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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1376 - (show annotations)
Sat Oct 12 18:02:11 2013 UTC (6 years, 1 month ago) by ph10
File size: 254523 byte(s)
Error occurred while calculating annotation data.
Add U+0085 and U+180E to what \s matches in UCP mode, to match Perl.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.34 xx-xxxx-201x
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_POSSESSIFY is added for testing purposes, and the -O and /O
89 options in pcretest are provided to set it.
90
91 18. The character VT has been added to the set of characters that match \s and
92 are generally treated as white space, following this same change in Perl
93 5.18. There is now no difference between "Perl space" and "POSIX space".
94
95 19. The code for checking named groups as conditions, either for being set or
96 for being recursed, has been refactored (this is related to 14 and 15
97 above). Processing unduplicated named groups should now be as fast at
98 numerical groups, and processing duplicated groups should be faster than
99 before.
100
101 20. Two patches to the CMake build system, by Alexander Barkov:
102
103 (1) Replace the "source" command by "." in CMakeLists.txt because
104 "source" is a bash-ism.
105
106 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
107 without these the CMake build does not work on Solaris.
108
109 21. Perl has changed its handling of \8 and \9. If there is no previously
110 encountered capturing group of those numbers, they are treated as the
111 literal characters 8 and 9 instead of a binary zero followed by the
112 literals. PCRE now does the same.
113
114 22. Following Perl, added \o{} to specify codepoints in octal, making it
115 possible to specify values greater than 0777 and also making them
116 unambiguous.
117
118 23. Perl now gives an error for missing closing braces after \x{... instead of
119 treating the string as literal. PCRE now does the same.
120
121 24. RunTest used to grumble if an inappropriate test was selected explicitly,
122 but just skip it when running all tests. This make it awkward to run ranges
123 of tests when one of them was inappropriate. Now it just skips any
124 inappropriate tests, as it always did when running all tests.
125
126 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
127 character types such as \d or \w, too many callouts were inserted, and the
128 data that they returned was rubbish.
129
130 26. In UCP mode, \s was not matching two of the characters that Perl matches,
131 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
132 were matched by \h. The code has now been refactored so that the lists of
133 the horizontal and vertical whitespace characters used for \h and \v (which
134 are defined only in one place) are now also used for \s.
135
136
137 Version 8.33 28-May-2013
138 ------------------------
139
140 1. Added 'U' to some constants that are compared to unsigned integers, to
141 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
142 variables that are added to signed variables, to ensure the result is
143 signed and can be negated.
144
145 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
146 CMake config files.
147
148 3. Revise the creation of config.h.generic so that all boolean macros are
149 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
150 overriding via -D on the command line possible.
151
152 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
153 to unsigned int is reported to make a quite noticeable speed difference in
154 a specific Windows environment. Testing on Linux did also appear to show
155 some benefit (and it is clearly not harmful). Also fixed the definition of
156 Xop which should be unsigned.
157
158 5. Related to (4), changing the definition of the intermediate variable cc
159 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
160 improvements.
161
162 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
163 unnecessary spaces.
164
165 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
166 and later.
167
168 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
169
170 9. Optimizing fast_forward_start_bits in JIT.
171
172 10. Adding support for callouts in JIT, and fixing some issues revealed
173 during this work. Namely:
174
175 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
176
177 (b) Minimum length was not checked before the matching is started.
178
179 11. The value of capture_last that is passed to callouts was incorrect in some
180 cases when there was a capture on one path that was subsequently abandoned
181 after a backtrack. Also, the capture_last value is now reset after a
182 recursion, since all captures are also reset in this case.
183
184 12. The interpreter no longer returns the "too many substrings" error in the
185 case when an overflowing capture is in a branch that is subsequently
186 abandoned after a backtrack.
187
188 13. In the pathological case when an offset vector of size 2 is used, pcretest
189 now prints out the matched string after a yield of 0 or 1.
190
191 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
192 Only supported by the JIT compiler at the moment.
193
194 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
195
196 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
197 offset of the starting point of the matching process, provided the offsets
198 vector is large enough.
199
200 17. The \A escape now records a lookbehind value of 1, though its execution
201 does not actually inspect the previous character. This is to ensure that,
202 in partial multi-segment matching, at least one character from the old
203 segment is retained when a new segment is processed. Otherwise, if there
204 are no lookbehinds in the pattern, \A might match incorrectly at the start
205 of a new segment.
206
207 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
208
209 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
210 gives some modest performance improvement in 8-bit mode.
211
212 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
213 be expressed in certain programming languages using Universal Character
214 Names.
215
216 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
217 which points out that "non characters" are not "characters that may not
218 appear in Unicode strings" but rather "characters that are reserved for
219 internal use and have only local meaning".
220
221 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
222 there was a conditional group that depended on an assertion, if the
223 assertion was false, the callout that immediately followed the alternation
224 in the condition was skipped when pcre_exec() was used for matching.
225
226 23. Allow an explicit callout to be inserted before an assertion that is the
227 condition for a conditional group, for compatibility with automatic
228 callouts, which always insert a callout at this point.
229
230 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
231 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
232
233 25. (*PRUNE) is now supported by the JIT compiler.
234
235 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
236
237 27. Fix the case where there are two or more SKIPs with arguments that may be
238 ignored.
239
240 28. (*SKIP) is now supported by the JIT compiler.
241
242 29. (*THEN) is now supported by the JIT compiler.
243
244 30. Update RunTest with additional test selector options.
245
246 31. The way PCRE handles backtracking verbs has been changed in two ways.
247
248 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
249 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
250 backtracking. In some cases this makes it more Perl-compatible, but Perl's
251 rather obscure rules do not always do the same thing.
252
253 (2) Previously, backtracking verbs were confined within assertions. This is
254 no longer the case for positive assertions, except for (*ACCEPT). Again,
255 this sometimes improves Perl compatibility, and sometimes does not.
256
257 32. A number of tests that were in test 2 because Perl did things differently
258 have been moved to test 1, because either Perl or PCRE has changed, and
259 these tests are now compatible.
260
261 32. Backtracking control verbs are now handled in the same way in JIT and
262 interpreter.
263
264 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
265 contained a forward subroutine reference caused a compile error.
266
267 34. Auto-detect and optimize limited repetitions in JIT.
268
269 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
270 blocking (*UTF) etc.
271
272 36. In the interpreter, maximizing pattern repetitions for characters and
273 character types now use tail recursion, which reduces stack usage.
274
275 37. The value of the max lookbehind was not correctly preserved if a compiled
276 and saved regex was reloaded on a host of different endianness.
277
278 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
279 of the compiled pattern block, expand the flags field from 16 to 32 bits
280 because it was almost full.
281
282 39. Try madvise first before posix_madvise.
283
284 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
285 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
286 original change locked out all matches of empty strings. This has been
287 changed so that one match of an empty string per line is recognized.
288 Subsequent searches on the same line (for colouring or for --only-matching,
289 for example) do not recognize empty strings.
290
291 41. Applied a user patch to fix a number of spelling mistakes in comments.
292
293 42. Data lines longer than 65536 caused pcretest to crash.
294
295 43. Clarified the data type for length and startoffset arguments for pcre_exec
296 and pcre_dfa_exec in the function-specific man pages, where they were
297 explicitly stated to be in bytes, never having been updated. I also added
298 some clarification to the pcreapi man page.
299
300 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
301 a segmentation fault.
302
303
304 Version 8.32 30-November-2012
305 -----------------------------
306
307 1. Improved JIT compiler optimizations for first character search and single
308 character iterators.
309
310 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
311 Patch by Daniel Richard G.
312
313 3. Single character iterator optimizations in the JIT compiler.
314
315 4. Improved JIT compiler optimizations for character ranges.
316
317 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
318 Reported by Giuseppe D'Angelo.
319
320 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
321 of a line, was being set incorrectly in cases where .* appeared inside
322 atomic brackets at the start of a pattern, or where there was a subsequent
323 *PRUNE or *SKIP.
324
325 7. Improved instruction cache flush for POWER/PowerPC.
326 Patch by Daniel Richard G.
327
328 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
329 grep:
330
331 (a) There is now no limit to the number of patterns to be matched.
332
333 (b) An error is given if a pattern is too long.
334
335 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
336 are now supported.
337
338 (d) --exclude-from and --include-from (multiple use) have been added.
339
340 (e) Exclusions and inclusions now apply to all files and directories, not
341 just to those obtained from scanning a directory recursively.
342
343 (f) Multiple uses of -f and --file-list are now supported.
344
345 (g) In a Windows environment, the default for -d has been changed from
346 "read" (the GNU grep default) to "skip", because otherwise the presence
347 of a directory in the file list provokes an error.
348
349 (h) The documentation has been revised and clarified in places.
350
351 9. Improve the matching speed of capturing brackets.
352
353 10. Changed the meaning of \X so that it now matches a Unicode extended
354 grapheme cluster.
355
356 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
357 out POSIX threads when JIT support is configured.
358
359 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
360
361 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
362 structure could go wrong in environments where size_t is not the same size
363 as int.
364
365 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
366 set.
367
368 15. The EBCDIC support had decayed; later updates to the code had included
369 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
370 general tidy up of EBCDIC-related issues, and the documentation was also
371 not quite right. There is now a test that can be run on ASCII systems to
372 check some of the EBCDIC-related things (but is it not a full test).
373
374 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
375 in a small tidy to the code.
376
377 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
378
379 18. If the --only-matching (-o) option in pcregrep is specified multiple
380 times, each one causes appropriate output. For example, -o1 -o2 outputs the
381 substrings matched by the 1st and 2nd capturing parentheses. A separating
382 string can be specified by --om-separator (default empty).
383
384 19. Improving the first n character searches.
385
386 20. Turn case lists for horizontal and vertical white space into macros so that
387 they are defined only once.
388
389 21. This set of changes together give more compatible Unicode case-folding
390 behaviour for characters that have more than one other case when UCP
391 support is available.
392
393 (a) The Unicode property table now has offsets into a new table of sets of
394 three or more characters that are case-equivalent. The MultiStage2.py
395 script that generates these tables (the pcre_ucd.c file) now scans
396 CaseFolding.txt instead of UnicodeData.txt for character case
397 information.
398
399 (b) The code for adding characters or ranges of characters to a character
400 class has been abstracted into a generalized function that also handles
401 case-independence. In UTF-mode with UCP support, this uses the new data
402 to handle characters with more than one other case.
403
404 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
405 whose other case is greater than 256 are now correctly matched
406 caselessly. Previously, the high codepoint matched the low one, but not
407 vice versa.
408
409 (d) The processing of \h, \H, \v, and \ in character classes now makes use
410 of the new class addition function, using character lists defined as
411 macros alongside the case definitions of 20 above.
412
413 (e) Caseless back references now work with characters that have more than
414 one other case.
415
416 (f) General caseless matching of characters with more than one other case
417 is supported.
418
419 22. Unicode character properties were updated from Unicode 6.2.0
420
421 23. Improved CMake support under Windows. Patch by Daniel Richard G.
422
423 24. Add support for 32-bit character strings, and UTF-32
424
425 25. Major JIT compiler update (code refactoring and bugfixing).
426 Experimental Sparc 32 support is added.
427
428 26. Applied a modified version of Daniel Richard G's patch to create
429 pcre.h.generic and config.h.generic by "make" instead of in the
430 PrepareRelease script.
431
432 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
433 pcre_compile.c when checking for a zero character.
434
435 28. Introducing a native interface for JIT. Through this interface, the compiled
436 machine code can be directly executed. The purpose of this interface is to
437 provide fast pattern matching, so several sanity checks are not performed.
438 However, feature tests are still performed. The new interface provides
439 1.4x speedup compared to the old one.
440
441 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
442 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
443 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
444
445 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
446 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
447 the "old" RFC 2279). Instead, it ended up passing a negative length to
448 pcre_exec().
