of this starting letter was not recognized
1435 (#976).
1436
1437 12. If a pattern that was studied started with a repeated Unicode property
1438 test, for example, \p{Nd}+, there was the theoretical possibility of
1439 setting up an incorrect bitmap of starting bytes, but fortunately it could
1440 not have actually happened in practice until change 8 above was made (it
1441 added property types that matched character-matching opcodes).
1442
1443 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1444 possible starting bytes for non-anchored patterns.
1445
1446 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1447 \R, and also a number of cases that involve Unicode properties, both
1448 explicit and implicit when PCRE_UCP is set.
1449
1450 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1451 input, it could crash or give wrong results if characters with values
1452 greater than 0xc0 were present in the subject string. (Detail: it assumed
1453 UTF-8 input when processing these items.)
1454
1455 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1456 size_t is 64-bit (#991).
1457
1458 17. Added a check for running out of memory when PCRE is compiled with
1459 --disable-stack-for-recursion (#990).
1460
1461 18. If the last data line in a file for pcretest does not have a newline on
1462 the end, a newline was missing in the output.
1463
1464 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1465 less than 128) in its various bitmaps. However, there is a facility for
1466 generating tables according to the current locale when PCRE is compiled. It
1467 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1468 characters, are recognized by isspace() and therefore were getting set in
1469 these tables, and indeed these tables seem to approximate to ISO 8859. This
1470 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1471 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1472 which of course cannot start UTF-8 characters. I have changed the code so
1473 that only real ASCII characters (less than 128) and the correct starting
1474 bytes for UTF-8 encodings are set for characters greater than 127 when in
1475 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1476 altogether.)
1477
1478 20. Added the /T option to pcretest so as to be able to run tests with non-
1479 standard character tables, thus making it possible to include the tests
1480 used for 19 above in the standard set of tests.
1481
1482 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1483 reference to a subpattern the other side of a comment that contains an
1484 opening parenthesis caused either an internal compiling error, or a
1485 reference to the wrong subpattern.
1486
1487
1488 Version 8.02 19-Mar-2010
1489 ------------------------
1490
1491 1. The Unicode data tables have been updated to Unicode 5.2.0.
1492
1493 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1494 configured.
1495
1496 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1497 original author of that file, following a query about its status.
1498
1499 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1500 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1501
1502 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1503 quantifier applied to a forward-referencing subroutine call, could compile
1504 incorrect code or give the error "internal error: previously-checked
1505 referenced subpattern not found".
1506
1507 6. Both MS Visual Studio and Symbian OS have problems with initializing
1508 variables to point to external functions. For these systems, therefore,
1509 pcre_malloc etc. are now initialized to local functions that call the
1510 relevant global functions.
1511
1512 7. There were two entries missing in the vectors called coptable and poptable
1513 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1514 I've fixed the data, and added a kludgy way of testing at compile time that
1515 the lengths are correct (equal to the number of opcodes).
1516
1517 8. Following on from 7, I added a similar kludge to check the length of the
1518 eint vector in pcreposix.c.
1519
1520 9. Error texts for pcre_compile() are held as one long string to avoid too
1521 much relocation at load time. To find a text, the string is searched,
1522 counting zeros. There was no check for running off the end of the string,
1523 which could happen if a new error number was added without updating the
1524 string.
1525
1526 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1527
1528 11. \K was not working if it appeared in an atomic group or in a group that
1529 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1530 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1531 the assertion is positive, but not if it is negative.
1532
1533 12. Change 11 fortuitously reduced the size of the stack frame used in the
1534 "match()" function of pcre_exec.c by one pointer. Forthcoming
1535 implementation of support for (*MARK) will need an extra pointer on the
1536 stack; I have reserved it now, so that the stack frame size does not
1537 decrease.
1538
1539 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1540 item in branch that calls a recursion is a subroutine call - as in the
1541 second branch in the above example - was incorrectly given the compile-
1542 time error "recursive call could loop indefinitely" because pcre_compile()
1543 was not correctly checking the subroutine for matching a non-empty string.
1544
1545 14. The checks for overrunning compiling workspace could trigger after an
1546 overrun had occurred. This is a "should never occur" error, but it can be
1547 triggered by pathological patterns such as hundreds of nested parentheses.
1548 The checks now trigger 100 bytes before the end of the workspace.
1549
1550 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1551
1552
1553 Version 8.01 19-Jan-2010
1554 ------------------------
1555
1556 1. If a pattern contained a conditional subpattern with only one branch (in
1557 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1558 computed the wrong minimum data length (which is of course zero for such
1559 subpatterns). This could cause incorrect "no match" results.
1560
1561 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1562 the pattern is reset in the first branch, pcre_compile() failed with
1563 "internal error: code overflow at offset...". This happened only when
1564 the reset was to the original external option setting. (An optimization
1565 abstracts leading options settings into an external setting, which was the
1566 cause of this.)
1567
1568 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1569 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1570 assertion pattern did not match (meaning that the assertion was true), it
1571 was incorrectly treated as false if the SKIP had been reached during the
1572 matching. This also applied to assertions used as conditions.
1573
1574 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1575 assertion subpattern, including such a pattern used as a condition,
1576 unpredictable results occurred, instead of the error return
1577 PCRE_ERROR_DFA_UITEM.
1578
1579 5. The C++ GlobalReplace function was not working like Perl for the special
1580 situation when an empty string is matched. It now does the fancy magic
1581 stuff that is necessary.
1582
1583 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1584 removed. (These were left over from very, very early versions of PCRE.)
1585
1586 7. Some cosmetic changes to the code to make life easier when compiling it
1587 as part of something else:
1588
1589 (a) Change DEBUG to PCRE_DEBUG.
1590
1591 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1592 called "current" as "current_branch", to prevent a collision with the
1593 Linux macro when compiled as a kernel module.
1594
1595 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1596 prevent a collision with the Linux macro when compiled as a kernel
1597 module.
1598
1599 8. In pcre_compile() there are some checks for integer overflows that used to
1600 cast potentially large values to (double). This has been changed to that
1601 when building, a check for int64_t is made, and if it is found, it is used
1602 instead, thus avoiding the use of floating point arithmetic. (There is no
1603 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1604 double.
1605
1606 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1607 2005 (difference between two addresses compared to an unsigned value).
1608
1609 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1610 custom one, because of the following reported problem in Windows:
1611
1612 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1613 under Win32.
1614 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1615 therefore missing the function definition.
1616 - The compiler thus generates a "C" signature for the test function.
1617 - The linker fails to find the "C" function.
1618 - PCRE fails to configure if asked to do so against libbz2.
1619
1620 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1621 messages were output:
1622
1623 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1624 rerunning libtoolize, to keep the correct libtool macros in-tree.
1625 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1626
1627 I have done both of these things.
1628
1629 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1630 most of the time, it *can* run out if it is given a pattern that contains a
1631 runaway infinite recursion. I updated the discussion in the pcrestack man
1632 page.
1633
1634 13. Now that we have gone to the x.xx style of version numbers, the minor
1635 version may start with zero. Using 08 or 09 is a bad idea because users
1636 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1637 interpreted as invalid octal numbers. I've updated the previous comment in
1638 configure.ac, and also added a check that gives an error if 08 or 09 are
1639 used.
1640
1641 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1642 causing partial matching to fail when the end of the subject matched \W
1643 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1644
1645 15. There were some discrepancies between the declarations in pcre_internal.h
1646 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1647 their definitions. The declarations used "const uschar *" and the
1648 definitions used USPTR. Even though USPTR is normally defined as "const
1649 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1650 reported that: "This difference in casting confuses some C++ compilers, for
1651 example, SunCC recognizes above declarations as different functions and
1652 generates broken code for hbpcre." I have changed the declarations to use
1653 USPTR.
1654
1655 16. GNU libtool is named differently on some systems. The autogen.sh script now
1656 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1657 (FreeBSD).
1658
1659 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1660 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1661 comment: "Figure out how to create a longlong from a string: strtoll and
1662 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1663 instance, but it only takes 2 args instead of 3!"
1664
1665 18. A subtle bug concerned with back references has been fixed by a change of
1666 specification, with a corresponding code fix. A pattern such as
1667 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1668 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1669 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1670 same bug. Such groups have to be quantified to be useful, or contained
1671 inside another quantified group. (If there's no repetition, the reference
1672 can never match.) The problem arises because, having left the group and
1673 moved on to the rest of the pattern, a later failure that backtracks into
1674 the group uses the captured value from the final iteration of the group
1675 rather than the correct earlier one. I have fixed this in PCRE by forcing
1676 any group that contains a reference to itself to be an atomic group; that
1677 is, there cannot be any backtracking into it once it has completed. This is
1678 similar to recursive and subroutine calls.
1679
1680
1681 Version 8.00 19-Oct-09
1682 ----------------------
1683
1684 1. The table for translating pcre_compile() error codes into POSIX error codes
1685 was out-of-date, and there was no check on the pcre_compile() error code
1686 being within the table. This could lead to an OK return being given in
1687 error.
1688
1689 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1690 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1691 in a Windows environment.
1692
1693 3. The pcregrep --count option prints the count for each file even when it is
1694 zero, as does GNU grep. However, pcregrep was also printing all files when
1695 --files-with-matches was added. Now, when both options are given, it prints
1696 counts only for those files that have at least one match. (GNU grep just
1697 prints the file name in this circumstance, but including the count seems
1698 more useful - otherwise, why use --count?) Also ensured that the
1699 combination -clh just lists non-zero counts, with no names.
1700
1701 4. The long form of the pcregrep -F option was incorrectly implemented as
1702 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1703 but it seems right to fix it, and I didn't think it was worth preserving
1704 the old behaviour.
1705
1706 5. The command line items --regex=pattern and --regexp=pattern were not
1707 recognized by pcregrep, which required --regex pattern or --regexp pattern
1708 (with a space rather than an '='). The man page documented the '=' forms,
1709 which are compatible with GNU grep; these now work.
1710
1711 6. No libpcreposix.pc file was created for pkg-config; there was just
1712 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1713
1714 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1715 when UCP support is not needed, by modifying the Python script that
1716 generates it from Unicode data files. This should not matter if the module
1717 is correctly used as a library, but I received one complaint about 50K of
1718 unwanted data. My guess is that the person linked everything into his
1719 program rather than using a library. Anyway, it does no harm.
1720
1721 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1722 was a minimum greater than 1 for a wide character in a possessive
1723 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1724 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1725 character. Chaos in the form of incorrect output or a compiling loop could
1726 result.
1727
1728 9. The restrictions on what a pattern can contain when partial matching is
1729 requested for pcre_exec() have been removed. All patterns can now be
1730 partially matched by this function. In addition, if there are at least two
1731 slots in the offset vector, the offset of the earliest inspected character
1732 for the match and the offset of the end of the subject are set in them when
1733 PCRE_ERROR_PARTIAL is returned.
1734
1735 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1736 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1737 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1738 and may be more useful for multi-segment matching.
1739
1740 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1741 used to be given if ever the end of the subject was reached; now it is
1742 given only if matching could not proceed because another character was
1743 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1744 string "Z", which now yields "no match" instead of "partial match". In the
1745 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1746 final character ended with (*FAIL).
1747
1748 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1749 if the pattern had a "must contain" character that was already found in the
1750 earlier partial match, unless partial matching was again requested. For
1751 example, with the pattern /dog.(body)?/, the "must contain" character is
1752 "g". If the first part-match was for the string "dog", restarting with
1753 "sbody" failed. This bug has been fixed.
1754
1755 13. The string returned by pcre_dfa_exec() after a partial match has been
1756 changed so that it starts at the first inspected character rather than the
1757 first character of the match. This makes a difference only if the pattern
1758 starts with a lookbehind assertion or \b or \B (\K is not supported by
1759 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1760 matching functions compatible, and I think it's the right thing to do.
1761
1762 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1763 so that the demonstration program is easily available in environments where
1764 PCRE has not been installed from source.
1765
1766 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1767 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1768 library.
1769
1770 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1771 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1772 is not the first non-POSIX option to be added. Clearly some people find
1773 these options useful.
1774
1775 17. If a caller to the POSIX matching function regexec() passes a non-zero
1776 value for nmatch with a NULL value for pmatch, the value of
1777 nmatch is forced to zero.
1778
1779 18. RunGrepTest did not have a test for the availability of the -u option of
1780 the diff command, as RunTest does. It now checks in the same way as
1781 RunTest, and also checks for the -b option.
1782
1783 19. If an odd number of negated classes containing just a single character
1784 interposed, within parentheses, between a forward reference to a named
1785 subpattern and the definition of the subpattern, compilation crashed with
1786 an internal error, complaining that it could not find the referenced
1787 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1788 [The bug was that it was starting one character too far in when skipping
1789 over the character class, thus treating the ] as data rather than
1790 terminating the class. This meant it could skip too much.]
1791
1792 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1793 /g option in pcretest when the pattern contains \K, which makes it possible
1794 to have an empty string match not at the start, even when the pattern is
1795 anchored. Updated pcretest and pcredemo to use this option.
1796
1797 21. If the maximum number of capturing subpatterns in a recursion was greater
1798 than the maximum at the outer level, the higher number was returned, but
1799 with unset values at the outer level. The correct (outer level) value is
1800 now given.
1801
1802 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1803 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1804 make it do so. The string so far is captured, making this feature
1805 compatible with Perl.
1806
1807 23. The tests have been re-organized, adding tests 11 and 12, to make it
1808 possible to check the Perl 5.10 features against Perl 5.10.
1809
1810 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1811 pattern matches a fixed length string. PCRE did not allow this; now it
1812 does. Neither allows recursion.
1813
1814 25. I finally figured out how to implement a request to provide the minimum
1815 length of subject string that was needed in order to match a given pattern.
1816 (It was back references and recursion that I had previously got hung up
1817 on.) This code has now been added to pcre_study(); it finds a lower bound
1818 to the length of subject needed. It is not necessarily the greatest lower
1819 bound, but using it to avoid searching strings that are too short does give
1820 some useful speed-ups. The value is available to calling programs via
1821 pcre_fullinfo().
1822
1823 26. While implementing 25, I discovered to my embarrassment that pcretest had
1824 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1825 study optimizations had never been tested with that matching function.
1826 Oops. What is worse, even when it was passed study data, there was a bug in
1827 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1828 were also very few tests of studied patterns with pcre_dfa_exec().
1829
1830 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1831 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1832 on the other side of the coin, they are no longer allowed to have different
1833 names, because these cannot be distinguished in PCRE, and this has caused
1834 confusion. (This is a difference from Perl.)
1835
1836 28. When duplicate subpattern names are present (necessarily with different
1837 numbers, as required by 27 above), and a test is made by name in a
1838 conditional pattern, either for a subpattern having been matched, or for
1839 recursion in such a pattern, all the associated numbered subpatterns are
1840 tested, and the overall condition is true if the condition is true for any
1841 one of them. This is the way Perl works, and is also more like the way
1842 testing by number works.
1843
1844
1845 Version 7.9 11-Apr-09
1846 ---------------------
1847
1848 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1849 (pcretest), all targets were linked against these libraries. This included
1850 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1851 libraries. This caused unwanted dependencies to be created. This problem
1852 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1853 pcretest is linked with readline.
1854
1855 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1856 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1857 moved outside it again, because FALSE and TRUE are already defined in AIX,
1858 but BOOL is not.
1859
1860 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1861 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1862
1863 4. The pcregrep documentation said spaces were inserted as well as colons (or
1864 hyphens) following file names and line numbers when outputting matching
1865 lines. This is not true; no spaces are inserted. I have also clarified the
1866 wording for the --colour (or --color) option.
1867
1868 5. In pcregrep, when --colour was used with -o, the list of matching strings
1869 was not coloured; this is different to GNU grep, so I have changed it to be
1870 the same.
1871
1872 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1873 each matching line was coloured. Now it goes on to look for further matches
1874 of any of the test patterns, which is the same behaviour as GNU grep.
1875
1876 7. A pattern that could match an empty string could cause pcregrep to loop; it
1877 doesn't make sense to accept an empty string match in pcregrep, so I have
1878 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1879 seems to be how GNU grep behaves. [But see later change 40 for release
1880 8.33.]
1881
1882 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1883 start or after a newline", because the conditional assertion was not being
1884 correctly handled. The rule now is that both the assertion and what follows
1885 in the first alternative must satisfy the test.
1886
1887 9. If auto-callout was enabled in a pattern with a conditional group whose
1888 condition was an assertion, PCRE could crash during matching, both with
1889 pcre_exec() and pcre_dfa_exec().
1890
1891 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1892 used for matching.
1893
1894 11. Unicode property support in character classes was not working for
1895 characters (bytes) greater than 127 when not in UTF-8 mode.
1896
1897 12. Added the -M command line option to pcretest.
1898
1899 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1900
1901 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1902
1903 16. Added comments and documentation about mis-use of no_arg in the C++
1904 wrapper.
1905
1906 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1907 from Martin Jerabek that uses macro names for all relevant character and
1908 string constants.
1909
1910 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1911 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1912 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1913 these, but not everybody uses configure.
1914
1915 19. A conditional group that had only one branch was not being correctly
1916 recognized as an item that could match an empty string. This meant that an
1917 enclosing group might also not be so recognized, causing infinite looping
1918 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1919 with the subject "ab", where knowledge that the repeated group can match
1920 nothing is needed in order to break the loop.
1921
1922 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1923 exec(), but without supplying a callout function, matching went wrong.
1924
1925 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1926 leak if the size of the offset vector was greater than 30. When the vector
1927 is smaller, the saved offsets during recursion go onto a local stack
1928 vector, but for larger vectors malloc() is used. It was failing to free
1929 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1930 error, in fact).
1931
1932 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1933 heapframe that is used only when UTF-8 support is enabled. This caused no
1934 problem, but was untidy.
1935
1936 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1937 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1938 included within another project.
1939
1940 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1941 slightly modified by me:
1942
1943 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1944 not building pcregrep.
1945
1946 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1947 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1948
1949 25. Forward references, both numeric and by name, in patterns that made use of
1950 duplicate group numbers, could behave incorrectly or give incorrect errors,
1951 because when scanning forward to find the reference group, PCRE was not
1952 taking into account the duplicate group numbers. A pattern such as
1953 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1954
1955 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1956 the feature of a custom pointer more persuasive (as requested by a user).
1957
1958 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1959 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1960 pre-defined.
1961
1962 28. Added support for (*UTF8) at the start of a pattern.
1963
1964 29. Arrange for flags added by the "release type" setting in CMake to be shown
1965 in the configuration summary.
1966
1967
1968 Version 7.8 05-Sep-08
1969 ---------------------
1970
1971 1. Replaced UCP searching code with optimized version as implemented for Ad
1972 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1973 stage table and inline lookup instead of a function, giving speed ups of 2
1974 to 5 times on some simple patterns that I tested. Permission was given to
1975 distribute the MultiStage2.py script that generates the tables (it's not in
1976 the tarball, but is in the Subversion repository).
1977
1978 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1979 scripts.
1980
1981 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1982 a group with a zero qualifier. The result of the study could be incorrect,
1983 or the function might crash, depending on the pattern.
1984
1985 4. Caseless matching was not working for non-ASCII characters in back
1986 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1987 It now works when Unicode Property Support is available.
1988
1989 5. In pcretest, an escape such as \x{de} in the data was always generating
1990 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1991 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1992 truncation.
1993
1994 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1995
1996 7. Added two (int) casts to pcregrep when printing the difference of two
1997 pointers, in case they are 64-bit values.
1998
1999 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2000 test 2 if it fails.
2001
2002 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2003 and a #define of that name to empty if it is not externally set. This is to
2004 allow users of MSVC to set it if necessary.
2005
2006 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2007 the convenience functions in the pcre_get.c source file.
2008
2009 11. An option change at the start of a pattern that had top-level alternatives
2010 could cause overwriting and/or a crash. This command provoked a crash in
2011 some environments:
2012
2013 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2014
2015 This potential security problem was recorded as CVE-2008-2371.
2016
2017 12. For a pattern where the match had to start at the beginning or immediately
2018 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2019 pcre_dfa_exec() could read past the end of the passed subject if there was
2020 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2021 pcretest so that it places the subject at the end of its malloc-ed buffer.
2022
2023 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2024 exec() might read past the end of the data buffer in UTF-8 mode.
2025
2026 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2027 the data contained the byte 0x85 as part of a UTF-8 character within its
2028 first line. This applied both to normal and DFA matching.
2029
2030 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2031 /^[^d]*?$/8 failed to match "abc".
2032
2033 16. Added a missing copyright notice to pcrecpp_internal.h.
2034
2035 17. Make it more clear in the documentation that values returned from
2036 pcre_exec() in ovector are byte offsets, not character counts.
2037
2038 18. Tidied a few places to stop certain compilers from issuing warnings.
2039
2040 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2041 supplied by Stefan Weber. I made a further small update for 7.8 because
2042 there is a change of source arrangements: the pcre_searchfuncs.c module is
2043 replaced by pcre_ucd.c.
2044
2045
2046 Version 7.7 07-May-08
2047 ---------------------
2048
2049 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2050 a string to a long long, pretend we don't even have a long long." This is
2051 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2052
2053 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2054 pre-7.6 versions, which defined a global no_arg variable instead of putting
2055 it in the RE class. (See also #8 below.)
2056
2057 3. Remove a line of dead code, identified by coverity and reported by Nuno
2058 Lopes.
2059
2060 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2061
2062 (1) The include/exclude patterns were being applied to the whole pathnames
2063 of files, instead of just to the final components.
2064
2065 (2) If there was more than one level of directory, the subdirectories were
2066 skipped unless they satisfied the include/exclude conditions. This is
2067 inconsistent with GNU grep (and could even be seen as contrary to the
2068 pcregrep specification - which I improved to make it absolutely clear).
2069 The action now is always to scan all levels of directory, and just
2070 apply the include/exclude patterns to regular files.
2071
2072 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2073 --exclude_dir in the tests to avoid scanning .svn directories.
2074
2075 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2076 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2077 doesn't support NULs in patterns.
2078
2079 7. Added some missing "const"s to declarations of static tables in
2080 pcre_compile.c and pcre_dfa_exec.c.
2081
2082 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2083 caused by fix #2 above. (Subsequently also a second patch to fix the
2084 first patch. And a third patch - this was a messy problem.)
2085
2086 9. Applied Craig's patch to remove the use of push_back().
2087
2088 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2089 matching function regexec().
2090
2091 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2092 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2093 references. PCRE supports relative numbers with this syntax (I don't think
2094 Oniguruma does).
2095
2096 12. Previously, a group with a zero repeat such as (...){0} was completely
2097 omitted from the compiled regex. However, this means that if the group
2098 was called as a subroutine from elsewhere in the pattern, things went wrong
2099 (an internal error was given). Such groups are now left in the compiled
2100 pattern, with a new opcode that causes them to be skipped at execution
2101 time.
2102
2103 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2104 to the way PCRE behaves:
2105
2106 (a) A lone ] character is dis-allowed (Perl treats it as data).
2107
2108 (b) A back reference to an unmatched subpattern matches an empty string
2109 (Perl fails the current match path).
2110
2111 (c) A data ] in a character class must be notated as \] because if the
2112 first data character in a class is ], it defines an empty class. (In
2113 Perl it is not possible to have an empty class.) The empty class []
2114 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2115 The negative empty class [^] matches any one character, independently
2116 of the DOTALL setting.
2117
2118 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2119 non-existent subpattern following a character class starting with ']' and
2120 containing () gave an internal compiling error instead of "reference to
2121 non-existent subpattern". Fortunately, when the pattern did exist, the
2122 compiled code was correct. (When scanning forwards to check for the
2123 existence of the subpattern, it was treating the data ']' as terminating
2124 the class, so got the count wrong. When actually compiling, the reference
2125 was subsequently set up correctly.)
2126
2127 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2128 it was being rejected as not supported by pcre_dfa_exec(), even though
2129 other assertions are supported. I have made pcre_dfa_exec() support
2130 (*FAIL).
2131
2132 16. The implementation of 13c above involved the invention of a new opcode,
2133 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2134 cannot be changed at match time, I realized I could make a small
2135 improvement to matching performance by compiling OP_ALLANY instead of
2136 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2137 on the OP_ANY path.
2138
2139 17. Compiling pcretest on Windows with readline support failed without the
2140 following two fixes: (1) Make the unistd.h include conditional on
2141 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2142
2143 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2144 ncurses library to be included for pcretest when ReadLine support is
2145 requested, but also to allow for it to be overridden. This patch came from
2146 Daniel Bergström.
2147
2148 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2149 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2150 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2151 spotting this.
2152
2153
2154 Version 7.6 28-Jan-08
2155 ---------------------
2156
2157 1. A character class containing a very large number of characters with
2158 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2159 overflow.
2160
2161 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2162 HAVE_LONG_LONG is not defined.
2163
2164 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2165 bring them up to date and include new features. This patch includes:
2166
2167 - Fixed PH's badly added libz and libbz2 support.
2168 - Fixed a problem with static linking.
2169 - Added pcredemo. [But later removed - see 7 below.]
2170 - Fixed dftables problem and added an option.
2171 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2172 HAVE_LONG_LONG.
2173 - Added readline support for pcretest.
2174 - Added an listing of the option settings after cmake has run.
2175
2176 4. A user submitted a patch to Makefile that makes it easy to create
2177 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2178 Makefile.am that cause it to include this special target, without
2179 affecting anything else. Note that the same mingw target plus all
2180 the other distribution libraries and programs are now supported
2181 when configuring with CMake (see 6 below) instead of with
2182 Configure/Make.
2183
2184 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2185 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2186 exported in the Windows port". It has not yet been confirmed that the patch
2187 solves the problem, but it does no harm.
2188
2189 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2190 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2191 with CMake, and also correct the comment about stack recursion.
2192
2193 7. Remove the automatic building of pcredemo from the ./configure system and
2194 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2195 of a program that users should build themselves after PCRE is installed, so
2196 building it automatically is not really right. What is more, it gave
2197 trouble in some build environments.
2198
2199 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2200
2201
2202 Version 7.5 10-Jan-08
2203 ---------------------
2204
2205 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2206 values in parens when parsing an RE using the C++ wrapper."
2207
2208 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2209 Characters greater than 255 were excluded from the class instead of being
2210 included.
2211
2212 3. The same bug as (2) above applied to negated POSIX classes such as
2213 [:^space:].
2214
2215 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2216 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2217 I have changed it.
2218
2219 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2220 first named subpattern) and a construct such as (?&a) would reference the
2221 first named subpattern whose name started with "a" (in other words, the
2222 length check was missing). Both these problems are fixed. "Subpattern name
2223 expected" is now given for (?&) (a zero-length name), and this patch also
2224 makes it give the same error for \k'' (previously it complained that that
2225 was a reference to a non-existent subpattern).
2226
2227 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2228 this is right because (?- can be followed by option settings as well as by
2229 digits. I have, however, made the messages clearer.
2230
2231 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2232 than the number used in the conditional) now cause a compile-time error.
2233 This is actually not compatible with Perl, which accepts such patterns, but
2234 treats the conditional as always being FALSE (as PCRE used to), but it
2235 seems to me that giving a diagnostic is better.
2236
2237 8. Change "alphameric" to the more common word "alphanumeric" in comments
2238 and messages.
2239
2240 9. Fix two occurrences of "backslash" in comments that should have been
2241 "backspace".
2242
2243 10. Remove two redundant lines of code that can never be obeyed (their function
2244 was moved elsewhere).
2245
2246 11. The program that makes PCRE's Unicode character property table had a bug
2247 which caused it to generate incorrect table entries for sequences of
2248 characters that have the same character type, but are in different scripts.
2249 It amalgamated them into a single range, with the script of the first of
2250 them. In other words, some characters were in the wrong script. There were
2251 thirteen such cases, affecting characters in the following ranges:
2252
2253 U+002b0 - U+002c1
2254 U+0060c - U+0060d
2255 U+0061e - U+00612
2256 U+0064b - U+0065e
2257 U+0074d - U+0076d
2258 U+01800 - U+01805
2259 U+01d00 - U+01d77
2260 U+01d9b - U+01dbf
2261 U+0200b - U+0200f
2262 U+030fc - U+030fe
2263 U+03260 - U+0327f
2264 U+0fb46 - U+0fbb1
2265 U+10450 - U+1049d
2266
2267 12. The -o option (show only the matching part of a line) for pcregrep was not
2268 compatible with GNU grep in that, if there was more than one match in a
2269 line, it showed only the first of them. It now behaves in the same way as
2270 GNU grep.
2271
2272 13. If the -o and -v options were combined for pcregrep, it printed a blank
2273 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2274 does the same. The return code can be used to tell if there were any
2275 non-matching lines.
2276
2277 14. Added --file-offsets and --line-offsets to pcregrep.
2278
2279 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2280 infinitely looping recursion. The bug was that positive lookaheads were not
2281 being skipped when checking for a possible empty match (negative lookaheads
2282 and both kinds of lookbehind were skipped).
2283
2284 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2285 inclusion of <windows.h> to before rather than after the definition of
2286 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2287
2288 17. Specifying a possessive quantifier with a specific limit for a Unicode
2289 character property caused pcre_compile() to compile bad code, which led at
2290 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2291 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2292 caused the error; without that there was no problem.
2293
2294 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2295
2296 19. Added --enable-pcretest-libreadline.
2297
2298 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2299 RE::GlobalReplace(). As a result, the number of replacements returned was
2300 double what it should be. I removed one of the increments, but Craig sent a
2301 later patch that removed the other one (the right fix) and added unit tests
2302 that check the return values (which was not done before).
2303
2304 21. Several CMake things:
2305
2306 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2307 the names libpcre and libpcreposix, not just pcre and pcreposix.
2308
2309 (2) The above change means that pcretest and pcregrep are now correctly
2310 linked with the newly-built libraries, not previously installed ones.
2311
2312 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2313
2314 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2315 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2316 UTF-8 newline character). The key issue is that the pattern starts .*;
2317 this means that the match must be either at the beginning, or after a
2318 newline. The bug was in the code for advancing after a failed match and
2319 checking that the new position followed a newline. It was not taking
2320 account of UTF-8 characters correctly.
2321
2322 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2323 character classes. PCRE was not treating the sequence [:...:] as a
2324 character class unless the ... were all letters. Perl, however, seems to
2325 allow any characters between [: and :], though of course it rejects as
2326 unknown any "names" that contain non-letters, because all the known class
2327 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2328 for example, whereas PCRE did not - it did not recognize a POSIX character
2329 class. This seemed a bit dangerous, so the code has been changed to be
2330 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2331 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2332 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2333 Perl does, and where it didn't before.
2334
2335 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2336 Windows environments %n is disabled by default.
2337
2338
2339 Version 7.4 21-Sep-07
2340 ---------------------
2341
2342 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2343 means that a class such as [\s] counted as "explicit reference to CR or
2344 LF". That isn't really right - the whole point of the change was to try to
2345 help when there was an actual mention of one of the two characters. So now
2346 the change happens only if \r or \n (or a literal CR or LF) character is
2347 encountered.
2348
2349 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2350 of both had grown to the point where there were only 3 bits left.
2351 Fortunately, there was spare space in the data structure, and so I have
2352 moved the internal flags into a new 16-bit field to free up more option
2353 bits.
2354
2355 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2356 but did not set the internal JCHANGED flag - either of these is enough to
2357 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2358 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2359 start sets both bits.
2360
2361 4. Added options (at build time, compile time, exec time) to change \R from
2362 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2363
2364 5. doc/pcresyntax.html was missing from the distribution.
2365
2366 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2367 compatibility, even though it is no longer used.
2368
2369 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2370 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2371 windows.h file is present (where different names are used). [This was
2372 reversed later after testing - see 16 below.]
2373
2374 8. Changed all #include <config.h> to #include "config.h". There were also
2375 some further <pcre.h> cases that I changed to "pcre.h".
2376
2377 9. When pcregrep was used with the --colour option, it missed the line ending
2378 sequence off the lines that it output.
2379
2380 10. It was pointed out to me that arrays of string pointers cause lots of
2381 relocations when a shared library is dynamically loaded. A technique of
2382 using a single long string with a table of offsets can drastically reduce
2383 these. I have refactored PCRE in four places to do this. The result is
2384 dramatic:
2385
2386 Originally: 290
2387 After changing UCP table: 187
2388 After changing error message table: 43
2389 After changing table of "verbs" 36
2390 After changing table of Posix names 22
2391
2392 Thanks to the folks working on Gregex for glib for this insight.
2393
2394 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2395 unicode-properties was also set.
2396
2397 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2398
2399 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2400 checked only for CRLF.
2401
2402 14. Added casts to pcretest.c to avoid compiler warnings.
2403
2404 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2405
2406 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2407 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2408 entirely. This removes changes made in 7 above.
2409
2410 17. The CMake files have been updated, and there is now more information about
2411 building with CMake in the NON-UNIX-USE document.