449
450 31. Add support for GCC's visibility feature to hide internal functions.
451
452 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
453 "unknown -C option" after outputting 0 or 1.
454
455 33. There is now support for generating a code coverage report for the test
456 suite in environments where gcc is the compiler and lcov is installed. This
457 is mainly for the benefit of the developers.
458
459 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
460 unaddressable using valgrind annotations, allowing valgrind to detect
461 invalid memory accesses. This is mainly for the benefit of the developers.
462
463 25. (*UTF) can now be used to start a pattern in any of the three libraries.
464
465 26. Give configure error if --enable-cpp but no C++ compiler found.
466
467
468 Version 8.31 06-July-2012
469 -------------------------
470
471 1. Fixing a wrong JIT test case and some compiler warnings.
472
473 2. Removed a bashism from the RunTest script.
474
475 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
476 to unsigned type, result still unsigned" that was given by an MS compiler
477 on encountering the code "-sizeof(xxx)".
478
479 4. Partial matching support is added to the JIT compiler.
480
481 5. Fixed several bugs concerned with partial matching of items that consist
482 of more than one character:
483
484 (a) /^(..)\1/ did not partially match "aba" because checking references was
485 done on an "all or nothing" basis. This also applied to repeated
486 references.
487
488 (b) \R did not give a hard partial match if \r was found at the end of the
489 subject.
490
491 (c) \X did not give a hard partial match after matching one or more
492 characters at the end of the subject.
493
494 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
495 a partial match for the string "\r".
496
497 (e) When newline was set to CRLF, the metacharacter "." did not recognize
498 a partial match for a CR character at the end of the subject string.
499
500 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
501 running pcretest, the text "(JIT)" added to the output whenever JIT is
502 actually used to run the match.
503
504 7. Individual JIT compile options can be set in pcretest by following -s+[+]
505 or /S+[+] with a digit between 1 and 7.
506
507 8. OP_NOT now supports any UTF character not just single-byte ones.
508
509 9. (*MARK) control verb is now supported by the JIT compiler.
510
511 10. The command "./RunTest list" lists the available tests without actually
512 running any of them. (Because I keep forgetting what they all are.)
513
514 11. Add PCRE_INFO_MAXLOOKBEHIND.
515
516 12. Applied a (slightly modified) user-supplied patch that improves performance
517 when the heap is used for recursion (compiled with --disable-stack-for-
518 recursion). Instead of malloc and free for each heap frame each time a
519 logical recursion happens, frames are retained on a chain and re-used where
520 possible. This sometimes gives as much as 30% improvement.
521
522 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
523 call.
524
525 14. As documented, (*COMMIT) is now confined to within a positive assertion.
526
527 15. It is now possible to link pcretest with libedit as an alternative to
528 libreadline.
529
530 16. (*COMMIT) control verb is now supported by the JIT compiler.
531
532 17. The Unicode data tables have been updated to Unicode 6.1.0.
533
534 18. Added --file-list option to pcregrep.
535
536 19. Added binary file support to pcregrep, including the -a, --binary-files,
537 -I, and --text options.
538
539 20. The madvise function is renamed for posix_madvise for QNX compatibility
540 reasons. Fixed by Giuseppe D'Angelo.
541
542 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
543
544 22. Changed the option for creating symbolic links for 16-bit man pages from
545 -s to -sf so that re-installing does not cause issues.
546
547 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
548
549 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
550 matches in certain environments (the workspace was not being correctly
551 retained). Also added to pcre_dfa_exec() a simple plausibility check on
552 some of the workspace data at the beginning of a restart.
553
554 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
555 was not doing so when it should - probably a typo introduced by SVN 528
556 (change 8.10/14).
557
558 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
559 \w+ when the character tables indicated that \x{c4} was a word character.
560 There were several related cases, all because the tests for doing a table
561 lookup were testing for characters less than 127 instead of 255.
562
563 27. If a pattern contains capturing parentheses that are not used in a match,
564 their slots in the ovector are set to -1. For those that are higher than
565 any matched groups, this happens at the end of processing. In the case when
566 there were back references that the ovector was too small to contain
567 (causing temporary malloc'd memory to be used during matching), and the
568 highest capturing number was not used, memory off the end of the ovector
569 was incorrectly being set to -1. (It was using the size of the temporary
570 memory instead of the true size.)
571
572 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
573 ovector size, it uses memory at the end of the block that it has got.
574
575 29. Check for an overlong MARK name and give an error at compile time. The
576 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
577
578 30. JIT compiler update.
579
580 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
581 Rill for the patch.
582
583 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
584
585 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
586
587 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
588 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
589 was enabled.
590
591 35. Improve JIT code generation for greedy plus quantifier.
592
593 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
594 1 to "aa" instead of to an empty string. The bug affected repeated groups
595 that could potentially match an empty string.
596
597 37. Optimizing single character iterators in JIT.
598
599 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
600 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
601 codepoints that are too big for the mode are faulted, and in a UTF mode,
602 disallowed codepoints are also faulted.
603
604 39. If PCRE was compiled with UTF support, in three places in the DFA
605 matcher there was code that should only have been obeyed in UTF mode, but
606 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
607 processing when bytes with values greater than 127 were present. In 16-bit
608 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
609 both cases the values are those that cannot be the first data item in a UTF
610 character. The three items that might have provoked this were recursions,
611 possessively repeated groups, and atomic groups.
612
613 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
614 and pcregrep, because some OS require shared objects to be explicitly
615 passed to ld, causing the link step to fail if they are not.
616
617 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
618 mode, patterns that started with \h* or \R* might be incorrectly matched.
619
620
621 Version 8.30 04-February-2012
622 -----------------------------
623
624 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
625 name is defined in ctype.h.
626
627 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
628 only in quite long subpatterns.
629
630 3. Removed the function pcre_info(), which has been obsolete and deprecated
631 since it was replaced by pcre_fullinfo() in February 2000.
632
633 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
634 match a (*MARK), and the match failed at the start of the subject, a
635 reference to memory before the start of the subject could occur. This bug
636 was introduced by fix 17 of release 8.21.
637
638 5. A reference to an unset group with zero minimum repetition was giving
639 totally wrong answers (in non-JavaScript-compatibility mode). For example,
640 /(another)?(\1?)test/ matched against "hello world test". This bug was
641 introduced in release 8.13.
642
643 6. Add support for 16-bit character strings (a large amount of work involving
644 many changes and refactorings).
645
646 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
647 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
648 from a file.
649
650 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
651 rounding is not applied in this particular case).
652
653 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
654 if they appear, or are escaped, in patterns.
655
656 10. Get rid of a number of -Wunused-but-set-variable warnings.
657
658 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
659 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
660 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
661 also returns the mark "x". This bug applied to capturing parentheses,
662 non-capturing parentheses, and atomic parentheses. It also applied to some
663 assertions.
664
665 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
666 information out of configure.ac instead of relying on pcre.h.generic, which
667 is not stored in the repository.
668
669 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
670 -lreadline.
671
672 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
673
674 15. Applied Graycode's patch to put the top-level frame on the stack rather
675 than the heap when not using the stack for recursion. This gives a
676 performance improvement in many cases when recursion is not deep.
677
678 16. Experimental code added to "pcretest -C" to output the stack frame size.
679
680
681 Version 8.21 12-Dec-2011
682 ------------------------
683
684 1. Updating the JIT compiler.
685
686 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
687 are added as well.
688
689 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
690 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
691 calling _pcre_jit_exec. Some extra comments are added.
692
693 4. (*MARK) settings inside atomic groups that do not contain any capturing
694 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
695 was introduced by change 18 for 8.20.
696
697 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
698 ECMA-262 standard.
699
700 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
701 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
702 This bug was probably introduced by change 9 of 8.13.
703
704 7. While fixing 6 above, I noticed that a number of other items were being
705 incorrectly rejected as "not fixed length". This arose partly because newer
706 opcodes had not been added to the fixed-length checking code. I have (a)
707 corrected the bug and added tests for these items, and (b) arranged for an
708 error to occur if an unknown opcode is encountered while checking for fixed
709 length instead of just assuming "not fixed length". The items that were
710 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
711 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
712 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
713
714 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
715 being incorrectly compiled and would have given unpredicatble results.
716
717 9. A possessively repeated subpattern with minimum repeat count greater than
718 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
719 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
720 the first (A) could occur when it should not.
721
722 10. Add a cast and remove a redundant test from the code.
723
724 11. JIT should use pcre_malloc/pcre_free for allocation.
725
726 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
727 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
728 is anything other than /usr, -L is still shown).
729
730 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
731
732 14. Perl does not support \N without a following name in a [] class; PCRE now
733 also gives an error.
734
735 15. If a forward reference was repeated with an upper limit of around 2000,
736 it caused the error "internal error: overran compiling workspace". The
737 maximum number of forward references (including repeats) was limited by the
738 internal workspace, and dependent on the LINK_SIZE. The code has been
739 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
740 the default depends on LINK_SIZE. There is a new upper limit (for safety)
741 of around 200,000 forward references. While doing this, I also speeded up
742 the filling in of repeated forward references.
743
744 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
745 incorrectly expecting the subject to contain another "a" after the start.
746
747 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
748 in the match, the SKIP should be ignored. This was not happening; instead
749 the SKIP was being treated as NOMATCH. For patterns such as
750 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
751 tested.
752
753 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
754 now much more compatible with Perl, in particular in cases where the result
755 is a non-match for a non-anchored pattern. For example, if
756 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
757 "m", where previously it did not return a name. A side effect of this
758 change is that for partial matches, the last encountered mark name is
759 returned, as for non matches. A number of tests that were previously not
760 Perl-compatible have been moved into the Perl-compatible test files. The
761 refactoring has had the pleasing side effect of removing one argument from
762 the match() function, thus reducing its stack requirements.
763
764 19. If the /S+ option was used in pcretest to study a pattern using JIT,
765 subsequent uses of /S (without +) incorrectly behaved like /S+.
766
767 21. Retrieve executable code size support for the JIT compiler and fixing
768 some warnings.
769
770 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
771 not work when the shorter character appeared right at the end of the
772 subject string.
773
774 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
775 systems.
776
777 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
778 output it when the /M option is used in pcretest.
779
780 25. The CheckMan script was not being included in the distribution. Also, added
781 an explicit "perl" to run Perl scripts from the PrepareRelease script
782 because this is reportedly needed in Windows.
783
784 26. If study data was being save in a file and studying had not found a set of
785 "starts with" bytes for the pattern, the data written to the file (though
786 never used) was taken from uninitialized memory and so caused valgrind to
787 complain.
788
789 27. Updated RunTest.bat as provided by Sheri Pierce.
790
791 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
792
793 29. Computation of memory usage for the table of capturing group names was
794 giving an unnecessarily large value.
795
796
797 Version 8.20 21-Oct-2011
798 ------------------------
799
800 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
801 a POSIX class. After further experiments with Perl, which convinced me that
802 Perl has bugs and confusions, a closing square bracket is no longer allowed
803 in a POSIX name. This bug also affected patterns with classes that started
804 with full stops.
805
806 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
807 captured substring, but while checking the failing first alternative,
808 substring 1 is temporarily captured. If the output vector supplied to
809 pcre_exec() was not big enough for this capture, the yield of the function
810 was still zero ("insufficient space for captured substrings"). This cannot
811 be totally fixed without adding another stack variable, which seems a lot
812 of expense for a edge case. However, I have improved the situation in cases
813 such as /(a)(b)x|abc/ matched against "abc", where the return code
814 indicates that fewer than the maximum number of slots in the ovector have
815 been set.
816
817 3. Related to (2) above: when there are more back references in a pattern than
818 slots in the output vector, pcre_exec() uses temporary memory during
819 matching, and copies in the captures as far as possible afterwards. It was
820 using the entire output vector, but this conflicts with the specification
821 that only 2/3 is used for passing back captured substrings. Now it uses
822 only the first 2/3, for compatibility. This is, of course, another edge
823 case.