2412
2413
2414 Version 7.3 28-Aug-07
2415 ---------------------
2416
2417 1. In the rejigging of the build system that eventually resulted in 7.1, the
2418 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2419 brackets there is not right, since it causes compilers to look for an
2420 installed pcre.h, not the version that is in the source that is being
2421 compiled (which of course may be different). I have changed it back to:
2422
2423 #include "pcre.h"
2424
2425 I have a vague recollection that the change was concerned with compiling in
2426 different directories, but in the new build system, that is taken care of
2427 by the VPATH setting the Makefile.
2428
2429 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2430 when the subject happened to end in the byte 0x85 (e.g. if the last
2431 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2432 characters but of course it shouldn't be taken as a newline when it is part
2433 of another character. The bug was that, for an unlimited repeat of . in
2434 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2435 characters when looking for a newline.
2436
2437 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2438
2439 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2440 in debug output.
2441
2442 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2443 long printing in the pcrecpp unittest when running under MinGW.
2444
2445 6. ESC_K was left out of the EBCDIC table.
2446
2447 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2448 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2449 limit also applies to "virtual nesting" when a pattern is recursive, and in
2450 this case 1000 isn't so big. I have been able to remove this limit at the
2451 expense of backing off one optimization in certain circumstances. Normally,
2452 when pcre_exec() would call its internal match() function recursively and
2453 immediately return the result unconditionally, it uses a "tail recursion"
2454 feature to save stack. However, when a subpattern that can match an empty
2455 string has an unlimited repetition quantifier, it no longer makes this
2456 optimization. That gives it a stack frame in which to save the data for
2457 checking that an empty string has been matched. Previously this was taken
2458 from the 1000-entry workspace that had been reserved. So now there is no
2459 explicit limit, but more stack is used.
2460
2461 8. Applied Daniel's patches to solve problems with the import/export magic
2462 syntax that is required for Windows, and which was going wrong for the
2463 pcreposix and pcrecpp parts of the library. These were overlooked when this
2464 problem was solved for the main library.
2465
2466 9. There were some crude static tests to avoid integer overflow when computing
2467 the size of patterns that contain repeated groups with explicit upper
2468 limits. As the maximum quantifier is 65535, the maximum group length was
2469 set at 30,000 so that the product of these two numbers did not overflow a
2470 32-bit integer. However, it turns out that people want to use groups that
2471 are longer than 30,000 bytes (though not repeat them that many times).
2472 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2473 made it possible to implement the integer overflow checks in a much more
2474 dynamic way, which I have now done. The artificial limitation on group
2475 length has been removed - we now have only the limit on the total length of
2476 the compiled pattern, which depends on the LINK_SIZE setting.
2477
2478 10. Fixed a bug in the documentation for get/copy named substring when
2479 duplicate names are permitted. If none of the named substrings are set, the
2480 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2481 empty string.
2482
2483 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2484 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2485 because the ] is interpreted as the first data character and the
2486 terminating ] is not found. PCRE has been made compatible with Perl in this
2487 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2488 cause memory overwriting.
2489
2490 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2491 string has been matched (to stop an infinite loop). It was not recognizing
2492 a conditional subpattern that could match an empty string if that
2493 subpattern was within another subpattern. For example, it looped when
2494 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2495 condition was not nested. This bug has been fixed.
2496
2497 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2498 past the start of the subject in the presence of bytes with the top bit
2499 set, for example "\x8aBCD".
2500
2501 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2502 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2503
2504 14. Optimized (?!) to (*FAIL).
2505
2506 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2507 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2508 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2509 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2510 does: it's just the validity check that is more restrictive.
2511
2512 16. Inserted checks for integer overflows during escape sequence (backslash)
2513 processing, and also fixed erroneous offset values for syntax errors during
2514 backslash processing.
2515
2516 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2517 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2518
2519 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2520 caused an overrun.
2521
2522 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2523 something other than just ASCII characters) inside a group that had an
2524 unlimited repeat caused a loop at compile time (while checking to see
2525 whether the group could match an empty string).
2526
2527 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2528 [\P{Any}] did so. (Error in the code for printing property names.)
2529
2530 21. An orphan \E inside a character class could cause a crash.
2531
2532 22. A repeated capturing bracket such as (A)? could cause a wild memory
2533 reference during compilation.
2534
2535 23. There are several functions in pcre_compile() that scan along a compiled
2536 expression for various reasons (e.g. to see if it's fixed length for look
2537 behind). There were bugs in these functions when a repeated \p or \P was
2538 present in the pattern. These operators have additional parameters compared
2539 with \d, etc, and these were not being taken into account when moving along
2540 the compiled data. Specifically:
2541
2542 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2543 length.
2544
2545 (b) An item such as \pL+ within a repeated group could cause crashes or
2546 loops.
2547
2548 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2549 "reference to non-existent subpattern" error.
2550
2551 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2552
2553 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2554 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2555
2556 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2557
2558 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2559 character were causing crashes (broken optimization).
2560
2561 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2562 \p or \P) caused a compile-time loop.
2563
2564 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2565 break. For example, the unstudied pattern [\r\n]A does not match the string
2566 "\r\nA" because change 7.0/46 below moves the current point on by two
2567 characters after failing to match at the start. However, the pattern \nA
2568 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2569 the same is true. There doesn't seem any very clean way out of this, but
2570 what I have chosen to do makes the common cases work: PCRE now takes note
2571 of whether there can be an explicit match for \r or \n anywhere in the
2572 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2573 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2574 pattern has explicit CR or LF references.
2575
2576 29. Added (*CR) etc for changing newline setting at start of pattern.
2577
2578
2579 Version 7.2 19-Jun-07
2580 ---------------------
2581
2582 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2583 which is apparently normally available under Windows.
2584
2585 2. Re-jig the pcregrep tests with different newline settings in an attempt
2586 to make them independent of the local environment's newline setting.
2587
2588 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2589
2590 4. Some of the "internals" tests were previously cut out when the link size
2591 was not 2, because the output contained actual offsets. The recent new
2592 "Z" feature of pcretest means that these can be cut out, making the tests
2593 usable with all link sizes.
2594
2595 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2596 stack recursion. This gives a massive performance boost under BSD, but just
2597 a small improvement under Linux. However, it saves one field in the frame
2598 in all cases.
2599
2600 6. Added more features from the forthcoming Perl 5.10:
2601
2602 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2603 recursion call. It refers to the nth most recently opened parentheses.
2604
2605 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2606 to be opened parentheses.
2607
2608 (c) Conditions that refer to capturing parentheses can be specified
2609 relatively, for example, (?(-2)... or (?(+3)...
2610
2611 (d) \K resets the start of the current match so that everything before
2612 is not part of it.
2613
2614 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2615
2616 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2617 reference syntax.
2618
2619 (g) (?| introduces a group in which the numbering of parentheses in each
2620 alternative starts with the same number.
2621
2622 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2623
2624 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2625 PCRE_INFO_JCHANGED.
2626
2627 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2628 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2629 for detecting groups that can match an empty string.
2630
2631 9. A pattern with a very large number of alternatives (more than several
2632 hundred) was running out of internal workspace during the pre-compile
2633 phase, where pcre_compile() figures out how much memory will be needed. A
2634 bit of new cunning has reduced the workspace needed for groups with
2635 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2636 workspace instead of running out of the 4096 that are available.
2637
2638 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2639
2640 11. Applied patch from Google to remove an optimization that didn't quite work.
2641 The report of the bug said:
2642
2643 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2644 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2645 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2646
2647 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2648 it matched the wrong number of bytes.
2649
2650
2651 Version 7.1 24-Apr-07
2652 ---------------------
2653
2654 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2655 that is more "standard", making use of automake and other Autotools. There
2656 is some re-arrangement of the files and adjustment of comments consequent
2657 on this.
2658
2659 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2660 for recursive directory scanning broke on some systems because the files
2661 are not scanned in any specific order and on different systems the order
2662 was different. A call to "sort" has been inserted into RunGrepTest for the
2663 approprate test as a short-term fix. In the longer term there may be an
2664 alternative.
2665
2666 3. I had an email from Eric Raymond about problems translating some of PCRE's
2667 man pages to HTML (despite the fact that I distribute HTML pages, some
2668 people do their own conversions for various reasons). The problems
2669 concerned the use of low-level troff macros .br and .in. I have therefore
2670 removed all such uses from the man pages (some were redundant, some could
2671 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2672 HTML has been updated to handle .nf/.fi and to complain if it encounters
2673 .br or .in.
2674
2675 4. Updated comments in configure.ac that get placed in config.h.in and also
2676 arranged for config.h to be included in the distribution, with the name
2677 config.h.generic, for the benefit of those who have to compile without
2678 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2679
2680 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2681 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2682 makevp.bat for the current PCRE, using the additional files
2683 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2684
2685 6. A Windows user reported a minor discrepancy with test 2, which turned out
2686 to be caused by a trailing space on an input line that had got lost in his
2687 copy. The trailing space was an accident, so I've just removed it.
2688
2689 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2690 that is needed.
2691
2692 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2693 as "const" (a) because they are and (b) because it helps the PHP
2694 maintainers who have recently made a script to detect big data structures
2695 in the php code that should be moved to the .rodata section. I remembered
2696 to update Builducptable as well, so it won't revert if ucptable.h is ever
2697 re-created.
2698
2699 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2700 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2701 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2702 support is not required. This saves 1.5-2K of code, which is important in
2703 some applications.
2704
2705 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2706 so as not to refer to the tables, even though these functions will never be
2707 called when UTF-8 support is disabled. Otherwise there are problems with a
2708 shared library.
2709
2710 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2711
2712 (a) It was defining its arguments as char * instead of void *.
2713
2714 (b) It was assuming that all moves were upwards in memory; this was true
2715 a long time ago when I wrote it, but is no longer the case.
2716
2717 The emulated memove() is provided for those environments that have neither
2718 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2719 is clearly not the case, as these two bugs were recently reported.
2720
2721 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2722 and Detrail to create the HTML documentation, the .txt form of the man
2723 pages, and it removes trailing spaces from listed files. It also creates
2724 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2725 case, it wraps all the #defines with #ifndefs. This script should be run
2726 before "make dist".
2727
2728 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2729 with Unicode property support.
2730
2731 (a) For a maximizing quantifier, if the two different cases of the
2732 character were of different lengths in their UTF-8 codings (there are
2733 some cases like this - I found 11), and the matching function had to
2734 back up over a mixture of the two cases, it incorrectly assumed they
2735 were both the same length.
2736
2737 (b) When PCRE was configured to use the heap rather than the stack for
2738 recursion during matching, it was not correctly preserving the data for
2739 the other case of a UTF-8 character when checking ahead for a match
2740 while processing a minimizing repeat. If the check also involved
2741 matching a wide character, but failed, corruption could cause an
2742 erroneous result when trying to check for a repeat of the original
2743 character.
2744
2745 13. Some tidying changes to the testing mechanism:
2746
2747 (a) The RunTest script now detects the internal link size and whether there
2748 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2749 values substituted by "configure". (The RunGrepTest script already did
2750 this for UTF-8.) The configure.ac script no longer substitutes the
2751 relevant variables.
2752
2753 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2754 with length and offset values. This means that the output is different
2755 for different internal link sizes. Test 2 is skipped for link sizes
2756 other than 2 because of this, bypassing the problem. Unfortunately,
2757 there was also a test in test 3 (the locale tests) that used /B and
2758 failed for link sizes other than 2. Rather than cut the whole test out,
2759 I have added a new /Z option to pcretest that replaces the length and
2760 offset values with spaces. This is now used to make test 3 independent
2761 of link size. (Test 2 will be tidied up later.)
2762
2763 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2764 segmentation fault instead of returning the appropriate error message.
2765
2766 15. In multiline mode when the newline sequence was set to "any", the pattern
2767 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2768 This doesn't seem right; it now treats the CRLF combination as the line
2769 ending, and so does not match in that case. It's only a pattern such as ^$
2770 that would hit this one: something like ^ABC$ would have failed after \r
2771 and then tried again after \r\n.
2772
2773 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2774 in an attempt to make files that differ only in their line terminators
2775 compare equal. This works on Linux.
2776
2777 17. Under certain error circumstances pcregrep might try to free random memory
2778 as it exited. This is now fixed, thanks to valgrind.
2779
2780 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2781 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2782 was because its rules for how to advance for /g after matching an empty
2783 string at the end of a line did not allow for this case. They now check for
2784 it specially.
2785
2786 20. pcretest is supposed to handle patterns and data of any length, by
2787 extending its buffers when necessary. It was getting this wrong when the
2788 buffer for a data line had to be extended.
2789
2790 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2791 CRLF as a newline sequence.
2792
2793 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2794 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2795 I have nevertheless tidied it up.
2796
2797 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2798
2799 24. Added a man page for pcre-config.
2800
2801
2802 Version 7.0 19-Dec-06
2803 ---------------------
2804
2805 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2806 moving to gcc 4.1.1.
2807
2808 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2809 sys/time.h, which is documented as needed for this function. It doesn't
2810 seem to matter on Linux, but it showed up on some releases of OS X.
2811
2812 3. It seems that there are systems where bytes whose values are greater than
2813 127 match isprint() in the "C" locale. The "C" locale should be the
2814 default when a C program starts up. In most systems, only ASCII printing
2815 characters match isprint(). This difference caused the output from pcretest
2816 to vary, making some of the tests fail. I have changed pcretest so that:
2817
2818 (a) When it is outputting text in the compiled version of a pattern, bytes
2819 other than 32-126 are always shown as hex escapes.
2820
2821 (b) When it is outputting text that is a matched part of a subject string,
2822 it does the same, unless a different locale has been set for the match
2823 (using the /L modifier). In this case, it uses isprint() to decide.
2824
2825 4. Fixed a major bug that caused incorrect computation of the amount of memory
2826 required for a compiled pattern when options that changed within the
2827 pattern affected the logic of the preliminary scan that determines the
2828 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2829 that the computed length was too small. The symptoms of this bug were
2830 either the PCRE error "internal error: code overflow" from pcre_compile(),
2831 or a glibc crash with a message such as "pcretest: free(): invalid next
2832 size (fast)". Examples of patterns that provoked this bug (shown in
2833 pcretest format) are:
2834
2835 /(?-x: )/x
2836 /(?x)(?-x: \s*#\s*)/
2837 /((?i)[\x{c0}])/8
2838 /(?i:[\x{c0}])/8
2839
2840 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2841 is now done differently.
2842
2843 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2844 wrapper classes; (b) implement a new function in the C++ scanner that is
2845 more efficient than the old way of doing things because it avoids levels of
2846 recursion in the regex matching; (c) add a paragraph to the documentation
2847 for the FullMatch() function.
2848
2849 6. The escape sequence \n was being treated as whatever was defined as
2850 "newline". Not only was this contrary to the documentation, which states
2851 that \n is character 10 (hex 0A), but it also went horribly wrong when
2852 "newline" was defined as CRLF. This has been fixed.
2853
2854 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2855 was being set to -1 for the "end of line" case (supposedly a value that no
2856 character can have). Though this value is never used (the check for end of
2857 line is "zero bytes in current character"), it caused compiler complaints.
2858 I've changed it to 0xffffffff.
2859
2860 8. In pcre_version.c, the version string was being built by a sequence of
2861 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2862 string (as it is for production releases) called a macro with an empty
2863 argument. The C standard says the result of this is undefined. The gcc
2864 compiler treats it as an empty string (which was what was wanted) but it is
2865 reported that Visual C gives an error. The source has been hacked around to
2866 avoid this problem.
2867
2868 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2869 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2870 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2871 of them did).
2872
2873 10. Originally, pcretest opened its input and output without "b"; then I was
2874 told that "b" was needed in some environments, so it was added for release
2875 5.0 to both the input and output. (It makes no difference on Unix-like
2876 systems.) Later I was told that it is wrong for the input on Windows. I've
2877 now abstracted the modes into two macros, to make it easier to fiddle with
2878 them, and removed "b" from the input mode under Windows.
2879
2880 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2881
2882 12. Added -help and --help to pcretest as an official way of being reminded
2883 of the options.
2884
2885 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2886 and pcrecpp.cc because they annoy compilers at high warning levels.
2887
2888 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2889
2890 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2891 scripts are not C programs :-) and which was not noticed because it works
2892 on Linux.
2893
2894 16. pcretest is supposed to handle any length of pattern and data line (as one
2895 line or as a continued sequence of lines) by extending its input buffer if
2896 necessary. This feature was broken for very long pattern lines, leading to
2897 a string of junk being passed to pcre_compile() if the pattern was longer
2898 than about 50K.
2899
2900 17. I have done a major re-factoring of the way pcre_compile() computes the
2901 amount of memory needed for a compiled pattern. Previously, there was code
2902 that made a preliminary scan of the pattern in order to do this. That was
2903 OK when PCRE was new, but as the facilities have expanded, it has become
2904 harder and harder to keep it in step with the real compile phase, and there
2905 have been a number of bugs (see for example, 4 above). I have now found a
2906 cunning way of running the real compile function in a "fake" mode that
2907 enables it to compute how much memory it would need, while actually only
2908 ever using a few hundred bytes of working memory and without too many
2909 tests of the mode. This should make future maintenance and development
2910 easier. A side effect of this work is that the limit of 200 on the nesting
2911 depth of parentheses has been removed (though this was never a serious
2912 limitation, I suspect). However, there is a downside: pcre_compile() now
2913 runs more slowly than before (30% or more, depending on the pattern). I
2914 hope this isn't a big issue. There is no effect on runtime performance.
2915
2916 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2917 newline (only possible for the last line of a file) and it was a
2918 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2919
2920 19. Added additional timing features to pcretest. (1) The -tm option now times
2921 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2922 separate command line item, by a number that specifies the number of
2923 repeats to use when timing. The default is 50000; this gives better
2924 precision, but takes uncomfortably long for very large patterns.
2925
2926 20. Extended pcre_study() to be more clever in cases where a branch of a
2927 subpattern has no definite first character. For example, (a*|b*)[cd] would
2928 previously give no result from pcre_study(). Now it recognizes that the
2929 first character must be a, b, c, or d.
2930
2931 21. There was an incorrect error "recursive call could loop indefinitely" if
2932 a subpattern (or the entire pattern) that was being tested for matching an
2933 empty string contained only one non-empty item after a nested subpattern.
2934 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2935 incorrectly, because the \d was being skipped in the check.
2936
2937 22. The pcretest program now has a new pattern option /B and a command line
2938 option -b, which is equivalent to adding /B to every pattern. This causes
2939 it to show the compiled bytecode, without the additional information that
2940 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2941 is the same as /B/I).
2942
2943 23. A new optimization is now able automatically to treat some sequences such
2944 as a*b as a*+b. More specifically, if something simple (such as a character
2945 or a simple class like \d) has an unlimited quantifier, and is followed by
2946 something that cannot possibly match the quantified thing, the quantifier
2947 is automatically "possessified".
2948
2949 24. A recursive reference to a subpattern whose number was greater than 39
2950 went wrong under certain circumstances in UTF-8 mode. This bug could also
2951 have affected the operation of pcre_study().
2952
2953 25. Realized that a little bit of performance could be had by replacing
2954 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2955
2956 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2957
2958 27. Possessive quantifiers such as a++ were previously implemented by turning
2959 them into atomic groups such as ($>a+). Now they have their own opcodes,
2960 which improves performance. This includes the automatically created ones
2961 from 23 above.
2962
2963 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2964 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2965 the first matched character to be a colon. This applied both to named and
2966 numbered groups.
2967
2968 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2969
2970 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2971 building PCRE on Windows easier, so I have included it in the distribution.
2972
2973 31. There is now a check in pcretest against a ridiculously large number being
2974 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2975 loop, the loop is abandoned.
2976
2977 32. Forward references to subpatterns in conditions such as (?(2)...) where
2978 subpattern 2 is defined later cause pcre_compile() to search forwards in
2979 the pattern for the relevant set of parentheses. This search went wrong
2980 when there were unescaped parentheses in a character class, parentheses
2981 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2982
2983 33. "Subroutine" calls and backreferences were previously restricted to
2984 referencing subpatterns earlier in the regex. This restriction has now
2985 been removed.
2986
2987 34. Added a number of extra features that are going to be in Perl 5.10. On the
2988 whole, these are just syntactic alternatives for features that PCRE had
2989 previously implemented using the Python syntax or my own invention. The
2990 other formats are all retained for compatibility.
2991
2992 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2993 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2994 also .NET compatible.
2995
2996 (b) A recursion or subroutine call to a named group can now be defined as
2997 (?&name) as well as (?P>name).
2998
2999 (c) A backreference to a named group can now be defined as \k<name> or
3000 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3001 5.10, are also .NET compatible.
3002
3003 (d) A conditional reference to a named group can now use the syntax
3004 (?(<name>) or (?('name') as well as (?(name).
3005
3006 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3007 groups (named and numbered) that are never evaluated inline, but can be
3008 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3009 is always false. There may be only one alternative in such a group.
3010
3011 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3012 as the simple (?(R). The condition is true only if the most recent
3013 recursion is that of the given number or name. It does not search out
3014 through the entire recursion stack.
3015
3016 (g) The escape \gN or \g{N} has been added, where N is a positive or
3017 negative number, specifying an absolute or relative reference.
3018
3019 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3020 some "unreachable code" warnings.
3021
3022 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3023 things, this adds five new scripts.
3024
3025 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3026 There were also incompatibilities regarding the handling of \Q..\E inside
3027 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3028 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3029
3030 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3031 matches an empty string, and forcibly breaks the loop. There were bugs in
3032 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3033 against aaaa the result was just "a" rather than "aaaa", for example. Two
3034 separate and independent bugs (that affected different cases) have been
3035 fixed.
3036
3037 39. Refactored the code to abolish the use of different opcodes for small
3038 capturing bracket numbers. This is a tidy that I avoided doing when I
3039 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3040 The new approach is not only tidier, it makes it possible to reduce the
3041 memory needed to fix the previous bug (38).
3042
3043 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3044 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3045 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3046 mode.
3047
3048 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3049 report.
3050
3051 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3052 copy construction and assignment in the C++ wrapper.
3053
3054 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3055 couple of bugs that could have given wrong results in the "--newline=crlf"
3056 case.
3057
3058 44. Added a number of casts and did some reorganization of signed/unsigned int
3059 variables following suggestions from Dair Grant. Also renamed the variable
3060 "this" as "item" because it is a C++ keyword.
3061
3062 45. Arranged for dftables to add
3063
3064 #include "pcre_internal.h"
3065
3066 to pcre_chartables.c because without it, gcc 4.x may remove the array
3067 definition from the final binary if PCRE is built into a static library and
3068 dead code stripping is activated.
3069
3070 46. For an unanchored pattern, if a match attempt fails at the start of a
3071 newline sequence, and the newline setting is CRLF or ANY, and the next two
3072 characters are CRLF, advance by two characters instead of one.
3073
3074
3075 Version 6.7 04-Jul-06
3076 ---------------------
3077
3078 1. In order to handle tests when input lines are enormously long, pcretest has
3079 been re-factored so that it automatically extends its buffers when
3080 necessary. The code is crude, but this _is_ just a test program. The
3081 default size has been increased from 32K to 50K.
3082
3083 2. The code in pcre_study() was using the value of the re argument before
3084 testing it for NULL. (Of course, in any sensible call of the function, it
3085 won't be NULL.)
3086
3087 3. The memmove() emulation function in pcre_internal.h, which is used on
3088 systems that lack both memmove() and bcopy() - that is, hardly ever -
3089 was missing a "static" storage class specifier.
3090
3091 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3092 containing an extended class (one that cannot be represented by a bitmap
3093 because it contains high-valued characters or Unicode property items, e.g.
3094 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3095 pattern, but PCRE should not loop if you do not (it no longer does).
3096 [Detail: two cases were found: (a) a repeated subpattern containing an
3097 extended class; (b) a recursive reference to a subpattern that followed a
3098 previous extended class. It wasn't skipping over the extended class
3099 correctly when UTF-8 mode was not set.]
3100
3101 5. A negated single-character class was not being recognized as fixed-length
3102 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3103 compile error "lookbehind assertion is not fixed length".
3104
3105 6. The RunPerlTest auxiliary script was showing an unexpected difference
3106 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3107 write a Perl script that can interpret lines of an input file either as
3108 byte characters or as UTF-8, which is what "perltest" was being required to
3109 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3110 can't do is switch easily at run time between having the "use utf8;" pragma
3111 or not. In the end, I fudged it by using the RunPerlTest script to insert
3112 "use utf8;" explicitly for the UTF-8 tests.
3113
3114 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3115 the end of the subject string, contrary to the documentation and to what
3116 Perl does. This was true of both matching functions. Now it matches only at
3117 the start of the subject and immediately after *internal* newlines.
3118
3119 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3120 a pointer to an int instead of a pointer to an unsigned long int. This
3121 caused problems on 64-bit systems.
3122
3123 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3124 instance of the 'standard' template library not being so standard".
3125
3126 10. There was no check on the number of named subpatterns nor the maximum
3127 length of a subpattern name. The product of these values is used to compute
3128 the size of the memory block for a compiled pattern. By supplying a very
3129 long subpattern name and a large number of named subpatterns, the size
3130 computation could be caused to overflow. This is now prevented by limiting
3131 the length of names to 32 characters, and the number of named subpatterns
3132 to 10,000.
3133
3134 11. Subpatterns that are repeated with specific counts have to be replicated in
3135 the compiled pattern. The size of memory for this was computed from the
3136 length of the subpattern and the repeat count. The latter is limited to
3137 65535, but there was no limit on the former, meaning that integer overflow
3138 could in principle occur. The compiled length of a repeated subpattern is
3139 now limited to 30,000 bytes in order to prevent this.
3140
3141 12. Added the optional facility to have named substrings with the same name.
3142
3143 13. Added the ability to use a named substring as a condition, using the
3144 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3145 are numbers (not recommended). Forward references are permitted.
3146
3147 14. Added forward references in named backreferences (if you see what I mean).
3148
3149 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3150 pattern could run off the end of the subject. For example, the pattern
3151 "(?s)(.{1,5})"8 did this with the subject "ab".
3152
3153 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3154 PCRE_CASELESS was set when matching characters that were quantified with ?
3155 or *.
3156
3157 17. A character class other than a single negated character that had a minimum
3158 but no maximum quantifier - for example [ab]{6,} - was not handled
3159 correctly by pce_dfa_exec(). It would match only one character.
3160
3161 18. A valid (though odd) pattern that looked like a POSIX character
3162 class but used an invalid character after [ (for example [[,abc,]]) caused
3163 pcre_compile() to give the error "Failed: internal error: code overflow" or
3164 in some cases to crash with a glibc free() error. This could even happen if
3165 the pattern terminated after [[ but there just happened to be a sequence of
3166 letters, a binary zero, and a closing ] in the memory that followed.
3167
3168 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3169 over the years. Originally (before any Unicode support), just the bottom 8
3170 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3171 output from "man perlunicode" includes this:
3172
3173 The regular expression compiler produces polymorphic opcodes. That
3174 is, the pattern adapts to the data and automatically switches to
3175 the Unicode character scheme when presented with Unicode data--or
3176 instead uses a traditional byte scheme when presented with byte
3177 data.
3178
3179 Sadly, a wide octal escape does not cause a switch, and in a string with
3180 no other multibyte characters, these octal escapes are treated as before.
3181 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3182 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3183 Unicode string.
3184
3185 I have not perpetrated such confusion in PCRE. Up till now, it took just
3186 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3187 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3188 translate to the appropriate multibyte character.
3189
3190 29. Applied some refactoring to reduce the number of warnings from Microsoft
3191 and Borland compilers. This has included removing the fudge introduced
3192 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3193 a warning about an unused variable.
3194
3195 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3196 characters since release 4.0, because Perl (from release 5.004) does not.
3197 [Or at least, is documented not to: some releases seem to be in conflict
3198 with the documentation.] However, when a pattern was studied with
3199 pcre_study() and all its branches started with \s, PCRE still included VT
3200 as a possible starting character. Of course, this did no harm; it just
3201 caused an unnecessary match attempt.
3202
3203 22. Removed a now-redundant internal flag bit that recorded the fact that case
3204 dependency changed within the pattern. This was once needed for "required
3205 byte" processing, but is no longer used. This recovers a now-scarce options
3206 bit. Also moved the least significant internal flag bit to the most-
3207 significant bit of the word, which was not previously used (hangover from
3208 the days when it was an int rather than a uint) to free up another bit for
3209 the future.
3210
3211 23. Added support for CRLF line endings as well as CR and LF. As well as the
3212 default being selectable at build time, it can now be changed at runtime
3213 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3214 specify that it is scanning data with non-default line endings.
3215
3216 24. Changed the definition of CXXLINK to make it agree with the definition of
3217 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3218
3219 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3220 recursions. This makes a big different to stack usage for some patterns.
3221
3222 26. If a subpattern containing a named recursion or subroutine reference such
3223 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3224 the space required for the compiled pattern went wrong and gave too small a
3225 value. Depending on the environment, this could lead to "Failed: internal
3226 error: code overflow at offset 49" or "glibc detected double free or
3227 corruption" errors.
3228
3229 27. Applied patches from Google (a) to support the new newline modes and (b) to
3230 advance over multibyte UTF-8 characters in GlobalReplace.
3231
3232 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3233 difference for some implementation of PCRE in some Windows version.
3234
3235 29. Added some extra testing facilities to pcretest:
3236
3237 \q<number> in a data line sets the "match limit" value
3238 \Q<number> in a data line sets the "match recursion limt" value
3239 -S <number> sets the stack size, where <number> is in megabytes
3240
3241 The -S option isn't available for Windows.
3242
3243
3244 Version 6.6 06-Feb-06
3245 ---------------------
3246
3247 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3248 in pcreposix.h. I have copied the definition from pcre.h.
3249
3250 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3251 because pcre.h is no longer a built file.
3252
3253 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3254 not normally included in the compiled code.
3255
3256
3257 Version 6.5 01-Feb-06
3258 ---------------------
3259
3260 1. When using the partial match feature with pcre_dfa_exec(), it was not
3261 anchoring the second and subsequent partial matches at the new starting
3262 point. This could lead to incorrect results. For example, with the pattern
3263 /1234/, partially matching against "123" and then "a4" gave a match.
3264
3265 2. Changes to pcregrep:
3266
3267 (a) All non-match returns from pcre_exec() were being treated as failures
3268 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3269 error message is output. Some extra information is given for the
3270 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3271 probably the only errors that are likely to be caused by users (by
3272 specifying a regex that has nested indefinite repeats, for instance).
3273 If there are more than 20 of these errors, pcregrep is abandoned.
3274
3275 (b) A binary zero was treated as data while matching, but terminated the
3276 output line if it was written out. This has been fixed: binary zeroes
3277 are now no different to any other data bytes.
3278
3279 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3280 used to set a locale for matching. The --locale=xxxx long option has
3281 been added (no short equivalent) to specify a locale explicitly on the
3282 pcregrep command, overriding the environment variables.
3283
3284 (d) When -B was used with -n, some line numbers in the output were one less
3285 than they should have been.
3286
3287 (e) Added the -o (--only-matching) option.
3288
3289 (f) If -A or -C was used with -c (count only), some lines of context were
3290 accidentally printed for the final match.
3291
3292 (g) Added the -H (--with-filename) option.
3293
3294 (h) The combination of options -rh failed to suppress file names for files
3295 that were found from directory arguments.
3296
3297 (i) Added the -D (--devices) and -d (--directories) options.
3298
3299 (j) Added the -F (--fixed-strings) option.
3300
3301 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3302
3303 (l) Added the --colo(u)r option.
3304
3305 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3306 is not present by default.
3307
3308 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3309 items such as (?R) or (?1), when the recursion could match a number of
3310 alternatives. If it matched one of the alternatives, but subsequently,
3311 outside the recursion, there was a failure, the code tried to back up into
3312 the recursion. However, because of the way PCRE is implemented, this is not
3313 possible, and the result was an incorrect result from the match.
3314
3315 In order to prevent this happening, the specification of recursion has
3316 been changed so that all such subpatterns are automatically treated as
3317 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3318
3319 4. I had overlooked the fact that, in some locales, there are characters for
3320 which isalpha() is true but neither isupper() nor islower() are true. In
3321 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3322 and ordfeminine) are like this. This affected the treatment of \w and \W
3323 when they appeared in character classes, but not when they appeared outside
3324 a character class. The bit map for "word" characters is now created
3325 separately from the results of isalnum() instead of just taking it from the
3326 upper, lower, and digit maps. (Plus the underscore character, of course.)
3327
3328 5. The above bug also affected the handling of POSIX character classes such as
3329 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3330 permanent tables. Instead, the bit maps for such a class were previously
3331 created as the appropriate unions of the upper, lower, and digit bitmaps.
3332 Now they are created by subtraction from the [[:word:]] class, which has
3333 its own bitmap.
3334
3335 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3336 It is created by subtracting the vertical space characters (\x09, \x0a,
3337 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3338 subtraction was done in the overall bitmap for a character class, meaning
3339 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3340 be recognized. This bug has been fixed.
3341
3342 7. Patches from the folks at Google:
3343
3344 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3345 real life, but is still worth protecting against".
3346
3347 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3348 regular expressions".
3349
3350 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3351 have it.