824
825 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
826 main code base, and can be used by building with --enable-jit. When this is
827 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
828 runtime --no-jit option is given.
829
830 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
831 ovector, the return from the function was zero, implying that there were
832 other matches that did not fit. The correct "exactly full" value is now
833 returned.
834
835 6. If a subpattern that was called recursively or as a subroutine contained
836 (*PRUNE) or any other control that caused it to give a non-standard return,
837 invalid errors such as "Error -26 (nested recursion at the same subject
838 position)" or even infinite loops could occur.
839
840 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
841 computing the minimum length on reaching *ACCEPT, and so ended up with the
842 wrong value of 1 rather than 0. Further investigation indicates that
843 computing a minimum subject length in the presence of *ACCEPT is difficult
844 (think back references, subroutine calls), and so I have changed the code
845 so that no minimum is registered for a pattern that contains *ACCEPT.
846
847 8. If (*THEN) was present in the first (true) branch of a conditional group,
848 it was not handled as intended. [But see 16 below.]
849
850 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
851 Sheri Pierce.
852
853 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
854 the first byte in a match must be "a".
855
856 11. Change 17 for 8.13 increased the recursion depth for patterns like
857 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
858 pattern contains any instances of (*THEN). If it does not, the old
859 optimizations are restored. It would be nice to do this on a per-group
860 basis, but at the moment that is not feasible.
861
862 12. In some environments, the output of pcretest -C is CRLF terminated. This
863 broke RunTest's code that checks for the link size. A single white space
864 character after the value is now allowed for.
865
866 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
867 For "fr", it uses the Windows-specific input and output files.
868
869 14. If (*THEN) appeared in a group that was called recursively or as a
870 subroutine, it did not work as intended. [But see next item.]
871
872 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
873 pattern fragments (but not containing any | characters). If A and B are
874 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
875 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
876 D. In other words, Perl considers parentheses that do not contain any |
877 characters to be part of a surrounding alternative, whereas PCRE was
878 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
879 differently. PCRE now behaves in the same way as Perl, except in the case
880 of subroutine/recursion calls such as (?1) which have in any case always
881 been different (but PCRE had them first :-).
882
883 16. Related to 15 above: Perl does not treat the | in a conditional group as
884 creating alternatives. Such a group is treated in the same way as an
885 ordinary group without any | characters when processing (*THEN). PCRE has
886 been changed to match Perl's behaviour.
887
888 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
889 RunGrepTest script failed.
890
891 18. Change 22 for version 13 caused atomic groups to use more stack. This is
892 inevitable for groups that contain captures, but it can lead to a lot of
893 stack use in large patterns. The old behaviour has been restored for atomic
894 groups that do not contain any capturing parentheses.
895
896 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
897 suppress the check for a minimum subject length at run time. (If it was
898 given to pcre_exec() or pcre_dfa_exec() it did work.)
899
900 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
901 fail to work when decoding hex characters in data strings in EBCDIC
902 environments.
903
904 21. It appears that in at least one Mac OS environment, the isxdigit() function
905 is implemented as a macro that evaluates to its argument more than once,
906 contravening the C 90 Standard (I haven't checked a later standard). There
907 was an instance in pcretest which caused it to go wrong when processing
908 \x{...} escapes in subject strings. The has been rewritten to avoid using
909 things like p++ in the argument of isxdigit().
910
911
912 Version 8.13 16-Aug-2011
913 ------------------------
914
915 1. The Unicode data tables have been updated to Unicode 6.0.0.
916
917 2. Two minor typos in pcre_internal.h have been fixed.
918
919 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
920 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
921 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
922
923 4. There were a number of related bugs in the code for matching backrefences
924 caselessly in UTF-8 mode when codes for the characters concerned were
925 different numbers of bytes. For example, U+023A and U+2C65 are an upper
926 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
927 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
928 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
929 2-byte code at the end of the subject (it thought there wasn't enough data
930 left).
931
932 5. Comprehensive information about what went wrong is now returned by
933 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
934 as the output vector has at least 2 elements. The offset of the start of
935 the failing character and a reason code are placed in the vector.
936
937 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
938 now returned is for the first byte of the failing character, instead of the
939 last byte inspected. This is an incompatible change, but I hope it is small
940 enough not to be a problem. It makes the returned offset consistent with
941 pcre_exec() and pcre_dfa_exec().
942
943 7. pcretest now gives a text phrase as well as the error number when
944 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
945 failure, the offset and reason code are output.
946
947 8. When \R was used with a maximizing quantifier it failed to skip backwards
948 over a \r\n pair if the subsequent match failed. Instead, it just skipped
949 back over a single character (\n). This seems wrong (because it treated the
950 two characters as a single entity when going forwards), conflicts with the
951 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
952 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
953 has been changed.
954
955 9. Some internal refactoring has changed the processing so that the handling
956 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
957 time (the PCRE_DOTALL option was changed this way some time ago: version
958 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
959 which was always a bit of a fudge. It also means that there is one less
960 argument for the match() function, which reduces its stack requirements
961 slightly. This change also fixes an incompatibility with Perl: the pattern
962 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
963
964 10. More internal refactoring has drastically reduced the number of recursive
965 calls to match() for possessively repeated groups such as (abc)++ when
966 using pcre_exec().
967
968 11. While implementing 10, a number of bugs in the handling of groups were
969 discovered and fixed:
970
971 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
972 (a|)*(?1) gave a compile-time internal error.
973 ((a|)+)+ did not notice that the outer group could match an empty string.
974 (^a|^)+ was not marked as anchored.
975 (.*a|.*)+ was not marked as matching at start or after a newline.
976
977 12. Yet more internal refactoring has removed another argument from the match()
978 function. Special calls to this function are now indicated by setting a
979 value in a variable in the "match data" data block.
980
981 13. Be more explicit in pcre_study() instead of relying on "default" for
982 opcodes that mean there is no starting character; this means that when new
983 ones are added and accidentally left out of pcre_study(), testing should
984 pick them up.
985
986 14. The -s option of pcretest has been documented for ages as being an old
987 synonym of -m (show memory usage). I have changed it to mean "force study
988 for every regex", that is, assume /S for every regex. This is similar to -i
989 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
990 using it. It makes it easier to run collections of tests with and without
991 study enabled, and thereby test pcre_study() more easily. All the standard
992 tests are now run with and without -s (but some patterns can be marked as
993 "never study" - see 20 below).
994
995 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
996 restoration of the capturing data to the outer values was not happening
997 correctly.
998
999 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1000 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1001 pattern had matched an empty string, and so incorrectly returned a no
1002 match.
1003
1004 17. There was optimizing code for the last branch of non-capturing parentheses,
1005 and also for the obeyed branch of a conditional subexpression, which used
1006 tail recursion to cut down on stack usage. Unfortunately, now that there is
1007 the possibility of (*THEN) occurring in these branches, tail recursion is
1008 no longer possible because the return has to be checked for (*THEN). These
1009 two optimizations have therefore been removed. [But see 8.20/11 above.]
1010
1011 18. If a pattern containing \R was studied, it was assumed that \R always
1012 matched two bytes, thus causing the minimum subject length to be
1013 incorrectly computed because \R can also match just one byte.
1014
1015 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1016 was incorrectly computed.
1017
1018 20. If /S is present twice on a test pattern in pcretest input, it now
1019 *disables* studying, thereby overriding the use of -s on the command line
1020 (see 14 above). This is necessary for one or two tests to keep the output
1021 identical in both cases.
1022
1023 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1024 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1025
1026 22. When an atomic group that contained a capturing parenthesis was
1027 successfully matched, but the branch in which it appeared failed, the
1028 capturing was not being forgotten if a higher numbered group was later
1029 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1030 group 1 to "a", when in fact it should be unset. This applied to multi-
1031 branched capturing and non-capturing groups, repeated or not, and also to
1032 positive assertions (capturing in negative assertions does not happen
1033 in PCRE) and also to nested atomic groups.
1034
1035 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1036 subject after a captured substring, to make it easier to tell which of a
1037 number of identical substrings has been captured.
1038
1039 24. The way atomic groups are processed by pcre_exec() has been changed so that
1040 if they are repeated, backtracking one repetition now resets captured
1041 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1042 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1043 "aaa". Previously, it would have been "a". As part of this code
1044 refactoring, the way recursive calls are handled has also been changed.
1045
1046 25. If an assertion condition captured any substrings, they were not passed
1047 back unless some other capturing happened later. For example, if
1048 (?(?=(a))a) was matched against "a", no capturing was returned.
1049
1050 26. When studying a pattern that contained subroutine calls or assertions,
1051 the code for finding the minimum length of a possible match was handling
1052 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1053 group 1 called group 2 while simultaneously a separate group 2 called group
1054 1). A stack overflow occurred in this case. I have fixed this by limiting
1055 the recursion depth to 10.
1056
1057 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1058 Fortmann. This supports explicit test numbers on the command line, and has
1059 argument validation and error reporting.
1060
1061 28. An instance of \X with an unlimited repeat could fail if at any point the
1062 first character it looked at was a mark character.
1063
1064 29. Some minor code refactoring concerning Unicode properties and scripts
1065 should reduce the stack requirement of match() slightly.
1066
1067 30. Added the '=' option to pcretest to check the setting of unused capturing
1068 slots at the end of the pattern, which are documented as being -1, but are
1069 not included in the return count.
1070
1071 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1072 compiled something random. Now it gives a compile-time error (as does
1073 Perl).
1074
1075 32. A *MARK encountered during the processing of a positive assertion is now
1076 recorded and passed back (compatible with Perl).
1077
1078 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1079 had alternative anchored branches, the search for a second match in a line
1080 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1081 matched the line "0102" twice. The same bug affected patterns that started
1082 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1083 twice.
1084
1085 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1086 does, and because of capturing effects, quantifying parenthesized
1087 assertions may at times be useful. Quantifiers are now allowed for
1088 parenthesized assertions.
1089
1090 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1091
1092 36. \g was being checked for fancy things in a character class, when it should
1093 just be a literal "g".
1094
1095 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1096 appearance of a nested POSIX class supersedes an apparent external class.
1097 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1098 unescaped square brackets may also appear as part of class names. For
1099 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1100 more like Perl. (But see 8.20/1 above.)
1101
1102 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1103 was because it thought it was \N{name}, which is not supported).
1104
1105 39. Add minix to OS list not supporting the -S option in pcretest.
1106
1107 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1108 cannot analyze patterns in sufficient detail to catch mutual recursions
1109 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1110 subgroup is called recursively as a subpattern for a second time at the
1111 same position in the subject string. In previous releases this might have
1112 been caught by the recursion limit, or it might have run out of stack.
1113
1114 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1115 happen only once. PCRE was, however incorrectly giving a compile time error
1116 "recursive call could loop indefinitely" because it cannot analyze the
1117 pattern in sufficient detail. The compile time test no longer happens when
1118 PCRE is compiling a conditional subpattern, but actual runaway loops are
1119 now caught at runtime (see 40 above).
1120
1121 42. It seems that Perl allows any characters other than a closing parenthesis
1122 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1123 has been changed to be the same.
1124
1125 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1126 as not to get warnings when autogen.sh is called. Also changed
1127 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1128
1129 44. To help people who use pcregrep to scan files containing exceedingly long
1130 lines, the following changes have been made:
1131
1132 (a) The default value of the buffer size parameter has been increased from
1133 8K to 20K. (The actual buffer used is three times this size.)
1134
1135 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1136 PCRE is built.
1137
1138 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1139 to be set at run time.
1140
1141 (d) Numerical values in pcregrep options can be followed by K or M, for
1142 example --buffer-size=50K.
1143
1144 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1145 given and the return code is set to 2.