3352
3353 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3354 "configure" and the latter not, in order to fix a problem somebody had
3355 with compiling the Arg class on HP-UX.
3356
3357 (e) Improve the error-handling of the C++ wrapper a little bit.
3358
3359 (f) New tests for checking recursion limiting.
3360
3361 8. The pcre_memmove() function, which is used only if the environment does not
3362 have a standard memmove() function (and is therefore rarely compiled),
3363 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3364 returning a result (though PCRE never actually uses the result).
3365
3366 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3367 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3368 returned instead of calling malloc() with an overflowing number that would
3369 most likely cause subsequent chaos.
3370
3371 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3372
3373 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3374 with this option is matched, the nmatch and pmatch options of regexec() are
3375 ignored.
3376
3377 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3378 provided in case anyone wants to the the POSIX interface with UTF-8
3379 strings.
3380
3381 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3382 C++ linking (needed for some HP-UX environments).
3383
3384 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3385 (unused parameter) and in the pcre_printint() function (omitted "default"
3386 switch label when the default is to do nothing).
3387
3388 15. Added some code to make it possible, when PCRE is compiled as a C++
3389 library, to replace subject pointers for pcre_exec() with a smart pointer
3390 class, thus making it possible to process discontinuous strings.
3391
3392 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3393 much the same function. They were added by different people who were trying
3394 to make PCRE easy to compile on non-Unix systems. It has been suggested
3395 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3396 for compiling on Windows systems. I have therefore replaced it with
3397 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3398 defaults to "extern" for C or "extern C" for C++, which works fine on
3399 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3400 SCOPE with something explicit in config.h. In addition:
3401
3402 (a) pcreposix.h still had just "extern" instead of either of these macros;
3403 I have replaced it with PCRE_DATA_SCOPE.
3404
3405 (b) Functions such as _pcre_xclass(), which are internal to the library,
3406 but external in the C sense, all had PCRE_EXPORT in their definitions.
3407 This is apparently wrong for the Windows case, so I have removed it.
3408 (It makes no difference on Unix-like systems.)
3409
3410 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3411 of recursive calls to match(). This is different to MATCH_LIMIT because
3412 that limits the total number of calls to match(), not all of which increase
3413 the depth of recursion. Limiting the recursion depth limits the amount of
3414 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3415 when PCRE is compiled, and changed at run time. A patch from Google adds
3416 this functionality to the C++ interface.
3417
3418 18. Changes to the handling of Unicode character properties:
3419
3420 (a) Updated the table to Unicode 4.1.0.
3421
3422 (b) Recognize characters that are not in the table as "Cn" (undefined).
3423
3424 (c) I revised the way the table is implemented to a much improved format
3425 which includes recognition of ranges. It now supports the ranges that
3426 are defined in UnicodeData.txt, and it also amalgamates other
3427 characters into ranges. This has reduced the number of entries in the
3428 table from around 16,000 to around 3,000, thus reducing its size
3429 considerably. I realized I did not need to use a tree structure after
3430 all - a binary chop search is just as efficient. Having reduced the
3431 number of entries, I extended their size from 6 bytes to 8 bytes to
3432 allow for more data.
3433
3434 (d) Added support for Unicode script names via properties such as \p{Han}.
3435
3436 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3437 matching that character.
3438
3439 20. When matching a repeated Unicode property with a minimum greater than zero,
3440 (for example \pL{2,}), PCRE could look past the end of the subject if it
3441 reached it while seeking the minimum number of characters. This could
3442 happen only if some of the characters were more than one byte long, because
3443 there is a check for at least the minimum number of bytes.
3444
3445 21. Refactored the implementation of \p and \P so as to be more general, to
3446 allow for more different types of property in future. This has changed the
3447 compiled form incompatibly. Anybody with saved compiled patterns that use
3448 \p or \P will have to recompile them.
3449
3450 22. Added "Any" and "L&" to the supported property types.
3451
3452 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3453 but give a compile time error if the value is greater than 0xff.
3454
3455 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3456 accidentally not being installed or uninstalled.
3457
3458 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3459 made were to insert the current release number. This seemed silly, because
3460 it made things harder for people building PCRE on systems that don't run
3461 "configure". I have turned pcre.h into a distributed file, no longer built
3462 by "configure", with the version identification directly included. There is
3463 no longer a pcre.h.in file.
3464
3465 However, this change necessitated a change to the pcre-config script as
3466 well. It is built from pcre-config.in, and one of the substitutions was the
3467 release number. I have updated configure.ac so that ./configure now finds
3468 the release number by grepping pcre.h.
3469
3470 26. Added the ability to run the tests under valgrind.
3471
3472
3473 Version 6.4 05-Sep-05
3474 ---------------------
3475
3476 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3477 "--" to be printed when multiple files were scanned, even when none of the
3478 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3479 consider it to be a bug, and have restored the previous behaviour.
3480
3481 2. A couple of code tidies to get rid of compiler warnings.
3482
3483 3. The pcretest program used to cheat by referring to symbols in the library
3484 whose names begin with _pcre_. These are internal symbols that are not
3485 really supposed to be visible externally, and in some environments it is
3486 possible to suppress them. The cheating is now confined to including
3487 certain files from the library's source, which is a bit cleaner.
3488
3489 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3490 file's purpose clearer.
3491
3492 5. Reorganized pcre_ucp_findchar().
3493
3494
3495 Version 6.3 15-Aug-05
3496 ---------------------
3497
3498 1. The file libpcre.pc.in did not have general read permission in the tarball.
3499
3500 2. There were some problems when building without C++ support:
3501
3502 (a) If C++ support was not built, "make install" and "make test" still
3503 tried to test it.
3504
3505 (b) There were problems when the value of CXX was explicitly set. Some
3506 changes have been made to try to fix these, and ...
3507
3508 (c) --disable-cpp can now be used to explicitly disable C++ support.
3509
3510 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3511 backslash in a target when C++ was disabled. This confuses some
3512 versions of "make", apparently. Using an intermediate variable solves
3513 this. (Same for CPP_LOBJ.)
3514
3515 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3516 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3517 necessary on certain architectures.
3518
3519 4. Added a setting of -export-symbols-regex to the link command to remove
3520 those symbols that are exported in the C sense, but actually are local
3521 within the library, and not documented. Their names all begin with
3522 "_pcre_". This is not a perfect job, because (a) we have to except some
3523 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3524 available (and never for static libraries). I have made a note to try to
3525 find a way round (a) in the future.
3526
3527
3528 Version 6.2 01-Aug-05
3529 ---------------------
3530
3531 1. There was no test for integer overflow of quantifier values. A construction
3532 such as {1111111111111111} would give undefined results. What is worse, if
3533 a minimum quantifier for a parenthesized subpattern overflowed and became
3534 negative, the calculation of the memory size went wrong. This could have
3535 led to memory overwriting.
3536
3537 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3538
3539 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3540 operating environments where this matters.
3541
3542 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3543 PCRE options from within the C++ wrapper.
3544
3545 5. Named capturing subpatterns were not being correctly counted when a pattern
3546 was compiled. This caused two problems: (a) If there were more than 100
3547 such subpatterns, the calculation of the memory needed for the whole
3548 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3549 back references of the form \12, where the number was greater than 9, were
3550 not recognized as back references, even though there were sufficient
3551 previous subpatterns.
3552
3553 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3554 versions of gcc, e.g. 2.95.4.
3555
3556
3557 Version 6.1 21-Jun-05
3558 ---------------------
3559
3560 1. There was one reference to the variable "posix" in pcretest.c that was not
3561 surrounded by "#if !defined NOPOSIX".
3562
3563 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3564 the cross-check on the old pcre_info() function, for the benefit of the
3565 cut-down version of PCRE that is currently imported into Exim.
3566
3567 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3568 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3569 patterns that start (?i) but I don't think that matters. The use of (?i) is
3570 just an example; this all applies to the other options as well.
3571
3572 4. Since libtool seems to echo the compile commands it is issuing, the output
3573 from "make" can be reduced a bit by putting "@" in front of each libtool
3574 compile command.
3575
3576 5. Patch from the folks at Google for configure.in to be a bit more thorough
3577 in checking for a suitable C++ installation before trying to compile the
3578 C++ stuff. This should fix a reported problem when a compiler was present,
3579 but no suitable headers.
3580
3581 6. The man pages all had just "PCRE" as their title. I have changed them to
3582 be the relevant file name. I have also arranged that these names are
3583 retained in the file doc/pcre.txt, which is a concatenation in text format
3584 of all the man pages except the little individual ones for each function.
3585
3586 7. The NON-UNIX-USE file had not been updated for the different set of source
3587 files that come with release 6. I also added a few comments about the C++
3588 wrapper.
3589
3590
3591 Version 6.0 07-Jun-05
3592 ---------------------
3593
3594 1. Some minor internal re-organization to help with my DFA experiments.
3595
3596 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3597 didn't matter for the library itself when fully configured, but did matter
3598 when compiling without UCP support, or within Exim, where the ucp files are
3599 not imported.
3600
3601 3. Refactoring of the library code to split up the various functions into
3602 different source modules. The addition of the new DFA matching code (see
3603 below) to a single monolithic source would have made it really too
3604 unwieldy, quite apart from causing all the code to be include in a
3605 statically linked application, when only some functions are used. This is
3606 relevant even without the DFA addition now that patterns can be compiled in
3607 one application and matched in another.
3608
3609 The downside of splitting up is that there have to be some external
3610 functions and data tables that are used internally in different modules of
3611 the library but which are not part of the API. These have all had their
3612 names changed to start with "_pcre_" so that they are unlikely to clash
3613 with other external names.
3614
3615 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3616 a different (DFA) algorithm. Although it is slower than the original
3617 function, it does have some advantages for certain types of matching
3618 problem.
3619
3620 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3621 including restarting after a partial match.
3622
3623 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3624 defined when compiling for Windows was sent to me. I have put it into the
3625 code, though I have no means of testing or verifying it.
3626
3627 7. Added the pcre_refcount() auxiliary function.
3628
3629 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3630 match before or at the first newline in the subject string. In pcretest,
3631 the /f option on a pattern can be used to set this.
3632
3633 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3634 would behave wrongly. This has been present in PCRE since release 4.0.
3635
3636 10. A number of changes to the pcregrep command:
3637
3638 (a) Refactored how -x works; insert ^(...)$ instead of setting
3639 PCRE_ANCHORED and checking the length, in preparation for adding
3640 something similar for -w.
3641
3642 (b) Added the -w (match as a word) option.
3643
3644 (c) Refactored the way lines are read and buffered so as to have more
3645 than one at a time available.
3646
3647 (d) Implemented a pcregrep test script.
3648
3649 (e) Added the -M (multiline match) option. This allows patterns to match
3650 over several lines of the subject. The buffering ensures that at least
3651 8K, or the rest of the document (whichever is the shorter) is available
3652 for matching (and similarly the previous 8K for lookbehind assertions).
3653
3654 (f) Changed the --help output so that it now says
3655
3656 -w, --word-regex(p)
3657
3658 instead of two lines, one with "regex" and the other with "regexp"
3659 because that confused at least one person since the short forms are the
3660 same. (This required a bit of code, as the output is generated
3661 automatically from a table. It wasn't just a text change.)
3662
3663 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3664 option but starts with a hyphen. Could be a pattern or a path name
3665 starting with a hyphen, for instance.
3666
3667 (h) "-" can be given as a file name to represent stdin.
3668
3669 (i) When file names are being printed, "(standard input)" is used for
3670 the standard input, for compatibility with GNU grep. Previously
3671 "<stdin>" was used.
3672
3673 (j) The option --label=xxx can be used to supply a name to be used for
3674 stdin when file names are being printed. There is no short form.
3675
3676 (k) Re-factored the options decoding logic because we are going to add
3677 two more options that take data. Such options can now be given in four
3678 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3679
3680 (l) Added the -A, -B, and -C options for requesting that lines of context
3681 around matches be printed.
3682
3683 (m) Added the -L option to print the names of files that do not contain
3684 any matching lines, that is, the complement of -l.
3685
3686 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3687 continue to scan other files.
3688
3689 (o) The -s option was incorrectly implemented. For compatibility with other
3690 greps, it now suppresses the error message for a non-existent or non-
3691 accessible file (but not the return code). There is a new option called
3692 -q that suppresses the output of matching lines, which was what -s was
3693 previously doing.
3694
3695 (p) Added --include and --exclude options to specify files for inclusion
3696 and exclusion when recursing.
3697
3698 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3699 Hopefully, it now does.
3700
3701 12. Missing cast in pcre_study().
3702
3703 13. Added an "uninstall" target to the makefile.
3704
3705 14. Replaced "extern" in the function prototypes in Makefile.in with
3706 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3707 world, but is set differently for Windows.
3708
3709 15. Added a second compiling function called pcre_compile2(). The only
3710 difference is that it has an extra argument, which is a pointer to an
3711 integer error code. When there is a compile-time failure, this is set
3712 non-zero, in addition to the error test pointer being set to point to an
3713 error message. The new argument may be NULL if no error number is required
3714 (but then you may as well call pcre_compile(), which is now just a
3715 wrapper). This facility is provided because some applications need a
3716 numeric error indication, but it has also enabled me to tidy up the way
3717 compile-time errors are handled in the POSIX wrapper.
3718
3719 16. Added VPATH=.libs to the makefile; this should help when building with one
3720 prefix path and installing with another. (Or so I'm told by someone who
3721 knows more about this stuff than I do.)
3722
3723 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3724 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3725 match everything, including newlines. This is not POSIX-compatible, but
3726 somebody wanted the feature. From pcretest it can be activated by using
3727 both the P and the s flags.
3728
3729 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3730
3731 19. libpcre.pc was being incorrectly installed as executable.
3732
3733 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3734 it now also looks for '\r' so that it will work unmodified on Windows.
3735
3736 21. Added Google's contributed C++ wrapper to the distribution.
3737
3738 22. Added some untidy missing memory free() calls in pcretest, to keep
3739 Electric Fence happy when testing.
3740
3741
3742
3743 Version 5.0 13-Sep-04
3744 ---------------------
3745
3746 1. Internal change: literal characters are no longer packed up into items
3747 containing multiple characters in a single byte-string. Each character
3748 is now matched using a separate opcode. However, there may be more than one
3749 byte in the character in UTF-8 mode.
3750
3751 2. The pcre_callout_block structure has two new fields: pattern_position and
3752 next_item_length. These contain the offset in the pattern to the next match
3753 item, and its length, respectively.
3754
3755 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3756 insertion of callouts before each pattern item. Added the /C option to
3757 pcretest to make use of this.
3758
3759 4. On the advice of a Windows user, the lines
3760
3761 #if defined(_WIN32) || defined(WIN32)
3762 _setmode( _fileno( stdout ), 0x8000 );
3763 #endif /* defined(_WIN32) || defined(WIN32) */
3764
3765 have been added to the source of pcretest. This apparently does useful
3766 magic in relation to line terminators.
3767
3768 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3769 for the benefit of those environments where the "b" makes a difference.
3770
3771 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3772 to know about it. I have put a hack into configure.in that adds in code
3773 to set GCC=yes if CC=icc. This seems to end up at a point in the
3774 generated configure script that is early enough to affect the setting of
3775 compiler options, which is what is needed, but I have no means of testing
3776 whether it really works. (The user who reported this had patched the
3777 generated configure script, which of course I cannot do.)
3778
3779 LATER: After change 22 below (new libtool files), the configure script
3780 seems to know about icc (and also ecc). Therefore, I have commented out
3781 this hack in configure.in.
3782
3783 7. Added support for pkg-config (2 patches were sent in).
3784
3785 8. Negated POSIX character classes that used a combination of internal tables
3786 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3787 [[:^ascii]]. Typically, they would match almost any characters. The other
3788 POSIX classes were not broken in this way.
3789
3790 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3791 to find the match, as PCRE was deluded into thinking that the match had to
3792 start at the start point or following a newline. The same bug applied to
3793 patterns with negative forward assertions or any backward assertions
3794 preceding ".*" at the start, unless the pattern required a fixed first
3795 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3796
3797 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3798 starting at the last subject character, bytes beyond the end of the subject
3799 string were read.
3800
3801 11. Renamed the variable "class" as "classbits" to make life easier for C++
3802 users. (Previously there was a macro definition, but it apparently wasn't
3803 enough.)
3804
3805 12. Added the new field "tables" to the extra data so that tables can be passed
3806 in at exec time, or the internal tables can be re-selected. This allows
3807 a compiled regex to be saved and re-used at a later time by a different
3808 program that might have everything at different addresses.
3809
3810 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3811 -R library as well as a -L library.
3812
3813 14. The debugging options of pcretest (-d on the command line or D on a
3814 pattern) showed incorrect output for anything following an extended class
3815 that contained multibyte characters and which was followed by a quantifier.
3816
3817 15. Added optional support for general category Unicode character properties
3818 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3819 support. It adds about 90K to the size of the library. The meanings of the
3820 inbuilt class escapes such as \d and \s have NOT been changed.
3821
3822 16. Updated pcredemo.c to include calls to free() to release the memory for the
3823 compiled pattern.
3824
3825 17. The generated file chartables.c was being created in the source directory
3826 instead of in the building directory. This caused the build to fail if the
3827 source directory was different from the building directory, and was
3828 read-only.
3829
3830 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3831 file. No doubt somebody will tell me if they don't make sense... Also added
3832 Dan Mooney's comments about building on OpenVMS.
3833
3834 19. Added support for partial matching via the PCRE_PARTIAL option for
3835 pcre_exec() and the \P data escape in pcretest.
3836
3837 20. Extended pcretest with 3 new pattern features:
3838
3839 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3840 write the compiled pattern to the file whose name is "rest-of-line".
3841 This is a straight binary dump of the data, with the saved pointer to
3842 the character tables forced to be NULL. The study data, if any, is
3843 written too. After writing, pcretest reads a new pattern.
3844
3845 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3846 compiled pattern from the given file. There must not be any
3847 occurrences of "<" in the file name (pretty unlikely); if there are,
3848 pcretest will instead treat the initial "<" as a pattern delimiter.
3849 After reading in the pattern, pcretest goes on to read data lines as
3850 usual.
3851
3852 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3853 and 16-bit fields in a compiled pattern, to simulate a pattern that
3854 was compiled on a host of opposite endianness.
3855
3856 21. The pcre-exec() function can now cope with patterns that were compiled on
3857 hosts of opposite endianness, with this restriction:
3858
3859 As for any compiled expression that is saved and used later, the tables
3860 pointer field cannot be preserved; the extra_data field in the arguments
3861 to pcre_exec() should be used to pass in a tables address if a value
3862 other than the default internal tables were used at compile time.
3863
3864 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3865 now diagnosed as an error. Previously, most of the time, a negative number
3866 would have been treated as zero, but if in addition "ovector" was passed as
3867 NULL, a crash could occur.
3868
3869 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3870 new versions from the libtool 1.5 distribution (the last one is a copy of
3871 a file called libtool.m4). This seems to have fixed the need to patch
3872 "configure" to support Darwin 1.3 (which I used to do). However, I still
3873 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3874 workstation).
3875
3876 24. Changed the PCRE licence to be the more standard "BSD" licence.
3877
3878
3879 Version 4.5 01-Dec-03
3880 ---------------------
3881
3882 1. There has been some re-arrangement of the code for the match() function so
3883 that it can be compiled in a version that does not call itself recursively.
3884 Instead, it keeps those local variables that need separate instances for
3885 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3886 needs to "recurse". Keeping track of where control must go is done by means
3887 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3888 hide most of the details from the main code, and operates only if
3889 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3890 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3891 operating.
3892
3893 To make it easier for callers to provide specially tailored get/free
3894 functions for this usage, two new functions, pcre_stack_malloc, and
3895 pcre_stack_free, are used. They are always called in strict stacking order,
3896 and the size of block requested is always the same.
3897
3898 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3899 PCRE has been compiled to use the stack or the heap for recursion. The
3900 -C option of pcretest uses this to show which version is compiled.
3901
3902 A new data escape \S, is added to pcretest; it causes the amounts of store
3903 obtained and freed by both kinds of malloc/free at match time to be added
3904 to the output.
3905
3906 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3907 what's available on my current Linux desktop machine.
3908
3909 3. When matching a UTF-8 string, the test for a valid string at the start has
3910 been extended. If start_offset is not zero, PCRE now checks that it points
3911 to a byte that is the start of a UTF-8 character. If not, it returns
3912 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3913 this is necessary because there may be backward assertions in the pattern.
3914 When matching the same subject several times, it may save resources to use
3915 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3916
3917 4. The code for checking the validity of UTF-8 strings has been tightened so
3918 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3919 containing "overlong sequences".
3920
3921 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3922 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3923 should have been "&", but it just so happened that all the cases this let
3924 through by mistake were picked up later in the function.
3925
3926 6. I had used a variable called "isblank" - this is a C99 function, causing
3927 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3928
3929 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3930 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3931 so the version is shown; (c) stop "make test" echoing "./RunTest".
3932
3933 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3934
3935 9. The prototype for memmove() for systems that don't have it was using
3936 size_t, but the inclusion of the header that defines size_t was later. I've
3937 moved the #includes for the C headers earlier to avoid this.
3938
3939 10. Added some adjustments to the code to make it easier to compiler on certain
3940 special systems:
3941
3942 (a) Some "const" qualifiers were missing.
3943 (b) Added the macro EXPORT before all exported functions; by default this
3944 is defined to be empty.
3945 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3946 that it reads its output file name as an argument instead of writing
3947 to the standard output and assuming this can be redirected.
3948
3949 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3950 class containing characters with values greater than 255, PCRE compilation
3951 went into a loop.
3952
3953 12. A recursive reference to a subpattern that was within another subpattern
3954 that had a minimum quantifier of zero caused PCRE to crash. For example,
3955 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3956 recursion. If the recursively-called subpattern itself had a zero repeat,
3957 that was OK.
3958
3959 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3960 buffer into which it was copied (for escape processing) was still set at
3961 1024, so long lines caused crashes.
3962
3963 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3964 "internal error: code overflow...". This applied to any character class
3965 that was followed by a possessive quantifier.
3966
3967 15. Modified the Makefile to add libpcre.la as a prerequisite for
3968 libpcreposix.la because I was told this is needed for a parallel build to
3969 work.
3970
3971 16. If a pattern that contained .* following optional items at the start was
3972 studied, the wrong optimizing data was generated, leading to matching
3973 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3974 matching string must start with a or b or c. The correct conclusion for
3975 this pattern is that a match can start with any character.
3976
3977
3978 Version 4.4 13-Aug-03
3979 ---------------------
3980
3981 1. In UTF-8 mode, a character class containing characters with values between
3982 127 and 255 was not handled correctly if the compiled pattern was studied.
3983 In fixing this, I have also improved the studying algorithm for such
3984 classes (slightly).
3985
3986 2. Three internal functions had redundant arguments passed to them. Removal
3987 might give a very teeny performance improvement.
3988
3989 3. Documentation bug: the value of the capture_top field in a callout is *one
3990 more than* the number of the hightest numbered captured substring.
3991
3992 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3993 in incorrectly linking with a previously installed version. They now link
3994 explicitly with libpcre.la.
3995
3996 5. configure.in no longer needs to recognize Cygwin specially.
3997
3998 6. A problem in pcre.in for Windows platforms is fixed.
3999
4000 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4001 pcretest, it used to include the size of the study block as part of its
4002 output. Unfortunately, the structure contains a field that has a different
4003 size on different hardware architectures. This meant that the tests that
4004 showed this size failed. As the block is currently always of a fixed size,
4005 this information isn't actually particularly useful in pcretest output, so
4006 I have just removed it.
4007
4008 8. Three pre-processor statements accidentally did not start in column 1.
4009 Sadly, there are *still* compilers around that complain, even though
4010 standard C has not required this for well over a decade. Sigh.
4011
4012 9. In pcretest, the code for checking callouts passed small integers in the
4013 callout_data field, which is a void * field. However, some picky compilers
4014 complained about the casts involved for this on 64-bit systems. Now
4015 pcretest passes the address of the small integer instead, which should get
4016 rid of the warnings.
4017
4018 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4019 both compile and run time, and gives an error if an invalid UTF-8 sequence
4020 is found. There is a option for disabling this check in cases where the
4021 string is known to be correct and/or the maximum performance is wanted.
4022
4023 11. In response to a bug report, I changed one line in Makefile.in from
4024
4025 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4026 to
4027 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4028
4029 to look similar to other lines, but I have no way of telling whether this
4030 is the right thing to do, as I do not use Windows. No doubt I'll get told
4031 if it's wrong...
4032
4033
4034 Version 4.3 21-May-03
4035 ---------------------
4036
4037 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4038 Makefile.
4039
4040 2. Some refactoring to improve the quality of the code:
4041
4042 (i) The utf8_table... variables are now declared "const".
4043
4044 (ii) The code for \cx, which used the "case flipping" table to upper case
4045 lower case letters, now just substracts 32. This is ASCII-specific,
4046 but the whole concept of \cx is ASCII-specific, so it seems
4047 reasonable.
4048
4049 (iii) PCRE was using its character types table to recognize decimal and
4050 hexadecimal digits in the pattern. This is silly, because it handles
4051 only 0-9, a-f, and A-F, but the character types table is locale-
4052 specific, which means strange things might happen. A private
4053 table is now used for this - though it costs 256 bytes, a table is
4054 much faster than multiple explicit tests. Of course, the standard
4055 character types table is still used for matching digits in subject
4056 strings against \d.
4057
4058 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4059 ending in _t are). So I've renamed it as ESC_tee.
4060
4061 3. The first argument for regexec() in the POSIX wrapper should have been
4062 defined as "const".
4063
4064 4. Changed pcretest to use malloc() for its buffers so that they can be
4065 Electric Fenced for debugging.
4066
4067 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4068 to read one or more bytes before the start of the subject string. Often this
4069 had no effect on PCRE's behaviour, but in some circumstances it could
4070 provoke a segmentation fault.
4071
4072 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4073 to try to read one or more bytes before the start of the subject string.
4074
4075 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4076 UTF-8 support could misbehave in various ways if the subject string
4077 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4078 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4079 back over UTF-8 characters.)
4080
4081
4082 Version 4.2 14-Apr-03
4083 ---------------------
4084
4085 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4086
4087 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4088 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4089 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4090 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4091 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4092 and BUILD_EXEEXT
4093 Note: automatic setting of the BUILD variables is not yet working
4094 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4095 compile-time but not at link-time
4096 [LINK]: use for linking executables only
4097 make different versions for Windows and non-Windows
4098 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4099 libraries
4100 [LINK_FOR_BUILD]: new variable
4101 [OBJEXT]: use throughout
4102 [EXEEXT]: use throughout
4103 <winshared>: new target
4104 <wininstall>: new target
4105 <dftables.o>: use native compiler
4106 <dftables>: use native linker
4107 <install>: handle Windows platform correctly
4108 <clean>: ditto
4109 <check>: ditto
4110 copy DLL to top builddir before testing
4111
4112 As part of these changes, -no-undefined was removed again. This was reported
4113 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4114 in any case.
4115
4116 3. Some tidies to get rid of compiler warnings:
4117
4118 . In the match_data structure, match_limit was an unsigned long int, whereas
4119 match_call_count was an int. I've made them both unsigned long ints.
4120
4121 . In pcretest the fact that a const uschar * doesn't automatically cast to
4122 a void * provoked a warning.
4123
4124 . Turning on some more compiler warnings threw up some "shadow" variables
4125 and a few more missing casts.
4126
4127 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4128 option, a class that contained a single character with a value between 128
4129 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4130
4131 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4132 option, a class that contained several characters, but with at least one
4133 whose value was between 128 and 255 caused PCRE to crash.
4134
4135
4136 Version 4.1 12-Mar-03
4137 ---------------------
4138
4139 1. Compiling with gcc -pedantic found a couple of places where casts were
4140 needed, and a string in dftables.c that was longer than standard compilers are
4141 required to support.
4142
4143 2. Compiling with Sun's compiler found a few more places where the code could
4144 be tidied up in order to avoid warnings.
4145
4146 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4147 first of these names is deprecated in the latest Autoconf in favour of the name
4148 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4149 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4150 analogy I have changed it to CFLAGS_FOR_BUILD.
4151
4152 4. Added -no-undefined to the linking command in the Makefile, because this is
4153 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4154 linking step for the pcreposix library.
4155
4156 5. PCRE was failing to diagnose the case of two named groups with the same
4157 name.
4158
4159 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4160 literal character that is needed in the subject for a match, and scans along to
4161 ensure that it is present before embarking on the full matching process. This
4162 saves time in cases of nested unlimited repeats that are never going to match.
4163 Problem: the scan can take a lot of time if the subject is very long (e.g.
4164 megabytes), thus penalizing straightforward matches. It is now done only if the
4165 amount of subject to be scanned is less than 1000 bytes.
4166
4167 7. A lesser problem with the same optimization is that it was recording the
4168 first character of an anchored pattern as "needed", thus provoking a search
4169 right along the subject, even when the first match of the pattern was going to
4170 fail. The "needed" character is now not set for anchored patterns, unless it
4171 follows something in the pattern that is of non-fixed length. Thus, it still
4172 fulfils its original purpose of finding quick non-matches in cases of nested
4173 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4174
4175
4176 Version 4.0 17-Feb-03
4177 ---------------------
4178
4179 1. If a comment in an extended regex that started immediately after a meta-item
4180 extended to the end of string, PCRE compiled incorrect data. This could lead to
4181 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4182
4183 2. Moved to autoconf 2.53 and libtool 1.4.2.
4184
4185 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4186 the special perltest8 script is no longer needed - all the tests can be run
4187 from a single perltest script.
4188
4189 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4190 by \s. It has now been removed in PCRE. This means it isn't recognized as
4191 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4192 class [:space:] *does* include VT, thereby creating a mess.
4193
4194 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4195 space and tab.
4196
4197 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4198 its new features into the main test script, reducing the number of scripts.
4199
4200 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4201 were backward compatible, and made the (?i) apply to the whole pattern, as if
4202 /i were given. Now it behaves more logically, and applies the option setting
4203 only to what follows. PCRE has been changed to follow suit. However, if it
4204 finds options settings right at the start of the pattern, it extracts them into
4205 the global options, as before. Thus, they show up in the info data.
4206
4207 8. Added support for the \Q...\E escape sequence. Characters in between are
4208 treated as literals. This is slightly different from Perl in that $ and @ are
4209 also handled as literals inside the quotes. In Perl, they will cause variable
4210 interpolation. Note the following examples:
4211
4212 Pattern PCRE matches Perl matches
4213
4214 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4215 \Qabc\$xyz\E abc\$xyz abc\$xyz
4216 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4217
4218 For compatibility with Perl, \Q...\E sequences are recognized inside character
4219 classes as well as outside them.
4220
4221 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4222 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4223 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4224 signed/unsigned warnings.
4225
4226 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4227 option for pcretest, so I've replaced it by a simple function that does just
4228 that job.
4229
4230 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4231 "pcregrep -".
4232
4233 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4234 Java package. This provides some syntactic sugar for simple cases of what my
4235 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4236 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4237 item, you can use this simplified notation. Note that only makes sense with
4238 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4239 greediness, whatever the setting of the PCRE_UNGREEDY option.
4240
4241 13. A change of greediness default within a pattern was not taking effect at
4242 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4243 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4244 was abstracted outside.
4245
4246 14. PCRE now supports the \G assertion. It is true when the current matching
4247 position is at the start point of the match. This differs from \A when the
4248 starting offset is non-zero. Used with the /g option of pcretest (or similar
4249 code), it works in the same way as it does for Perl's /g option. If all
4250 alternatives of a regex begin with \G, the expression is anchored to the start
4251 match position, and the "anchored" flag is set in the compiled expression.
4252
4253 15. Some bugs concerning the handling of certain option changes within patterns
4254 have been fixed. These applied to options other than (?ims). For example,
4255 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4256 been the other way round. Some of this was related to change 7 above.
4257
4258 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4259 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4260 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4261 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4262
4263 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4264 mode. Unlike ".", it always matches newline, whatever the setting of
4265 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4266 assertions. Perl allows it, but it doesn't (in general) work because it can't
4267 calculate the length of the lookbehind. At least, that's the case for Perl
4268 5.8.0 - I've been told they are going to document that it doesn't work in
4269 future.
4270
4271 18. Added an error diagnosis for escapes that PCRE does not support: these are
4272 \L, \l, \N, \P, \p, \U, \u, and \X.
4273
4274 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4275 reading past the end of the pattern in cases such as /[abcd/.
4276
4277 20. PCRE was getting more memory than necessary for patterns with classes that
4278 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4279
4280 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4281 compiling PCRE for use with Virtual Pascal.
4282
4283 22. Small fix to the Makefile to make it work properly if the build is done
4284 outside the source tree.
4285
4286 23. Added a new extension: a condition to go with recursion. If a conditional
4287 subpattern starts with (?(R) the "true" branch is used if recursion has
4288 happened, whereas the "false" branch is used only at the top level.