1146
1147 45. Add a pointer to the latest mark to the callout data block.
1148
1149 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1150 partial match of an empty string instead of no match. This was specific to
1151 the use of ".".
1152
1153 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1154 complete match instead of a partial match. This bug was dependent on both
1155 the PCRE_UTF8 and PCRE_DOTALL options being set.
1156
1157 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1158 starting byte set, because \b was not being ignored.
1159
1160
1161 Version 8.12 15-Jan-2011
1162 ------------------------
1163
1164 1. Fixed some typos in the markup of the man pages, and wrote a script that
1165 checks for such things as part of the documentation building process.
1166
1167 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1168 --match-limit and --recursion-limit options (added for 8.11). In
1169 particular, this made one of the standard tests fail. (The integer value
1170 went into the wrong half of a long int.)
1171
1172 3. If the --colour option was given to pcregrep with -v (invert match), it
1173 did strange things, either producing crazy output, or crashing. It should,
1174 of course, ignore a request for colour when reporting lines that do not
1175 match.
1176
1177 4. Another pcregrep bug caused similar problems if --colour was specified with
1178 -M (multiline) and the pattern match finished with a line ending.
1179
1180 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1181 matched in multiline mode, the following line was shown as part of the
1182 match. This seems wrong, so I have changed it.
1183
1184 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1185 the check for further matches in the same line (so they could be coloured)
1186 to overrun the end of the current line. If another match was found, it was
1187 incorrectly shown (and then shown again when found in the next line).
1188
1189 7. If pcregrep was compiled under Windows, there was a reference to the
1190 function pcregrep_exit() before it was defined. I am assuming this was
1191 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1192 reported by a user. I've moved the definition above the reference.
1193
1194
1195 Version 8.11 10-Dec-2010
1196 ------------------------
1197
1198 1. (*THEN) was not working properly if there were untried alternatives prior
1199 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1200 backtracked to try for "b" instead of moving to the next alternative branch
1201 at the same level (in this case, to look for "c"). The Perl documentation
1202 is clear that when (*THEN) is backtracked onto, it goes to the "next
1203 alternative in the innermost enclosing group".
1204
1205 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1206 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1207 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1208 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1209 (*THEN).
1210
1211 3. If \s appeared in a character class, it removed the VT character from
1212 the class, even if it had been included by some previous item, for example
1213 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1214 of \s, but is part of the POSIX "space" class.)
1215
1216 4. A partial match never returns an empty string (because you can always
1217 match an empty string at the end of the subject); however the checking for
1218 an empty string was starting at the "start of match" point. This has been
1219 changed to the "earliest inspected character" point, because the returned
1220 data for a partial match starts at this character. This means that, for
1221 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1222 (previously it gave "no match").
1223
1224 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1225 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1226 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1227 has an implication that the given string is incomplete (because a partial
1228 match is preferred over a full match). For this reason, these items now
1229 give a partial match in this situation. [Aside: previously, the one case
1230 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1231 match rather than a full match, which was wrong by the old rules, but is
1232 now correct.]
1233
1234 6. There was a bug in the handling of #-introduced comments, recognized when
1235 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1236 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1237 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1238 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1239 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1240 places in pcre_compile().
1241
1242 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1243 comments when looking ahead for named forward references to subpatterns,
1244 the only newline sequence it recognized was NL. It now handles newlines
1245 according to the set newline convention.
1246
1247 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1248 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1249 cater for a lack of strerror(). These oversights have been fixed.
1250
1251 9. Added --match-limit and --recursion-limit to pcregrep.
1252
1253 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1254
1255 11. When the -o option was used, pcregrep was setting a return code of 1, even
1256 when matches were found, and --line-buffered was not being honoured.
1257
1258 12. Added an optional parentheses number to the -o and --only-matching options
1259 of pcregrep.
1260
1261 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1262 can match an empty string. The code to do it in pcretest and pcredemo
1263 needed fixing:
1264
1265 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1266 only one byte after an empty string match just before CRLF (this case
1267 just got forgotten; "any" and "anycrlf" were OK).
1268
1269 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1270 mode when an empty string match preceded an ASCII character followed by
1271 a non-ASCII character. (The code for advancing by one character rather
1272 than one byte was nonsense.)
1273
1274 (c) The pcredemo.c sample program did not have any code at all to handle
1275 the cases when CRLF is a valid newline sequence.
1276
1277 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1278 as a starting offset was within the subject string. There is now a new
1279 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1280 negative or greater than the length of the string. In order to test this,
1281 pcretest is extended to allow the setting of negative starting offsets.
1282
1283 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1284 starting offset points to the beginning of a UTF-8 character was
1285 unnecessarily clumsy. I tidied it up.
1286
1287 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1288 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1289
1290 17. Nobody had reported that the --include_dir option, which was added in
1291 release 7.7 should have been called --include-dir (hyphen, not underscore)
1292 for compatibility with GNU grep. I have changed it to --include-dir, but
1293 left --include_dir as an undocumented synonym, and the same for
1294 --exclude-dir, though that is not available in GNU grep, at least as of
1295 release 2.5.4.
1296
1297 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1298 characters from a string of bytes) have been redefined so as not to use
1299 loops, in order to improve performance in some environments. At the same
1300 time, I abstracted some of the common code into auxiliary macros to save
1301 repetition (this should not affect the compiled code).
1302
1303 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1304 compile-time error is now given if \c is not followed by an ASCII
1305 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1306 different, and any byte value is allowed.)
1307
1308 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1309 START_OPTIMIZE option, which is now allowed at compile time - but just
1310 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1311 to pcregrep and other applications that have no direct access to PCRE
1312 options. The new /Y option in pcretest sets this option when calling
1313 pcre_compile().
1314
1315 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1316 back references. Groups containing recursive back references were forced to
1317 be atomic by that change, but in the case of named groups, the amount of
1318 memory required was incorrectly computed, leading to "Failed: internal
1319 error: code overflow". This has been fixed.
1320
1321 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1322 pcretest.c, to avoid build problems in some Borland environments.
1323
1324
1325 Version 8.10 25-Jun-2010
1326 ------------------------
1327
1328 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1329 THEN.
1330
1331 2. (*ACCEPT) was not working when inside an atomic group.
1332
1333 3. Inside a character class, \B is treated as a literal by default, but
1334 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1335 causes the error). The code is unchanged, but I tidied the documentation.
1336
1337 4. Inside a character class, PCRE always treated \R and \X as literals,
1338 whereas Perl faults them if its -w option is set. I have changed PCRE so
1339 that it faults them when PCRE_EXTRA is set.
1340
1341 5. Added support for \N, which always matches any character other than
1342 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1343
1344 6. When compiling pcregrep with newer versions of gcc which may have
1345 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1346 declared with attribute warn_unused_result" were given. Just casting the
1347 result to (void) does not stop the warnings; a more elaborate fudge is
1348 needed. I've used a macro to implement this.
1349
1350 7. Minor change to pcretest.c to avoid a compiler warning.
1351
1352 8. Added four artifical Unicode properties to help with an option to make
1353 \s etc use properties (see next item). The new properties are: Xan
1354 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1355
1356 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1357 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1358 this option. Modified pcretest to add /W to test this facility. Added
1359 REG_UCP to make it available via the POSIX interface.
1360
1361 10. Added --line-buffered to pcregrep.
1362
1363 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1364 studied, and the match started with a letter with a code point greater than
1365 127 whose first byte was different to the first byte of the other case of
1366 the letter, the other case of this starting letter was not recognized
1367 (#976).
1368
1369 12. If a pattern that was studied started with a repeated Unicode property
1370 test, for example, \p{Nd}+, there was the theoretical possibility of
1371 setting up an incorrect bitmap of starting bytes, but fortunately it could
1372 not have actually happened in practice until change 8 above was made (it
1373 added property types that matched character-matching opcodes).
1374
1375 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1376 possible starting bytes for non-anchored patterns.
1377
1378 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1379 \R, and also a number of cases that involve Unicode properties, both
1380 explicit and implicit when PCRE_UCP is set.
1381
1382 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1383 input, it could crash or give wrong results if characters with values
1384 greater than 0xc0 were present in the subject string. (Detail: it assumed
1385 UTF-8 input when processing these items.)
1386
1387 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1388 size_t is 64-bit (#991).
1389
1390 17. Added a check for running out of memory when PCRE is compiled with
1391 --disable-stack-for-recursion (#990).
1392
1393 18. If the last data line in a file for pcretest does not have a newline on
1394 the end, a newline was missing in the output.
1395
1396 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1397 less than 128) in its various bitmaps. However, there is a facility for
1398 generating tables according to the current locale when PCRE is compiled. It
1399 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1400 characters, are recognized by isspace() and therefore were getting set in
1401 these tables, and indeed these tables seem to approximate to ISO 8859. This
1402 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1403 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1404 which of course cannot start UTF-8 characters. I have changed the code so
1405 that only real ASCII characters (less than 128) and the correct starting
1406 bytes for UTF-8 encodings are set for characters greater than 127 when in
1407 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1408 altogether.)
1409
1410 20. Added the /T option to pcretest so as to be able to run tests with non-
1411 standard character tables, thus making it possible to include the tests
1412 used for 19 above in the standard set of tests.
1413
1414 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1415 reference to a subpattern the other side of a comment that contains an
1416 opening parenthesis caused either an internal compiling error, or a
1417 reference to the wrong subpattern.
1418
1419
1420 Version 8.02 19-Mar-2010
1421 ------------------------
1422
1423 1. The Unicode data tables have been updated to Unicode 5.2.0.
1424
1425 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1426 configured.
1427
1428 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1429 original author of that file, following a query about its status.
1430
1431 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1432 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1433
1434 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1435 quantifier applied to a forward-referencing subroutine call, could compile
1436 incorrect code or give the error "internal error: previously-checked
1437 referenced subpattern not found".
1438
1439 6. Both MS Visual Studio and Symbian OS have problems with initializing
1440 variables to point to external functions. For these systems, therefore,
1441 pcre_malloc etc. are now initialized to local functions that call the
1442 relevant global functions.
1443
1444 7. There were two entries missing in the vectors called coptable and poptable
1445 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1446 I've fixed the data, and added a kludgy way of testing at compile time that
1447 the lengths are correct (equal to the number of opcodes).
1448
1449 8. Following on from 7, I added a similar kludge to check the length of the
1450 eint vector in pcreposix.c.
1451
1452 9. Error texts for pcre_compile() are held as one long string to avoid too
1453 much relocation at load time. To find a text, the string is searched,
1454 counting zeros. There was no check for running off the end of the string,
1455 which could happen if a new error number was added without updating the
1456 string.
1457
1458 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1459
1460 11. \K was not working if it appeared in an atomic group or in a group that
1461 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1462 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1463 the assertion is positive, but not if it is negative.
1464
1465 12. Change 11 fortuitously reduced the size of the stack frame used in the
1466 "match()" function of pcre_exec.c by one pointer. Forthcoming
1467 implementation of support for (*MARK) will need an extra pointer on the
1468 stack; I have reserved it now, so that the stack frame size does not
1469 decrease.
1470
1471 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1472 item in branch that calls a recursion is a subroutine call - as in the
1473 second branch in the above example - was incorrectly given the compile-
1474 time error "recursive call could loop indefinitely" because pcre_compile()
1475 was not correctly checking the subroutine for matching a non-empty string.
1476
1477 14. The checks for overrunning compiling workspace could trigger after an
1478 overrun had occurred. This is a "should never occur" error, but it can be
1479 triggered by pathological patterns such as hundreds of nested parentheses.
1480 The checks now trigger 100 bytes before the end of the workspace.
1481
1482 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1483
1484
1485 Version 8.01 19-Jan-2010
1486 ------------------------
1487
1488 1. If a pattern contained a conditional subpattern with only one branch (in
1489 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1490 computed the wrong minimum data length (which is of course zero for such
1491 subpatterns). This could cause incorrect "no match" results.