4289
4290 24. When there was a very long string of literal characters (over 255 bytes
4291 without UTF support, over 250 bytes with UTF support), the computation of how
4292 much memory was required could be incorrect, leading to segfaults or other
4293 strange effects.
4294
4295 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4296 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4297 there was a subsequent back reference to those brackets. This meant that, for
4298 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4299 possible to check for precisely this case. All we can do is abandon the
4300 optimization if .* occurs inside capturing brackets when there are any back
4301 references whatsoever. (See below for a better fix that came later.)
4302
4303 26. The handling of the optimization for finding the first character of a
4304 non-anchored pattern, and for finding a character that is required later in the
4305 match were failing in some cases. This didn't break the matching; it just
4306 failed to optimize when it could. The way this is done has been re-implemented.
4307
4308 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4309
4310 28. Added a new feature that provides some of the functionality that Perl
4311 provides with (?{...}). The facility is termed a "callout". The way it is done
4312 in PCRE is for the caller to provide an optional function, by setting
4313 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4314 global variable. By default it is unset, which disables all calling out. To get
4315 the function called, the regex must include (?C) at appropriate points. This
4316 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4317 This provides a means of identifying different callout points. When PCRE
4318 reaches such a point in the regex, if pcre_callout has been set, the external
4319 function is called. It is provided with data in a structure called
4320 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4321 matching continues; if it returns a non-zero value, the match at the current
4322 point fails. However, backtracking will occur if possible. [This was changed
4323 later and other features added - see item 49 below.]
4324
4325 29. pcretest is upgraded to test the callout functionality. It provides a
4326 callout function that displays information. By default, it shows the start of
4327 the match and the current position in the text. There are some new data escapes
4328 to vary what happens:
4329
4330 \C+ in addition, show current contents of captured substrings
4331 \C- do not supply a callout function
4332 \C!n return 1 when callout number n is reached
4333 \C!n!m return 1 when callout number n is reached for the mth time
4334
4335 30. If pcregrep was called with the -l option and just a single file name, it
4336 output "<stdin>" if a match was found, instead of the file name.
4337
4338 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4339 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4340 pcre_exec(). This saves a malloc/free per call. The default value of
4341 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4342 when configuring.
4343
4344 32. The default maximum size of a compiled pattern is 64K. There have been a
4345 few cases of people hitting this limit. The code now uses macros to handle the
4346 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4347 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4348 configuring. Tests 2 and 5 work only with 2-byte links because they output
4349 debugging information about compiled patterns.
4350
4351 33. Internal code re-arrangements:
4352
4353 (a) Moved the debugging function for printing out a compiled regex into
4354 its own source file (printint.c) and used #include to pull it into
4355 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4356 separate copies.
4357
4358 (b) Defined the list of op-code names for debugging as a macro in
4359 internal.h so that it is next to the definition of the opcodes.
4360
4361 (c) Defined a table of op-code lengths for simpler skipping along compiled
4362 code. This is again a macro in internal.h so that it is next to the
4363 definition of the opcodes.
4364
4365 34. Added support for recursive calls to individual subpatterns, along the
4366 lines of Robin Houston's patch (but implemented somewhat differently).
4367
4368 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4369 allow it to read and process whole directories in Win32. This code was
4370 contributed by Lionel Fourquaux; it has not been tested by me.
4371
4372 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4373 used to name a group. Names consist of alphanumerics and underscores, and must
4374 be unique. Back references use the syntax (?P=name) and recursive calls use
4375 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4376 numbers. The function pcre_fullinfo() can be used after compilation to extract
4377 a name/number map. There are three relevant calls:
4378
4379 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4380 PCRE_INFO_NAMECOUNT yields the number of entries
4381 PCRE_INFO_NAMETABLE yields a pointer to the map.
4382
4383 The map is a vector of fixed-size entries. The size of each entry depends on
4384 the length of the longest name used. The first two bytes of each entry are the
4385 group number, most significant byte first. There follows the corresponding
4386 name, zero terminated. The names are in alphabetical order.
4387
4388 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4389 case instead of 255. Making it the same both with and without UTF-8 support
4390 means that the same test output works with both.
4391
4392 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4393 calling malloc() with a zero argument.
4394
4395 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4396 optimization. I've improved things by keeping a bitmap of backreferences with
4397 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4398 fact referenced, the optimization can be applied. It is unlikely that a
4399 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4400 the match to follow \n) will appear inside brackets with a number greater than
4401 31, but if it does, any back reference > 31 suppresses the optimization.
4402
4403 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4404 of disabling numbered capturing parentheses. Any opening parenthesis that is
4405 not followed by ? behaves as if it were followed by ?: but named parentheses
4406 can still be used for capturing (and they will acquire numbers in the usual
4407 way).
4408
4409 41. Redesigned the return codes from the match() function into yes/no/error so
4410 that errors can be passed back from deep inside the nested calls. A malloc
4411 failure while inside a recursive subpattern call now causes the
4412 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4413
4414 42. It is now possible to set a limit on the number of times the match()
4415 function is called in a call to pcre_exec(). This facility makes it possible to
4416 limit the amount of recursion and backtracking, though not in a directly
4417 obvious way, because the match() function is used in a number of different
4418 circumstances. The count starts from zero for each position in the subject
4419 string (for non-anchored patterns). The default limit is, for compatibility, a
4420 large number, namely 10 000 000. You can change this in two ways:
4421
4422 (a) When configuring PCRE before making, you can use --with-match-limit=n
4423 to set a default value for the compiled library.
4424
4425 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4426 a different value is set. See 45 below.
4427
4428 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4429
4430 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4431 of things that can be changed at compile time. The first argument specifies
4432 what is wanted and the second points to where the information is to be placed.
4433 The current list of available information is:
4434
4435 PCRE_CONFIG_UTF8
4436
4437 The output is an integer that is set to one if UTF-8 support is available;
4438 otherwise it is set to zero.
4439
4440 PCRE_CONFIG_NEWLINE
4441
4442 The output is an integer that it set to the value of the code that is used for
4443 newline. It is either LF (10) or CR (13).
4444
4445 PCRE_CONFIG_LINK_SIZE
4446
4447 The output is an integer that contains the number of bytes used for internal
4448 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4449
4450 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4451
4452 The output is an integer that contains the threshold above which the POSIX
4453 interface uses malloc() for output vectors. See item 31 above.
4454
4455 PCRE_CONFIG_MATCH_LIMIT
4456
4457 The output is an unsigned integer that contains the default limit of the number
4458 of match() calls in a pcre_exec() execution. See 42 above.
4459
4460 44. pcretest has been upgraded by the addition of the -C option. This causes it
4461 to extract all the available output from the new pcre_config() function, and to
4462 output it. The program then exits immediately.
4463
4464 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4465 order to support additional features. One way would have been to define
4466 pcre_exec2() (for example) with extra arguments, but this would not have been
4467 extensible, and would also have required all calls to the original function to
4468 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4469 is used for passing in "extra" data from pcre_study().
4470
4471 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4472 contains the following fields:
4473
4474 flags a bitmap indicating which of the following fields are set
4475 study_data opaque data from pcre_study()
4476 match_limit a way of specifying a limit on match() calls for a specific
4477 call to pcre_exec()
4478 callout_data data for callouts (see 49 below)
4479
4480 The flag bits are also defined in pcre.h, and are
4481
4482 PCRE_EXTRA_STUDY_DATA
4483 PCRE_EXTRA_MATCH_LIMIT
4484 PCRE_EXTRA_CALLOUT_DATA
4485
4486 The pcre_study() function now returns one of these new pcre_extra blocks, with
4487 the actual study data pointed to by the study_data field, and the
4488 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4489 before. That is, this change is entirely upwards-compatible and requires no
4490 change to existing code.
4491
4492 If you want to pass in additional data to pcre_exec(), you can either place it
4493 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4494 block.
4495
4496 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4497 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4498 times with different match limits, until it finds the minimum value needed for
4499 pcre_exec() to complete. The value is then output. This can be instructive; for
4500 most simple matches the number is quite small, but for pathological cases it
4501 gets very large very quickly.
4502
4503 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4504 returns the size of the data block pointed to by the study_data field in a
4505 pcre_extra block, that is, the value that was passed as the argument to
4506 pcre_malloc() when PCRE was getting memory in which to place the information
4507 created by pcre_study(). The fourth argument should point to a size_t variable.
4508 pcretest has been extended so that this information is shown after a successful
4509 pcre_study() call when information about the compiled regex is being displayed.
4510
4511 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4512 because what follows is always an absolute path. (Later: it turns out that this
4513 is more than cosmetic for MinGW, because it doesn't like empty path
4514 components.)
4515
4516 49. Some changes have been made to the callout feature (see 28 above):
4517
4518 (i) A callout function now has three choices for what it returns:
4519
4520 0 => success, carry on matching
4521 > 0 => failure at this point, but backtrack if possible
4522 < 0 => serious error, return this value from pcre_exec()
4523
4524 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4525 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4526 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4527 use by callout functions. It will never be used by PCRE itself.
4528
4529 (ii) The pcre_extra structure (see 45 above) has a void * field called
4530 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4531 pcre_callout_block structure has a field of the same name. The contents of
4532 the field passed in the pcre_extra structure are passed to the callout
4533 function in the corresponding field in the callout block. This makes it
4534 easier to use the same callout-containing regex from multiple threads. For
4535 testing, the pcretest program has a new data escape
4536
4537 \C*n pass the number n (may be negative) as callout_data
4538
4539 If the callout function in pcretest receives a non-zero value as
4540 callout_data, it returns that value.
4541
4542 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4543 there were some redundant $(CFLAGS) in commands that are now specified as
4544 $(LINK), which already includes $(CFLAGS).
4545
4546 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4547 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4548 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4549 one-byte characters throughout. Note that case-insensitive matching applies
4550 only to characters whose values are less than 256. PCRE doesn't support the
4551 notion of cases for higher-valued characters.
4552
4553 (i) A character class whose characters are all within 0-255 is handled as
4554 a bit map, and the map is inverted for negative classes. Previously, a
4555 character > 255 always failed to match such a class; however it should
4556 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4557
4558 (ii) A negated character class with a single character < 255 is coded as
4559 "not this character" (OP_NOT). This wasn't working properly when the test
4560 character was multibyte, either singly or repeated.
4561
4562 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4563 mode, for example: \x{100}{2,3}.
4564
4565 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4566 singly or repeated) now correctly test multibyte characters. However,
4567 PCRE doesn't recognize any characters with values greater than 255 as
4568 digits, spaces, or word characters. Such characters always match \D, \S,
4569 and \W, and never match \d, \s, or \w.
4570
4571 (v) Classes may now contain characters and character ranges with values
4572 greater than 255. For example: [ab\x{100}-\x{400}].
4573
4574 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4575 PCRE in UTF-8 mode.
4576
4577 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4578 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4579 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4580 value.)
4581
4582 53. The single man page has become too large. I have therefore split it up into
4583 a number of separate man pages. These also give rise to individual HTML pages;
4584 these are now put in a separate directory, and there is an index.html page that
4585 lists them all. Some hyperlinking between the pages has been installed.
4586
4587 54. Added convenience functions for handling named capturing parentheses.
4588
4589 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4590 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4591 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4592 are faulted.
4593
4594 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4595 calling configure. These values are used when compiling the dftables.c program
4596 which is run to generate the source of the default character tables. They
4597 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4598 you will need to set these values.
4599
4600 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4601
4602
4603 Version 3.9 02-Jan-02
4604 ---------------------
4605
4606 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4607
4608 2. If --disable-static was given, the building process failed when trying to
4609 build pcretest and pcregrep. (For some reason it was using libtool to compile
4610 them, which is not right, as they aren't part of the library.)
4611
4612
4613 Version 3.8 18-Dec-01
4614 ---------------------
4615
4616 1. The experimental UTF-8 code was completely screwed up. It was packing the
4617 bytes in the wrong order. How dumb can you get?
4618
4619
4620 Version 3.7 29-Oct-01
4621 ---------------------
4622
4623 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4624 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4625 this didn't happen under Solaris 8, where I normally test things.
4626
4627 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4628 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4629 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4630 this fix an hour or so after the initial 3.7 release.)
4631
4632
4633 Version 3.6 23-Oct-01
4634 ---------------------
4635
4636 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4637 offsets passed as NULL with zero offset count.
4638
4639 2. The config.guess and config.sub files had not been updated when I moved to
4640 the latest autoconf.
4641
4642
4643 Version 3.5 15-Aug-01
4644 ---------------------
4645
4646 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4647 had been forgotten.
4648
4649 2. By using declared but undefined structures, we can avoid using "void"
4650 definitions in pcre.h while keeping the internal definitions of the structures
4651 private.
4652
4653 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4654 user point of view, this means that both static and shared libraries are built
4655 by default, but this can be individually controlled. More of the work of
4656 handling this static/shared cases is now inside libtool instead of PCRE's make
4657 file.
4658
4659 4. The pcretest utility is now installed along with pcregrep because it is
4660 useful for users (to test regexs) and by doing this, it automatically gets
4661 relinked by libtool. The documentation has been turned into a man page, so
4662 there are now .1, .txt, and .html versions in /doc.
4663
4664 5. Upgrades to pcregrep:
4665 (i) Added long-form option names like gnu grep.
4666 (ii) Added --help to list all options with an explanatory phrase.
4667 (iii) Added -r, --recursive to recurse into sub-directories.
4668 (iv) Added -f, --file to read patterns from a file.
4669
4670 6. pcre_exec() was referring to its "code" argument before testing that
4671 argument for NULL (and giving an error if it was NULL).
4672
4673 7. Upgraded Makefile.in to allow for compiling in a different directory from
4674 the source directory.
4675
4676 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4677 options bits, the pointer it was passed was to an int instead of to an unsigned
4678 long int. This mattered only on 64-bit systems.
4679
4680 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4681 generated) instead of pcre.in, which it its source. Also made the same change
4682 in several of the .c files.
4683
4684 10. A new release of gcc defines printf() as a macro, which broke pcretest
4685 because it had an ifdef in the middle of a string argument for printf(). Fixed
4686 by using separate calls to printf().
4687
4688 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4689 script, to force use of CR or LF instead of \n in the source. On non-Unix
4690 systems, the value can be set in config.h.
4691
4692 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4693 absolute limit. Changed the text of the error message to make this clear, and
4694 likewise updated the man page.
4695
4696 13. The limit of 99 on the number of capturing subpatterns has been removed.
4697 The new limit is 65535, which I hope will not be a "real" limit.
4698
4699
4700 Version 3.4 22-Aug-00
4701 ---------------------
4702
4703 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4704
4705 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4706
4707
4708 Version 3.3 01-Aug-00
4709 ---------------------
4710
4711 1. If an octal character was given, but the value was greater than \377, it
4712 was not getting masked to the least significant bits, as documented. This could
4713 lead to crashes in some systems.
4714
4715 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4716 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4717
4718 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4719 These just pass their arguments on to (pcre_free)(), but they are provided
4720 because some uses of PCRE bind it to non-C systems that can call its functions,
4721 but cannot call free() or pcre_free() directly.
4722
4723 4. Add "make test" as a synonym for "make check". Corrected some comments in
4724 the Makefile.
4725
4726 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4727 Makefile.
4728
4729 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4730 command called pgrep for grepping around the active processes.
4731
4732 7. Added the beginnings of support for UTF-8 character strings.
4733
4734 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4735 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4736 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4737 out for the ar command.)
4738
4739
4740 Version 3.2 12-May-00
4741 ---------------------
4742
4743 This is purely a bug fixing release.
4744
4745 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4746 of ZA. This was just one example of several cases that could provoke this bug,
4747 which was introduced by change 9 of version 2.00. The code for breaking
4748 infinite loops after an iteration that matches an empty string was't working
4749 correctly.
4750
4751 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4752 when matched against abbab (for example). After matching an empty string, it
4753 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4754 caused it to match further down the string than it should.
4755
4756 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4757 was there because it doesn't seem to be needed, and it causes trouble on some
4758 systems, as it is not a Standard C header. It has been removed.
4759
4760 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4761 were reported on the Macintosh. The changes were from
4762
4763 while ((c = *(++ptr)) != 0 && c != '\n');
4764 to
4765 while ((c = *(++ptr)) != 0 && c != '\n') ;
4766
4767 Totally extraordinary, but if that's what it takes...
4768
4769 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4770 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4771 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4772 assumes the way PCRE uses memmove() (always moving upwards).
4773
4774 6. PCRE is being used in one environment where strchr() is not available. There
4775 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4776 faster code anyway.
4777
4778
4779 Version 3.1 09-Feb-00
4780 ---------------------
4781
4782 The only change in this release is the fixing of some bugs in Makefile.in for
4783 the "install" target:
4784
4785 (1) It was failing to install pcreposix.h.