1492
1493 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1494 the pattern is reset in the first branch, pcre_compile() failed with
1495 "internal error: code overflow at offset...". This happened only when
1496 the reset was to the original external option setting. (An optimization
1497 abstracts leading options settings into an external setting, which was the
1498 cause of this.)
1499
1500 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1501 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1502 assertion pattern did not match (meaning that the assertion was true), it
1503 was incorrectly treated as false if the SKIP had been reached during the
1504 matching. This also applied to assertions used as conditions.
1505
1506 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1507 assertion subpattern, including such a pattern used as a condition,
1508 unpredictable results occurred, instead of the error return
1509 PCRE_ERROR_DFA_UITEM.
1510
1511 5. The C++ GlobalReplace function was not working like Perl for the special
1512 situation when an empty string is matched. It now does the fancy magic
1513 stuff that is necessary.
1514
1515 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1516 removed. (These were left over from very, very early versions of PCRE.)
1517
1518 7. Some cosmetic changes to the code to make life easier when compiling it
1519 as part of something else:
1520
1521 (a) Change DEBUG to PCRE_DEBUG.
1522
1523 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1524 called "current" as "current_branch", to prevent a collision with the
1525 Linux macro when compiled as a kernel module.
1526
1527 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1528 prevent a collision with the Linux macro when compiled as a kernel
1529 module.
1530
1531 8. In pcre_compile() there are some checks for integer overflows that used to
1532 cast potentially large values to (double). This has been changed to that
1533 when building, a check for int64_t is made, and if it is found, it is used
1534 instead, thus avoiding the use of floating point arithmetic. (There is no
1535 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1536 double.
1537
1538 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1539 2005 (difference between two addresses compared to an unsigned value).
1540
1541 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1542 custom one, because of the following reported problem in Windows:
1543
1544 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1545 under Win32.
1546 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1547 therefore missing the function definition.
1548 - The compiler thus generates a "C" signature for the test function.
1549 - The linker fails to find the "C" function.
1550 - PCRE fails to configure if asked to do so against libbz2.
1551
1552 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1553 messages were output:
1554
1555 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1556 rerunning libtoolize, to keep the correct libtool macros in-tree.
1557 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1558
1559 I have done both of these things.
1560
1561 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1562 most of the time, it *can* run out if it is given a pattern that contains a
1563 runaway infinite recursion. I updated the discussion in the pcrestack man
1564 page.
1565
1566 13. Now that we have gone to the x.xx style of version numbers, the minor
1567 version may start with zero. Using 08 or 09 is a bad idea because users
1568 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1569 interpreted as invalid octal numbers. I've updated the previous comment in
1570 configure.ac, and also added a check that gives an error if 08 or 09 are
1571 used.
1572
1573 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1574 causing partial matching to fail when the end of the subject matched \W
1575 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1576
1577 15. There were some discrepancies between the declarations in pcre_internal.h
1578 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1579 their definitions. The declarations used "const uschar *" and the
1580 definitions used USPTR. Even though USPTR is normally defined as "const
1581 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1582 reported that: "This difference in casting confuses some C++ compilers, for
1583 example, SunCC recognizes above declarations as different functions and
1584 generates broken code for hbpcre." I have changed the declarations to use
1585 USPTR.
1586
1587 16. GNU libtool is named differently on some systems. The autogen.sh script now
1588 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1589 (FreeBSD).
1590
1591 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1592 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1593 comment: "Figure out how to create a longlong from a string: strtoll and
1594 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1595 instance, but it only takes 2 args instead of 3!"
1596
1597 18. A subtle bug concerned with back references has been fixed by a change of
1598 specification, with a corresponding code fix. A pattern such as
1599 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1600 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1601 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1602 same bug. Such groups have to be quantified to be useful, or contained
1603 inside another quantified group. (If there's no repetition, the reference
1604 can never match.) The problem arises because, having left the group and
1605 moved on to the rest of the pattern, a later failure that backtracks into
1606 the group uses the captured value from the final iteration of the group
1607 rather than the correct earlier one. I have fixed this in PCRE by forcing
1608 any group that contains a reference to itself to be an atomic group; that
1609 is, there cannot be any backtracking into it once it has completed. This is
1610 similar to recursive and subroutine calls.
1611
1612
1613 Version 8.00 19-Oct-09
1614 ----------------------
1615
1616 1. The table for translating pcre_compile() error codes into POSIX error codes
1617 was out-of-date, and there was no check on the pcre_compile() error code
1618 being within the table. This could lead to an OK return being given in
1619 error.
1620
1621 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1622 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1623 in a Windows environment.
1624
1625 3. The pcregrep --count option prints the count for each file even when it is
1626 zero, as does GNU grep. However, pcregrep was also printing all files when
1627 --files-with-matches was added. Now, when both options are given, it prints
1628 counts only for those files that have at least one match. (GNU grep just
1629 prints the file name in this circumstance, but including the count seems
1630 more useful - otherwise, why use --count?) Also ensured that the
1631 combination -clh just lists non-zero counts, with no names.
1632
1633 4. The long form of the pcregrep -F option was incorrectly implemented as
1634 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1635 but it seems right to fix it, and I didn't think it was worth preserving
1636 the old behaviour.
1637
1638 5. The command line items --regex=pattern and --regexp=pattern were not
1639 recognized by pcregrep, which required --regex pattern or --regexp pattern
1640 (with a space rather than an '='). The man page documented the '=' forms,
1641 which are compatible with GNU grep; these now work.
1642
1643 6. No libpcreposix.pc file was created for pkg-config; there was just
1644 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1645
1646 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1647 when UCP support is not needed, by modifying the Python script that
1648 generates it from Unicode data files. This should not matter if the module
1649 is correctly used as a library, but I received one complaint about 50K of
1650 unwanted data. My guess is that the person linked everything into his
1651 program rather than using a library. Anyway, it does no harm.
1652
1653 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1654 was a minimum greater than 1 for a wide character in a possessive
1655 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1656 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1657 character. Chaos in the form of incorrect output or a compiling loop could
1658 result.
1659
1660 9. The restrictions on what a pattern can contain when partial matching is
1661 requested for pcre_exec() have been removed. All patterns can now be
1662 partially matched by this function. In addition, if there are at least two
1663 slots in the offset vector, the offset of the earliest inspected character
1664 for the match and the offset of the end of the subject are set in them when
1665 PCRE_ERROR_PARTIAL is returned.
1666
1667 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1668 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1669 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1670 and may be more useful for multi-segment matching.
1671
1672 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1673 used to be given if ever the end of the subject was reached; now it is
1674 given only if matching could not proceed because another character was
1675 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1676 string "Z", which now yields "no match" instead of "partial match". In the
1677 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1678 final character ended with (*FAIL).
1679
1680 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1681 if the pattern had a "must contain" character that was already found in the
1682 earlier partial match, unless partial matching was again requested. For
1683 example, with the pattern /dog.(body)?/, the "must contain" character is
1684 "g". If the first part-match was for the string "dog", restarting with
1685 "sbody" failed. This bug has been fixed.
1686
1687 13. The string returned by pcre_dfa_exec() after a partial match has been
1688 changed so that it starts at the first inspected character rather than the
1689 first character of the match. This makes a difference only if the pattern
1690 starts with a lookbehind assertion or \b or \B (\K is not supported by
1691 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1692 matching functions compatible, and I think it's the right thing to do.
1693
1694 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1695 so that the demonstration program is easily available in environments where
1696 PCRE has not been installed from source.
1697
1698 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1699 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1700 library.
1701
1702 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1703 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1704 is not the first non-POSIX option to be added. Clearly some people find
1705 these options useful.
1706
1707 17. If a caller to the POSIX matching function regexec() passes a non-zero
1708 value for nmatch with a NULL value for pmatch, the value of
1709 nmatch is forced to zero.
1710
1711 18. RunGrepTest did not have a test for the availability of the -u option of
1712 the diff command, as RunTest does. It now checks in the same way as
1713 RunTest, and also checks for the -b option.
1714
1715 19. If an odd number of negated classes containing just a single character
1716 interposed, within parentheses, between a forward reference to a named
1717 subpattern and the definition of the subpattern, compilation crashed with
1718 an internal error, complaining that it could not find the referenced
1719 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1720 [The bug was that it was starting one character too far in when skipping
1721 over the character class, thus treating the ] as data rather than
1722 terminating the class. This meant it could skip too much.]
1723
1724 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1725 /g option in pcretest when the pattern contains \K, which makes it possible
1726 to have an empty string match not at the start, even when the pattern is
1727 anchored. Updated pcretest and pcredemo to use this option.
1728
1729 21. If the maximum number of capturing subpatterns in a recursion was greater
1730 than the maximum at the outer level, the higher number was returned, but
1731 with unset values at the outer level. The correct (outer level) value is
1732 now given.
1733
1734 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1735 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1736 make it do so. The string so far is captured, making this feature
1737 compatible with Perl.
1738
1739 23. The tests have been re-organized, adding tests 11 and 12, to make it
1740 possible to check the Perl 5.10 features against Perl 5.10.
1741
1742 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1743 pattern matches a fixed length string. PCRE did not allow this; now it
1744 does. Neither allows recursion.
1745
1746 25. I finally figured out how to implement a request to provide the minimum
1747 length of subject string that was needed in order to match a given pattern.
1748 (It was back references and recursion that I had previously got hung up
1749 on.) This code has now been added to pcre_study(); it finds a lower bound
1750 to the length of subject needed. It is not necessarily the greatest lower
1751 bound, but using it to avoid searching strings that are too short does give
1752 some useful speed-ups. The value is available to calling programs via
1753 pcre_fullinfo().
1754
1755 26. While implementing 25, I discovered to my embarrassment that pcretest had
1756 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1757 study optimizations had never been tested with that matching function.
1758 Oops. What is worse, even when it was passed study data, there was a bug in
1759 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1760 were also very few tests of studied patterns with pcre_dfa_exec().
1761
1762 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1763 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1764 on the other side of the coin, they are no longer allowed to have different
1765 names, because these cannot be distinguished in PCRE, and this has caused
1766 confusion. (This is a difference from Perl.)
1767
1768 28. When duplicate subpattern names are present (necessarily with different
1769 numbers, as required by 27 above), and a test is made by name in a
1770 conditional pattern, either for a subpattern having been matched, or for
1771 recursion in such a pattern, all the associated numbered subpatterns are
1772 tested, and the overall condition is true if the condition is true for any
1773 one of them. This is the way Perl works, and is also more like the way
1774 testing by number works.
1775
1776
1777 Version 7.9 11-Apr-09
1778 ---------------------
1779
1780 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1781 (pcretest), all targets were linked against these libraries. This included
1782 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1783 libraries. This caused unwanted dependencies to be created. This problem
1784 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1785 pcretest is linked with readline.
1786
1787 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1788 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1789 moved outside it again, because FALSE and TRUE are already defined in AIX,
1790 but BOOL is not.
1791
1792 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1793 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1794
1795 4. The pcregrep documentation said spaces were inserted as well as colons (or
1796 hyphens) following file names and line numbers when outputting matching
1797 lines. This is not true; no spaces are inserted. I have also clarified the
1798 wording for the --colour (or --color) option.
1799
1800 5. In pcregrep, when --colour was used with -o, the list of matching strings
1801 was not coloured; this is different to GNU grep, so I have changed it to be
1802 the same.
1803
1804 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1805 each matching line was coloured. Now it goes on to look for further matches
1806 of any of the test patterns, which is the same behaviour as GNU grep.
1807
1808 7. A pattern that could match an empty string could cause pcregrep to loop; it
1809 doesn't make sense to accept an empty string match in pcregrep, so I have
1810 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1811 seems to be how GNU grep behaves. [But see later change 40 for release
1812 8.33.]