4786
4787 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4788
4789
4790 Version 3.0 01-Feb-00
4791 ---------------------
4792
4793 1. Add support for the /+ modifier to perltest (to output $` like it does in
4794 pcretest).
4795
4796 2. Add support for the /g modifier to perltest.
4797
4798 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4799 matches null strings.
4800
4801 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4802 pattern. Perl treats empty patterns specially - it reuses the most recent
4803 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4804 effect.
4805
4806 5. The POSIX interface was broken in that it was just handing over the POSIX
4807 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4808 required a bigger vector, with some working space on the end. This means that
4809 the POSIX wrapper now has to get and free some memory, and copy the results.
4810
4811 6. Added some simple autoconf support, placing the test data and the
4812 documentation in separate directories, re-organizing some of the
4813 information files, and making it build pcre-config (a GNU standard). Also added
4814 libtool support for building PCRE as a shared library, which is now the
4815 default.
4816
4817 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4818 09 are not valid octal constants. Single digits will be used for minor values
4819 less than 10.
4820
4821 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4822 existing programs that set these in the POSIX interface can use PCRE without
4823 modification.
4824
4825 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4826 return all that pcre_info() returns, plus additional data. The pcre_info()
4827 function is retained for compatibility, but is considered to be obsolete.
4828
4829 10. Added experimental recursion feature (?R) to handle one common case that
4830 Perl 5.6 will be able to do with (?p{...}).
4831
4832 11. Added support for POSIX character classes like [:alpha:], which Perl is
4833 adopting.
4834
4835
4836 Version 2.08 31-Aug-99
4837 ----------------------
4838
4839 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4840 trying to match at the startoffset position, but instead was moving forward to
4841 the next newline as if a previous match had failed.
4842
4843 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4844 and could get into a loop if a null string was matched other than at the start
4845 of the subject.
4846
4847 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4848 be distinguished at compile time, and for completeness also added PCRE_DATE.
4849
4850 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4851 in GnuWin32 environments.
4852
4853
4854 Version 2.07 29-Jul-99
4855 ----------------------
4856
4857 1. The documentation is now supplied in plain text form and HTML as well as in
4858 the form of man page sources.
4859
4860 2. C++ compilers don't like assigning (void *) values to other pointer types.
4861 In particular this affects malloc(). Although there is no problem in Standard
4862 C, I've put in casts to keep C++ compilers happy.
4863
4864 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4865 should be (const char *).
4866
4867 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4868 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4869 However, I haven't made this a standard facility. The documentation doesn't
4870 mention it, and the Makefile doesn't support it.
4871
4872 5. The Makefile now contains an "install" target, with editable destinations at
4873 the top of the file. The pcretest program is not installed.
4874
4875 6. pgrep -V now gives the PCRE version number and date.
4876
4877 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4878 causing the entire string to be ignored, instead of just the last character.
4879
4880 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4881 non-matching string, it can take a very, very long time, even for strings of
4882 quite modest length, because of the nested recursion. PCRE now does better in
4883 some of these cases. It does this by remembering the last required literal
4884 character in the pattern, and pre-searching the subject to ensure it is present
4885 before running the real match. In other words, it applies a heuristic to detect
4886 some types of certain failure quickly, and in the above example, if presented
4887 with a string that has no trailing " it gives "no match" very quickly.
4888
4889 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4890 other alternatives are tried instead.
4891
4892
4893 Version 2.06 09-Jun-99
4894 ----------------------
4895
4896 1. Change pcretest's output for amount of store used to show just the code
4897 space, because the remainder (the data block) varies in size between 32-bit and
4898 64-bit systems.
4899
4900 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4901 start matching at. This allows lookbehinds to work when searching for multiple
4902 occurrences in a string.
4903
4904 3. Added additional options to pcretest for testing multiple occurrences:
4905
4906 /+ outputs the rest of the string that follows a match
4907 /g loops for multiple occurrences, using the new startoffset argument
4908 /G loops for multiple occurrences by passing an incremented pointer
4909
4910 4. PCRE wasn't doing the "first character" optimization for patterns starting
4911 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4912 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4913 the letter 'x'. On long subject strings, this gives a significant speed-up.
4914
4915
4916 Version 2.05 21-Apr-99
4917 ----------------------
4918
4919 1. Changed the type of magic_number from int to long int so that it works
4920 properly on 16-bit systems.
4921
4922 2. Fixed a bug which caused patterns starting with .* not to work correctly
4923 when the subject string contained newline characters. PCRE was assuming
4924 anchoring for such patterns in all cases, which is not correct because .* will
4925 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4926 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4927 must be retried after every newline in the subject.
4928
4929
4930 Version 2.04 18-Feb-99
4931 ----------------------
4932
4933 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4934 computation of the store needed to hold the pattern was incorrect (too large).
4935 If such patterns were nested a few deep, this could multiply and become a real
4936 problem.
4937
4938 2. Added /M option to pcretest to show the memory requirement of a specific
4939 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4940
4941 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4942 compiled in such a way that the backtracking after subsequent failure was
4943 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4944 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4945
4946
4947 Version 2.03 02-Feb-99
4948 ----------------------
4949
4950 1. Fixed typo and small mistake in man page.
4951
4952 2. Added 4th condition (GPL supersedes if conflict) and created separate
4953 LICENCE file containing the conditions.
4954
4955 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4956 Perl, that is the internal \ allows the delimiter to be included in the
4957 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4958 the final delimiter, add \ to the end of the pattern (to test the error).
4959
4960 4. Added the convenience functions for extracting substrings after a successful
4961 match. Updated pcretest to make it able to test these functions.
4962
4963
4964 Version 2.02 14-Jan-99
4965 ----------------------
4966
4967 1. Initialized the working variables associated with each extraction so that
4968 their saving and restoring doesn't refer to uninitialized store.
4969
4970 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4971 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4972 fix the problem.
4973
4974 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4975 calls, and wasn't printing the correct value for compiling calls. Increased the
4976 default value of LOOPREPEAT, and the number of significant figures in the
4977 times.
4978
4979 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4980
4981 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4982 a building problem on Windows NT with a FAT file system.
4983
4984
4985 Version 2.01 21-Oct-98
4986 ----------------------
4987
4988 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4989 to character tables built by pcre_maketables() in the current locale. If NULL
4990 is passed, the default tables are used.
4991
4992
4993 Version 2.00 24-Sep-98
4994 ----------------------
4995
4996 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4997 it any more.
4998
4999 2. Allow quantification of (?>) groups, and make it work correctly.
5000
5001 3. The first character computation wasn't working for (?>) groups.
5002
5003 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5004 end of the subject) and add 5.005's \z, which really does match only at the
5005 very end of the subject.
5006
5007 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5008
5009 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5010 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5011 localized options. All options to pcre_study() were also removed.
5012
5013 7. Add other new features from 5.005:
5014
5015 $(?<= positive lookbehind
5016 $(?<! negative lookbehind
5017 (?imsx-imsx) added the unsetting capability
5018 such a setting is global if at outer level; local otherwise
5019 (?imsx-imsx:) non-capturing groups with option setting
5020 (?(cond)re|re) conditional pattern matching
5021
5022 A backreference to itself in a repeated group matches the previous
5023 captured string.
5024
5025 8. General tidying up of studying (both automatic and via "study")
5026 consequential on the addition of new assertions.
5027
5028 9. As in 5.005, unlimited repeated groups that could match an empty substring
5029 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5030 runtime if any iteration does actually match an empty substring.
5031
5032 10. Include the RunTest script in the distribution.
5033
5034 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5035 discrepancies, some of which were old and were also with respect to 5.004. They
5036 have now been fixed.
5037
5038
5039 Version 1.09 28-Apr-98
5040 ----------------------
5041
5042 1. A negated single character class followed by a quantifier with a minimum
5043 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5044 program crashes, or just wrong answers. This did not apply to negated classes
5045 containing more than one character, or to minima other than one.
5046
5047
5048 Version 1.08 27-Mar-98
5049 ----------------------
5050
5051 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5052
5053 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5054 latter must appear before anything that relies on it in the pattern.
5055
5056
5057 Version 1.07 16-Feb-98
5058 ----------------------
5059
5060 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5061 repeat of a potentially empty string).
5062
5063
5064 Version 1.06 23-Jan-98
5065 ----------------------
5066
5067 1. Added Markus Oberhumer's little patches for C++.
5068
5069 2. Literal strings longer than 255 characters were broken.
5070
5071
5072 Version 1.05 23-Dec-97
5073 ----------------------
5074
5075 1. Negated character classes containing more than one character were failing if
5076 PCRE_CASELESS was set at run time.
5077
5078
5079 Version 1.04 19-Dec-97
5080 ----------------------
5081
5082 1. Corrected the man page, where some "const" qualifiers had been omitted.
5083
5084 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5085 input syntax.
5086
5087 3. Fixed memory leak which occurred when a regex with back references was
5088 matched with an offsets vector that wasn't big enough. The temporary memory
5089 that is used in this case wasn't being freed if the match failed.
5090
5091 4. Tidied pcretest to ensure it frees memory that it gets.
5092
5093 5. Temporary memory was being obtained in the case where the passed offsets
5094 vector was exactly big enough.
5095
5096 6. Corrected definition of offsetof() from change 5 below.
5097
5098 7. I had screwed up change 6 below and broken the rules for the use of
5099 setjmp(). Now fixed.
5100
5101
5102 Version 1.03 18-Dec-97
5103 ----------------------
5104
5105 1. A erroneous regex with a missing opening parenthesis was correctly
5106 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5107 on some systems.
5108
5109 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5110 it was reported that one broken compiler failed on the former because "code" is
5111 also an independent variable.
5112
5113 3. The erroneous regex a[]b caused an array overrun reference.
5114
5115 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5116 fail on data ending with that character. (It was going on too far, and checking
5117 the next character, typically a binary zero.) This was specific to the
5118 optimized code for single-character negative classes.
5119
5120 5. Added a contributed patch from the TIN world which does the following:
5121
5122 + Add an undef for memmove, in case the the system defines a macro for it.
5123
5124 + Add a definition of offsetof(), in case there isn't one. (I don't know
5125 the reason behind this - offsetof() is part of the ANSI standard - but
5126 it does no harm).
5127
5128 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5129 most of the places where whitespace preceded '#'. I have given up and
5130 allowed the remaining 2 cases to be at the margin.
5131
5132 + Rename some variables in pcre to eliminate shadowing. This seems very
5133 pedantic, but does no harm, of course.
5134
5135 6. Moved the call to setjmp() into its own function, to get rid of warnings
5136 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5137
5138 7. Constructs such as \d{8,} were compiling into the equivalent of
5139 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5140 outcome, but in this particular case used more store than had been allocated,
5141 which caused the bug to be discovered because it threw up an internal error.
5142
5143 8. The debugging code in both pcre and pcretest for outputting the compiled
5144 form of a regex was going wrong in the case of back references followed by
5145 curly-bracketed repeats.
5146
5147
5148 Version 1.02 12-Dec-97
5149 ----------------------
5150
5151 1. Typos in pcre.3 and comments in the source fixed.
5152
5153 2. Applied a contributed patch to get rid of places where it used to remove
5154 'const' from variables, and fixed some signed/unsigned and uninitialized
5155 variable warnings.
5156
5157 3. Added the "runtest" target to Makefile.
5158
5159 4. Set default compiler flag to -O2 rather than just -O.
5160
5161
5162 Version 1.01 19-Nov-97
5163 ----------------------
5164
5165 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5166 like /([ab]*)*/, that is, for classes with more than one character in them.
5167
5168 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5169 as /((?>a*))*/ (a PCRE_EXTRA facility).
5170
5171
5172 Version 1.00 18-Nov-97
5173 ----------------------
5174
5175 1. Added compile-time macros to support systems such as SunOS4 which don't have
5176 memmove() or strerror() but have other things that can be used instead.
5177
5178 2. Arranged that "make clean" removes the executables.
5179
5180
5181 Version 0.99 27-Oct-97
5182 ----------------------
5183
5184 1. Fixed bug in code for optimizing classes with only one character. It was
5185 initializing a 32-byte map regardless, which could cause it to run off the end
5186 of the memory it had got.
5187
5188 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5189
5190
5191 Version 0.98 22-Oct-97
5192 ----------------------
5193
5194 1. Fixed bug in code for handling temporary memory usage when there are more
5195 back references than supplied space in the ovector. This could cause segfaults.
5196
5197
5198 Version 0.97 21-Oct-97
5199 ----------------------
5200
5201 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5202
5203 2. Optimized negated single characters not to use a bit map.
5204
5205 3. Brought error texts together as macro definitions; clarified some of them;
5206 fixed one that was wrong - it said "range out of order" when it meant "invalid
5207 escape sequence".
5208
5209 4. Changed some char * arguments to const char *.
5210
5211 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5212
5213 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5214 pcretest.
5215
5216
5217 Version 0.96 16-Oct-97
5218 ----------------------
5219
5220 1. Added a simple "pgrep" utility to the distribution.
5221
5222 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5223 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5224 where "ddd" means "one or more decimal digits".
5225
5226 3. Fixed serious bug. If a pattern had a back reference, but the call to
5227 pcre_exec() didn't supply a large enough ovector to record the related
5228 identifying subpattern, the match always failed. PCRE now remembers the number
5229 of the largest back reference, and gets some temporary memory in which to save
5230 the offsets during matching if necessary, in order to ensure that
5231 backreferences always work.
5232
5233 4. Increased the compatibility with Perl in a number of ways:
5234
5235 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5236 to request this handling. The option can be set at compile or exec time.
5237
5238 (b) $ matches before a terminating newline by default; an option
5239 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5240 mode). The option can be set at compile or exec time.
5241
5242 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5243 the same as Perl's. If the decimal number it represents is less than 10
5244 or there aren't that many previous left capturing parentheses, an octal
5245 escape is read. Inside a character class, it's always an octal escape,
5246 even if it is a single digit.
5247
5248 (d) An escaped but undefined alphabetic character is taken as a literal,
5249 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5250 escapes.
5251
5252 (e) {0} is now permitted. (The previous item is removed from the compiled
5253 pattern).
5254
5255 5. Changed all the names of code files so that the basic parts are no longer
5256 than 10 characters, and abolished the teeny "globals.c" file.
5257
5258 6. Changed the handling of character classes; they are now done with a 32-byte
5259 bit map always.
5260
5261 7. Added the -d and /D options to pcretest to make it possible to look at the
5262 internals of compilation without having to recompile pcre.
5263
5264
5265 Version 0.95 23-Sep-97
5266 ----------------------
5267
5268 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5269 \x20 at the start of a run of normal characters. These were being treated as
5270 real characters, instead of the source characters being re-checked.
5271
5272
5273 Version 0.94 18-Sep-97
5274 ----------------------
5275
5276 1. The functions are now thread-safe, with the caveat that the global variables
5277 containing pointers to malloc() and free() or alternative functions are the
5278 same for all threads.
5279
5280 2. Get pcre_study() to generate a bitmap of initial characters for non-
5281 anchored patterns when this is possible, and use it if passed to pcre_exec().
5282
5283
5284 Version 0.93 15-Sep-97
5285 ----------------------
5286
5287 1. /(b)|(:+)/ was computing an incorrect first character.
5288
5289 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5290 but not actually doing anything yet.
5291
5292 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5293 as Perl does (e.g. [-az] or [az-]).
5294
5295 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5296 all possible positions.
5297
5298 5. Split up into different modules to avoid including unneeded functions in a
5299 compiled binary. However, compile and exec are still in one module. The "study"
5300 function is split off.
5301
5302 6. The character tables are now in a separate module whose source is generated
5303 by an auxiliary program - but can then be edited by hand if required. There are
5304 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5305 toupper() in the code.
5306
5307 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5308 make them global. Abolish the function for setting them, as the caller can now
5309 set them directly.
5310
5311
5312 Version 0.92 11-Sep-97
5313 ----------------------
5314
5315 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5316 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5317
5318 2. Caseless matching was not working in character classes if the characters in
5319 the pattern were in upper case.
5320
5321 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5322
5323 4. Make PCRE_ANCHORED public and accept as a compile option.
5324
5325 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5326 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5327 pass them.
5328
5329 6. Give an error if bad option bits passed at compile or run time.
5330
5331 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5332 pcretest to cause it to pass that flag.
5333
5334 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5335 options, and the first character, if set.
5336
5337 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5338
5339
5340 Version 0.91 10-Sep-97
5341 ----------------------
5342
5343 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5344 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5345
5346 2. PCRE was looping on encountering an indefinitely repeated back reference to
5347 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5348 Perl does - treats the match as successful.
5349
5350 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5