1813
1814 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1815 start or after a newline", because the conditional assertion was not being
1816 correctly handled. The rule now is that both the assertion and what follows
1817 in the first alternative must satisfy the test.
1818
1819 9. If auto-callout was enabled in a pattern with a conditional group whose
1820 condition was an assertion, PCRE could crash during matching, both with
1821 pcre_exec() and pcre_dfa_exec().
1822
1823 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1824 used for matching.
1825
1826 11. Unicode property support in character classes was not working for
1827 characters (bytes) greater than 127 when not in UTF-8 mode.
1828
1829 12. Added the -M command line option to pcretest.
1830
1831 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1832
1833 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1834
1835 16. Added comments and documentation about mis-use of no_arg in the C++
1836 wrapper.
1837
1838 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1839 from Martin Jerabek that uses macro names for all relevant character and
1840 string constants.
1841
1842 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1843 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1844 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1845 these, but not everybody uses configure.
1846
1847 19. A conditional group that had only one branch was not being correctly
1848 recognized as an item that could match an empty string. This meant that an
1849 enclosing group might also not be so recognized, causing infinite looping
1850 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1851 with the subject "ab", where knowledge that the repeated group can match
1852 nothing is needed in order to break the loop.
1853
1854 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1855 exec(), but without supplying a callout function, matching went wrong.
1856
1857 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1858 leak if the size of the offset vector was greater than 30. When the vector
1859 is smaller, the saved offsets during recursion go onto a local stack
1860 vector, but for larger vectors malloc() is used. It was failing to free
1861 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1862 error, in fact).
1863
1864 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1865 heapframe that is used only when UTF-8 support is enabled. This caused no
1866 problem, but was untidy.
1867
1868 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1869 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1870 included within another project.
1871
1872 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1873 slightly modified by me:
1874
1875 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1876 not building pcregrep.
1877
1878 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1879 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1880
1881 25. Forward references, both numeric and by name, in patterns that made use of
1882 duplicate group numbers, could behave incorrectly or give incorrect errors,
1883 because when scanning forward to find the reference group, PCRE was not
1884 taking into account the duplicate group numbers. A pattern such as
1885 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1886
1887 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1888 the feature of a custom pointer more persuasive (as requested by a user).
1889
1890 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1891 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1892 pre-defined.
1893
1894 28. Added support for (*UTF8) at the start of a pattern.
1895
1896 29. Arrange for flags added by the "release type" setting in CMake to be shown
1897 in the configuration summary.
1898
1899
1900 Version 7.8 05-Sep-08
1901 ---------------------
1902
1903 1. Replaced UCP searching code with optimized version as implemented for Ad
1904 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1905 stage table and inline lookup instead of a function, giving speed ups of 2
1906 to 5 times on some simple patterns that I tested. Permission was given to
1907 distribute the MultiStage2.py script that generates the tables (it's not in
1908 the tarball, but is in the Subversion repository).
1909
1910 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1911 scripts.
1912
1913 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1914 a group with a zero qualifier. The result of the study could be incorrect,
1915 or the function might crash, depending on the pattern.
1916
1917 4. Caseless matching was not working for non-ASCII characters in back
1918 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1919 It now works when Unicode Property Support is available.
1920
1921 5. In pcretest, an escape such as \x{de} in the data was always generating
1922 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1923 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1924 truncation.
1925
1926 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1927
1928 7. Added two (int) casts to pcregrep when printing the difference of two
1929 pointers, in case they are 64-bit values.
1930
1931 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1932 test 2 if it fails.
1933
1934 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1935 and a #define of that name to empty if it is not externally set. This is to
1936 allow users of MSVC to set it if necessary.
1937
1938 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1939 the convenience functions in the pcre_get.c source file.
1940
1941 11. An option change at the start of a pattern that had top-level alternatives
1942 could cause overwriting and/or a crash. This command provoked a crash in
1943 some environments:
1944
1945 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1946
1947 This potential security problem was recorded as CVE-2008-2371.
1948
1949 12. For a pattern where the match had to start at the beginning or immediately
1950 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1951 pcre_dfa_exec() could read past the end of the passed subject if there was
1952 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1953 pcretest so that it places the subject at the end of its malloc-ed buffer.
1954
1955 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1956 exec() might read past the end of the data buffer in UTF-8 mode.
1957
1958 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1959 the data contained the byte 0x85 as part of a UTF-8 character within its
1960 first line. This applied both to normal and DFA matching.
1961
1962 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1963 /^[^d]*?$/8 failed to match "abc".
1964
1965 16. Added a missing copyright notice to pcrecpp_internal.h.
1966
1967 17. Make it more clear in the documentation that values returned from
1968 pcre_exec() in ovector are byte offsets, not character counts.
1969
1970 18. Tidied a few places to stop certain compilers from issuing warnings.
1971
1972 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1973 supplied by Stefan Weber. I made a further small update for 7.8 because
1974 there is a change of source arrangements: the pcre_searchfuncs.c module is
1975 replaced by pcre_ucd.c.
1976
1977
1978 Version 7.7 07-May-08
1979 ---------------------
1980
1981 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1982 a string to a long long, pretend we don't even have a long long." This is
1983 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1984
1985 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1986 pre-7.6 versions, which defined a global no_arg variable instead of putting
1987 it in the RE class. (See also #8 below.)
1988
1989 3. Remove a line of dead code, identified by coverity and reported by Nuno
1990 Lopes.
1991
1992 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1993
1994 (1) The include/exclude patterns were being applied to the whole pathnames
1995 of files, instead of just to the final components.
1996
1997 (2) If there was more than one level of directory, the subdirectories were
1998 skipped unless they satisfied the include/exclude conditions. This is
1999 inconsistent with GNU grep (and could even be seen as contrary to the
2000 pcregrep specification - which I improved to make it absolutely clear).
2001 The action now is always to scan all levels of directory, and just
2002 apply the include/exclude patterns to regular files.
2003
2004 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2005 --exclude_dir in the tests to avoid scanning .svn directories.
2006
2007 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2008 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2009 doesn't support NULs in patterns.
2010
2011 7. Added some missing "const"s to declarations of static tables in
2012 pcre_compile.c and pcre_dfa_exec.c.
2013
2014 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2015 caused by fix #2 above. (Subsequently also a second patch to fix the
2016 first patch. And a third patch - this was a messy problem.)
2017
2018 9. Applied Craig's patch to remove the use of push_back().
2019
2020 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2021 matching function regexec().
2022
2023 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2024 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2025 references. PCRE supports relative numbers with this syntax (I don't think
2026 Oniguruma does).
2027
2028 12. Previously, a group with a zero repeat such as (...){0} was completely
2029 omitted from the compiled regex. However, this means that if the group
2030 was called as a subroutine from elsewhere in the pattern, things went wrong
2031 (an internal error was given). Such groups are now left in the compiled
2032 pattern, with a new opcode that causes them to be skipped at execution
2033 time.
2034
2035 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2036 to the way PCRE behaves:
2037
2038 (a) A lone ] character is dis-allowed (Perl treats it as data).
2039
2040 (b) A back reference to an unmatched subpattern matches an empty string
2041 (Perl fails the current match path).
2042
2043 (c) A data ] in a character class must be notated as \] because if the
2044 first data character in a class is ], it defines an empty class. (In
2045 Perl it is not possible to have an empty class.) The empty class []
2046 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2047 The negative empty class [^] matches any one character, independently
2048 of the DOTALL setting.
2049
2050 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2051 non-existent subpattern following a character class starting with ']' and
2052 containing () gave an internal compiling error instead of "reference to
2053 non-existent subpattern". Fortunately, when the pattern did exist, the
2054 compiled code was correct. (When scanning forwards to check for the
2055 existence of the subpattern, it was treating the data ']' as terminating
2056 the class, so got the count wrong. When actually compiling, the reference
2057 was subsequently set up correctly.)
2058
2059 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2060 it was being rejected as not supported by pcre_dfa_exec(), even though
2061 other assertions are supported. I have made pcre_dfa_exec() support
2062 (*FAIL).
2063
2064 16. The implementation of 13c above involved the invention of a new opcode,
2065 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2066 cannot be changed at match time, I realized I could make a small
2067 improvement to matching performance by compiling OP_ALLANY instead of
2068 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2069 on the OP_ANY path.
2070
2071 17. Compiling pcretest on Windows with readline support failed without the
2072 following two fixes: (1) Make the unistd.h include conditional on
2073 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2074
2075 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2076 ncurses library to be included for pcretest when ReadLine support is
2077 requested, but also to allow for it to be overridden. This patch came from
2078 Daniel Bergström.
2079
2080 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2081 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2082 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2083 spotting this.
2084
2085
2086 Version 7.6 28-Jan-08
2087 ---------------------
2088
2089 1. A character class containing a very large number of characters with
2090 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2091 overflow.
2092
2093 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2094 HAVE_LONG_LONG is not defined.
2095
2096 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2097 bring them up to date and include new features. This patch includes:
2098
2099 - Fixed PH's badly added libz and libbz2 support.
2100 - Fixed a problem with static linking.
2101 - Added pcredemo. [But later removed - see 7 below.]
2102 - Fixed dftables problem and added an option.
2103 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2104 HAVE_LONG_LONG.
2105 - Added readline support for pcretest.
2106 - Added an listing of the option settings after cmake has run.
2107
2108 4. A user submitted a patch to Makefile that makes it easy to create
2109 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2110 Makefile.am that cause it to include this special target, without
2111 affecting anything else. Note that the same mingw target plus all
2112 the other distribution libraries and programs are now supported
2113 when configuring with CMake (see 6 below) instead of with
2114 Configure/Make.
2115
2116 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2117 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2118 exported in the Windows port". It has not yet been confirmed that the patch
2119 solves the problem, but it does no harm.
2120
2121 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2122 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2123 with CMake, and also correct the comment about stack recursion.
2124
2125 7. Remove the automatic building of pcredemo from the ./configure system and
2126 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2127 of a program that users should build themselves after PCRE is installed, so
2128 building it automatically is not really right. What is more, it gave
2129 trouble in some build environments.
2130
2131 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2132
2133
2134 Version 7.5 10-Jan-08
2135 ---------------------
2136
2137 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2138 values in parens when parsing an RE using the C++ wrapper."
2139
2140 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2141 Characters greater than 255 were excluded from the class instead of being
2142 included.
2143
2144 3. The same bug as (2) above applied to negated POSIX classes such as
2145 [:^space:].
2146
2147 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2148 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2149 I have changed it.
2150
2151 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2152 first named subpattern) and a construct such as (?&a) would reference the
2153 first named subpattern whose name started with "a" (in other words, the
2154 length check was missing). Both these problems are fixed. "Subpattern name
2155 expected" is now given for (?&) (a zero-length name), and this patch also
2156 makes it give the same error for \k'' (previously it complained that that
2157 was a reference to a non-existent subpattern).
2158
2159 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2160 this is right because (?- can be followed by option settings as well as by
2161 digits. I have, however, made the messages clearer.
2162
2163 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2164 than the number used in the conditional) now cause a compile-time error.
2165 This is actually not compatible with Perl, which accepts such patterns, but
2166 treats the conditional as always being FALSE (as PCRE used to), but it
2167 seems to me that giving a diagnostic is better.
2168
2169 8. Change "alphameric" to the more common word "alphanumeric" in comments
2170 and messages.
2171
2172 9. Fix two occurrences of "backslash" in comments that should have been
2173 "backspace".
2174
2175 10. Remove two redundant lines of code that can never be obeyed (their function
2176 was moved elsewhere).
2177
2178 11. The program that makes PCRE's Unicode character property table had a bug
2179 which caused it to generate incorrect table entries for sequences of
2180 characters that have the same character type, but are in different scripts.
2181 It amalgamated them into a single range, with the script of the first of
2182 them. In other words, some characters were in the wrong script. There were
2183 thirteen such cases, affecting characters in the following ranges:
2184
2185 U+002b0 - U+002c1
2186 U+0060c - U+0060d
2187 U+0061e - U+00612
2188 U+0064b - U+0065e
2189 U+0074d - U+0076d
2190 U+01800 - U+01805
2191 U+01d00 - U+01d77
2192 U+01d9b - U+01dbf
2193 U+0200b - U+0200f
2194 U+030fc - U+030fe
2195 U+03260 - U+0327f
2196 U+0fb46 - U+0fbb1
2197 U+10450 - U+1049d
2198
2199 12. The -o option (show only the matching part of a line) for pcregrep was not
2200 compatible with GNU grep in that, if there was more than one match in a
2201 line, it showed only the first of them. It now behaves in the same way as
2202 GNU grep.
2203
2204 13. If the -o and -v options were combined for pcregrep, it printed a blank
2205 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2206 does the same. The return code can be used to tell if there were any
2207 non-matching lines.
2208
2209 14. Added --file-offsets and --line-offsets to pcregrep.
2210
2211 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2212 infinitely looping recursion. The bug was that positive lookaheads were not
2213 being skipped when checking for a possible empty match (negative lookaheads
2214 and both kinds of lookbehind were skipped).
2215
2216 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2217 inclusion of <windows.h> to before rather than after the definition of
2218 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2219
2220 17. Specifying a possessive quantifier with a specific limit for a Unicode
2221 character property caused pcre_compile() to compile bad code, which led at
2222 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2223 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2224 caused the error; without that there was no problem.
2225
2226 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2227
2228 19. Added --enable-pcretest-libreadline.
2229
2230 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2231 RE::GlobalReplace(). As a result, the number of replacements returned was
2232 double what it should be. I removed one of the increments, but Craig sent a
2233 later patch that removed the other one (the right fix) and added unit tests
2234 that check the return values (which was not done before).
2235
2236 21. Several CMake things:
2237
2238 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2239 the names libpcre and libpcreposix, not just pcre and pcreposix.
2240
2241 (2) The above change means that pcretest and pcregrep are now correctly
2242 linked with the newly-built libraries, not previously installed ones.
2243
2244 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2245
2246 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2247 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2248 UTF-8 newline character). The key issue is that the pattern starts .*;
2249 this means that the match must be either at the beginning, or after a
2250 newline. The bug was in the code for advancing after a failed match and
2251 checking that the new position followed a newline. It was not taking
2252 account of UTF-8 characters correctly.
2253
2254 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2255 character classes. PCRE was not treating the sequence [:...:] as a
2256 character class unless the ... were all letters. Perl, however, seems to
2257 allow any characters between [: and :], though of course it rejects as
2258 unknown any "names" that contain non-letters, because all the known class
2259 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2260 for example, whereas PCRE did not - it did not recognize a POSIX character
2261 class. This seemed a bit dangerous, so the code has been changed to be
2262 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2263 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2264 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2265 Perl does, and where it didn't before.
2266
2267 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2268 Windows environments %n is disabled by default.
2269
2270
2271 Version 7.4 21-Sep-07
2272 ---------------------
2273
2274 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2275 means that a class such as [\s] counted as "explicit reference to CR or
2276 LF". That isn't really right - the whole point of the change was to try to
2277 help when there was an actual mention of one of the two characters. So now
2278 the change happens only if \r or \n (or a literal CR or LF) character is
2279 encountered.
2280
2281 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2282 of both had grown to the point where there were only 3 bits left.
2283 Fortunately, there was spare space in the data structure, and so I have
2284 moved the internal flags into a new 16-bit field to free up more option
2285 bits.
2286
2287 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2288 but did not set the internal JCHANGED flag - either of these is enough to
2289 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2290 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2291 start sets both bits.
2292
2293 4. Added options (at build time, compile time, exec time) to change \R from
2294 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2295
2296 5. doc/pcresyntax.html was missing from the distribution.
2297
2298 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2299 compatibility, even though it is no longer used.
2300
2301 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2302 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2303 windows.h file is present (where different names are used). [This was
2304 reversed later after testing - see 16 below.]
2305
2306 8. Changed all #include <config.h> to #include "config.h". There were also
2307 some further <pcre.h> cases that I changed to "pcre.h".
2308
2309 9. When pcregrep was used with the --colour option, it missed the line ending
2310 sequence off the lines that it output.
2311
2312 10. It was pointed out to me that arrays of string pointers cause lots of
2313 relocations when a shared library is dynamically loaded. A technique of
2314 using a single long string with a table of offsets can drastically reduce
2315 these. I have refactored PCRE in four places to do this. The result is
2316 dramatic:
2317
2318 Originally: 290
2319 After changing UCP table: 187
2320 After changing error message table: 43
2321 After changing table of "verbs" 36
2322 After changing table of Posix names 22
2323
2324 Thanks to the folks working on Gregex for glib for this insight.
2325
2326 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2327 unicode-properties was also set.
2328
2329 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2330
2331 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2332 checked only for CRLF.
2333
2334 14. Added casts to pcretest.c to avoid compiler warnings.
2335
2336 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2337
2338 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2339 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2340 entirely. This removes changes made in 7 above.
2341
2342 17. The CMake files have been updated, and there is now more information about
2343 building with CMake in the NON-UNIX-USE document.
2344
2345
2346 Version 7.3 28-Aug-07
2347 ---------------------
2348
2349 1. In the rejigging of the build system that eventually resulted in 7.1, the
2350 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2351 brackets there is not right, since it causes compilers to look for an
2352 installed pcre.h, not the version that is in the source that is being
2353 compiled (which of course may be different). I have changed it back to:
2354
2355 #include "pcre.h"
2356
2357 I have a vague recollection that the change was concerned with compiling in
2358 different directories, but in the new build system, that is taken care of
2359 by the VPATH setting the Makefile.
2360
2361 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2362 when the subject happened to end in the byte 0x85 (e.g. if the last
2363 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2364 characters but of course it shouldn't be taken as a newline when it is part
2365 of another character. The bug was that, for an unlimited repeat of . in
2366 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2367 characters when looking for a newline.
2368
2369 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2370
2371 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2372 in debug output.
2373
2374 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2375 long printing in the pcrecpp unittest when running under MinGW.
2376
2377 6. ESC_K was left out of the EBCDIC table.
2378
2379 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2380 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2381 limit also applies to "virtual nesting" when a pattern is recursive, and in
2382 this case 1000 isn't so big. I have been able to remove this limit at the
2383 expense of backing off one optimization in certain circumstances. Normally,
2384 when pcre_exec() would call its internal match() function recursively and
2385 immediately return the result unconditionally, it uses a "tail recursion"
2386 feature to save stack. However, when a subpattern that can match an empty
2387 string has an unlimited repetition quantifier, it no longer makes this
2388 optimization. That gives it a stack frame in which to save the data for
2389 checking that an empty string has been matched. Previously this was taken
2390 from the 1000-entry workspace that had been reserved. So now there is no
2391 explicit limit, but more stack is used.
2392
2393 8. Applied Daniel's patches to solve problems with the import/export magic
2394 syntax that is required for Windows, and which was going wrong for the
2395 pcreposix and pcrecpp parts of the library. These were overlooked when this
2396 problem was solved for the main library.
2397
2398 9. There were some crude static tests to avoid integer overflow when computing
2399 the size of patterns that contain repeated groups with explicit upper
2400 limits. As the maximum quantifier is 65535, the maximum group length was
2401 set at 30,000 so that the product of these two numbers did not overflow a
2402 32-bit integer. However, it turns out that people want to use groups that
2403 are longer than 30,000 bytes (though not repeat them that many times).
2404 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2405 made it possible to implement the integer overflow checks in a much more
2406 dynamic way, which I have now done. The artificial limitation on group
2407 length has been removed - we now have only the limit on the total length of
2408 the compiled pattern, which depends on the LINK_SIZE setting.
2409
2410 10. Fixed a bug in the documentation for get/copy named substring when
2411 duplicate names are permitted. If none of the named substrings are set, the
2412 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2413 empty string.
2414
2415 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2416 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2417 because the ] is interpreted as the first data character and the
2418 terminating ] is not found. PCRE has been made compatible with Perl in this
2419 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2420 cause memory overwriting.
2421
2422 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2423 string has been matched (to stop an infinite loop). It was not recognizing
2424 a conditional subpattern that could match an empty string if that
2425 subpattern was within another subpattern. For example, it looped when
2426 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2427 condition was not nested. This bug has been fixed.
2428
2429 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2430 past the start of the subject in the presence of bytes with the top bit
2431 set, for example "\x8aBCD".
2432
2433 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2434 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2435
2436 14. Optimized (?!) to (*FAIL).
2437
2438 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2439 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2440 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2441 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2442 does: it's just the validity check that is more restrictive.
2443
2444 16. Inserted checks for integer overflows during escape sequence (backslash)
2445 processing, and also fixed erroneous offset values for syntax errors during
2446 backslash processing.
2447
2448 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2449 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2450
2451 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2452 caused an overrun.
2453
2454 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2455 something other than just ASCII characters) inside a group that had an
2456 unlimited repeat caused a loop at compile time (while checking to see
2457 whether the group could match an empty string).
2458
2459 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2460 [\P{Any}] did so. (Error in the code for printing property names.)
2461
2462 21. An orphan \E inside a character class could cause a crash.
2463
2464 22. A repeated capturing bracket such as (A)? could cause a wild memory
2465 reference during compilation.
2466
2467 23. There are several functions in pcre_compile() that scan along a compiled
2468 expression for various reasons (e.g. to see if it's fixed length for look
2469 behind). There were bugs in these functions when a repeated \p or \P was
2470 present in the pattern. These operators have additional parameters compared
2471 with \d, etc, and these were not being taken into account when moving along
2472 the compiled data. Specifically:
2473
2474 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2475 length.
2476
2477 (b) An item such as \pL+ within a repeated group could cause crashes or
2478 loops.
2479
2480 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2481 "reference to non-existent subpattern" error.
2482
2483 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2484
2485 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2486 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2487
2488 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2489
2490 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2491 character were causing crashes (broken optimization).
2492
2493 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2494 \p or \P) caused a compile-time loop.
2495
2496 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2497 break. For example, the unstudied pattern [\r\n]A does not match the string
2498 "\r\nA" because change 7.0/46 below moves the current point on by two
2499 characters after failing to match at the start. However, the pattern \nA
2500 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2501 the same is true. There doesn't seem any very clean way out of this, but
2502 what I have chosen to do makes the common cases work: PCRE now takes note
2503 of whether there can be an explicit match for \r or \n anywhere in the
2504 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2505 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2506 pattern has explicit CR or LF references.
2507
2508 29. Added (*CR) etc for changing newline setting at start of pattern.
2509
2510
2511 Version 7.2 19-Jun-07
2512 ---------------------
2513
2514 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2515 which is apparently normally available under Windows.
2516
2517 2. Re-jig the pcregrep tests with different newline settings in an attempt
2518 to make them independent of the local environment's newline setting.
2519
2520 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2521
2522 4. Some of the "internals" tests were previously cut out when the link size
2523 was not 2, because the output contained actual offsets. The recent new
2524 "Z" feature of pcretest means that these can be cut out, making the tests
2525 usable with all link sizes.
2526
2527 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2528 stack recursion. This gives a massive performance boost under BSD, but just
2529 a small improvement under Linux. However, it saves one field in the frame
2530 in all cases.
2531
2532 6. Added more features from the forthcoming Perl 5.10:
2533
2534 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2535 recursion call. It refers to the nth most recently opened parentheses.
2536
2537 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2538 to be opened parentheses.
2539
2540 (c) Conditions that refer to capturing parentheses can be specified
2541 relatively, for example, (?(-2)... or (?(+3)...
2542
2543 (d) \K resets the start of the current match so that everything before
2544 is not part of it.
2545
2546 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2547
2548 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2549 reference syntax.
2550
2551 (g) (?| introduces a group in which the numbering of parentheses in each
2552 alternative starts with the same number.
2553
2554 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2555
2556 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2557 PCRE_INFO_JCHANGED.
2558
2559 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2560 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2561 for detecting groups that can match an empty string.
2562
2563 9. A pattern with a very large number of alternatives (more than several
2564 hundred) was running out of internal workspace during the pre-compile
2565 phase, where pcre_compile() figures out how much memory will be needed. A
2566 bit of new cunning has reduced the workspace needed for groups with
2567 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2568 workspace instead of running out of the 4096 that are available.
2569
2570 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2571
2572 11. Applied patch from Google to remove an optimization that didn't quite work.
2573 The report of the bug said:
2574
2575 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2576 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2577 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2578
2579 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2580 it matched the wrong number of bytes.
2581
2582
2583 Version 7.1 24-Apr-07
2584 ---------------------
2585
2586 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2587 that is more "standard", making use of automake and other Autotools. There
2588 is some re-arrangement of the files and adjustment of comments consequent
2589 on this.
2590
2591 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2592 for recursive directory scanning broke on some systems because the files
2593 are not scanned in any specific order and on different systems the order
2594 was different. A call to "sort" has been inserted into RunGrepTest for the
2595 approprate test as a short-term fix. In the longer term there may be an
2596 alternative.
2597
2598 3. I had an email from Eric Raymond about problems translating some of PCRE's
2599 man pages to HTML (despite the fact that I distribute HTML pages, some
2600 people do their own conversions for various reasons). The problems
2601 concerned the use of low-level troff macros .br and .in. I have therefore
2602 removed all such uses from the man pages (some were redundant, some could
2603 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2604 HTML has been updated to handle .nf/.fi and to complain if it encounters
2605 .br or .in.
2606
2607 4. Updated comments in configure.ac that get placed in config.h.in and also
2608 arranged for config.h to be included in the distribution, with the name
2609 config.h.generic, for the benefit of those who have to compile without
2610 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2611
2612 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2613 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2614 makevp.bat for the current PCRE, using the additional files
2615 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2616
2617 6. A Windows user reported a minor discrepancy with test 2, which turned out
2618 to be caused by a trailing space on an input line that had got lost in his
2619 copy. The trailing space was an accident, so I've just removed it.
2620
2621 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2622 that is needed.
2623
2624 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2625 as "const" (a) because they are and (b) because it helps the PHP
2626 maintainers who have recently made a script to detect big data structures
2627 in the php code that should be moved to the .rodata section. I remembered
2628 to update Builducptable as well, so it won't revert if ucptable.h is ever
2629 re-created.
2630
2631 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2632 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2633 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2634 support is not required. This saves 1.5-2K of code, which is important in
2635 some applications.
2636
2637 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2638 so as not to refer to the tables, even though these functions will never be
2639 called when UTF-8 support is disabled. Otherwise there are problems with a
2640 shared library.
2641
2642 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2643
2644 (a) It was defining its arguments as char * instead of void *.
2645
2646 (b) It was assuming that all moves were upwards in memory; this was true
2647 a long time ago when I wrote it, but is no longer the case.
2648
2649 The emulated memove() is provided for those environments that have neither
2650 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2651 is clearly not the case, as these two bugs were recently reported.
2652
2653 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2654 and Detrail to create the HTML documentation, the .txt form of the man
2655 pages, and it removes trailing spaces from listed files. It also creates
2656 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2657 case, it wraps all the #defines with #ifndefs. This script should be run
2658 before "make dist".
2659
2660 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2661 with Unicode property support.
2662
2663 (a) For a maximizing quantifier, if the two different cases of the
2664 character were of different lengths in their UTF-8 codings (there are
2665 some cases like this - I found 11), and the matching function had to
2666 back up over a mixture of the two cases, it incorrectly assumed they
2667 were both the same length.
2668
2669 (b) When PCRE was configured to use the heap rather than the stack for
2670 recursion during matching, it was not correctly preserving the data for
2671 the other case of a UTF-8 character when checking ahead for a match
2672 while processing a minimizing repeat. If the check also involved
2673 matching a wide character, but failed, corruption could cause an
2674 erroneous result when trying to check for a repeat of the original
2675 character.
2676
2677 13. Some tidying changes to the testing mechanism:
2678
2679 (a) The RunTest script now detects the internal link size and whether there
2680 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2681 values substituted by "configure". (The RunGrepTest script already did
2682 this for UTF-8.) The configure.ac script no longer substitutes the
2683 relevant variables.
2684
2685 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2686 with length and offset values. This means that the output is different
2687 for different internal link sizes. Test 2 is skipped for link sizes
2688 other than 2 because of this, bypassing the problem. Unfortunately,
2689 there was also a test in test 3 (the locale tests) that used /B and
2690 failed for link sizes other than 2. Rather than cut the whole test out,
2691 I have added a new /Z option to pcretest that replaces the length and
2692 offset values with spaces. This is now used to make test 3 independent
2693 of link size. (Test 2 will be tidied up later.)
2694
2695 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2696 segmentation fault instead of returning the appropriate error message.
2697
2698 15. In multiline mode when the newline sequence was set to "any", the pattern
2699 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2700 This doesn't seem right; it now treats the CRLF combination as the line
2701 ending, and so does not match in that case. It's only a pattern such as ^$
2702 that would hit this one: something like ^ABC$ would have failed after \r
2703 and then tried again after \r\n.
2704
2705 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2706 in an attempt to make files that differ only in their line terminators
2707 compare equal. This works on Linux.
2708
2709 17. Under certain error circumstances pcregrep might try to free random memory
2710 as it exited. This is now fixed, thanks to valgrind.
2711
2712 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2713 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2714 was because its rules for how to advance for /g after matching an empty
2715 string at the end of a line did not allow for this case. They now check for
2716 it specially.
2717
2718 20. pcretest is supposed to handle patterns and data of any length, by
2719 extending its buffers when necessary. It was getting this wrong when the
2720 buffer for a data line had to be extended.
2721
2722 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2723 CRLF as a newline sequence.
2724
2725 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2726 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2727 I have nevertheless tidied it up.
2728
2729 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2730
2731 24. Added a man page for pcre-config.
2732
2733
2734 Version 7.0 19-Dec-06
2735 ---------------------
2736
2737 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2738 moving to gcc 4.1.1.
2739
2740 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2741 sys/time.h, which is documented as needed for this function. It doesn't
2742 seem to matter on Linux, but it showed up on some releases of OS X.
2743
2744 3. It seems that there are systems where bytes whose values are greater than
2745 127 match isprint() in the "C" locale. The "C" locale should be the
2746 default when a C program starts up. In most systems, only ASCII printing
2747 characters match isprint(). This difference caused the output from pcretest
2748 to vary, making some of the tests fail. I have changed pcretest so that:
2749
2750 (a) When it is outputting text in the compiled version of a pattern, bytes
2751 other than 32-126 are always shown as hex escapes.
2752
2753 (b) When it is outputting text that is a matched part of a subject string,
2754 it does the same, unless a different locale has been set for the match
2755 (using the /L modifier). In this case, it uses isprint() to decide.
2756
2757 4. Fixed a major bug that caused incorrect computation of the amount of memory
2758 required for a compiled pattern when options that changed within the
2759 pattern affected the logic of the preliminary scan that determines the
2760 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2761 that the computed length was too small. The symptoms of this bug were
2762 either the PCRE error "internal error: code overflow" from pcre_compile(),
2763 or a glibc crash with a message such as "pcretest: free(): invalid next
2764 size (fast)". Examples of patterns that provoked this bug (shown in
2765 pcretest format) are:
2766
2767 /(?-x: )/x
2768 /(?x)(?-x: \s*#\s*)/
2769 /((?i)[\x{c0}])/8
2770 /(?i:[\x{c0}])/8
2771
2772 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2773 is now done differently.
2774
2775 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2776 wrapper classes; (b) implement a new function in the C++ scanner that is
2777 more efficient than the old way of doing things because it avoids levels of
2778 recursion in the regex matching; (c) add a paragraph to the documentation
2779 for the FullMatch() function.
2780
2781 6. The escape sequence \n was being treated as whatever was defined as
2782 "newline". Not only was this contrary to the documentation, which states
2783 that \n is character 10 (hex 0A), but it also went horribly wrong when
2784 "newline" was defined as CRLF. This has been fixed.
2785
2786 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2787 was being set to -1 for the "end of line" case (supposedly a value that no
2788 character can have). Though this value is never used (the check for end of
2789 line is "zero bytes in current character"), it caused compiler complaints.
2790 I've changed it to 0xffffffff.
2791
2792 8. In pcre_version.c, the version string was being built by a sequence of
2793 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2794 string (as it is for production releases) called a macro with an empty
2795 argument. The C standard says the result of this is undefined. The gcc
2796 compiler treats it as an empty string (which was what was wanted) but it is
2797 reported that Visual C gives an error. The source has been hacked around to
2798 avoid this problem.
2799
2800 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2801 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2802 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2803 of them did).
2804
2805 10. Originally, pcretest opened its input and output without "b"; then I was
2806 told that "b" was needed in some environments, so it was added for release
2807 5.0 to both the input and output. (It makes no difference on Unix-like
2808 systems.) Later I was told that it is wrong for the input on Windows. I've
2809 now abstracted the modes into two macros, to make it easier to fiddle with
2810 them, and removed "b" from the input mode under Windows.
2811
2812 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2813
2814 12. Added -help and --help to pcretest as an official way of being reminded
2815 of the options.
2816
2817 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2818 and pcrecpp.cc because they annoy compilers at high warning levels.
2819
2820 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2821
2822 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2823 scripts are not C programs :-) and which was not noticed because it works
2824 on Linux.
2825
2826 16. pcretest is supposed to handle any length of pattern and data line (as one
2827 line or as a continued sequence of lines) by extending its input buffer if
2828 necessary. This feature was broken for very long pattern lines, leading to
2829 a string of junk being passed to pcre_compile() if the pattern was longer
2830 than about 50K.
2831
2832 17. I have done a major re-factoring of the way pcre_compile() computes the
2833 amount of memory needed for a compiled pattern. Previously, there was code
2834 that made a preliminary scan of the pattern in order to do this. That was
2835 OK when PCRE was new, but as the facilities have expanded, it has become
2836 harder and harder to keep it in step with the real compile phase, and there
2837 have been a number of bugs (see for example, 4 above). I have now found a
2838 cunning way of running the real compile function in a "fake" mode that
2839 enables it to compute how much memory it would need, while actually only
2840 ever using a few hundred bytes of working memory and without too many
2841 tests of the mode. This should make future maintenance and development
2842 easier. A side effect of this work is that the limit of 200 on the nesting
2843 depth of parentheses has been removed (though this was never a serious
2844 limitation, I suspect). However, there is a downside: pcre_compile() now
2845 runs more slowly than before (30% or more, depending on the pattern). I
2846 hope this isn't a big issue. There is no effect on runtime performance.
2847
2848 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2849 newline (only possible for the last line of a file) and it was a
2850 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2851
2852 19. Added additional timing features to pcretest. (1) The -tm option now times
2853 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2854 separate command line item, by a number that specifies the number of
2855 repeats to use when timing. The default is 50000; this gives better
2856 precision, but takes uncomfortably long for very large patterns.
2857
2858 20. Extended pcre_study() to be more clever in cases where a branch of a
2859 subpattern has no definite first character. For example, (a*|b*)[cd] would
2860 previously give no result from pcre_study(). Now it recognizes that the
2861 first character must be a, b, c, or d.
2862
2863 21. There was an incorrect error "recursive call could loop indefinitely" if
2864 a subpattern (or the entire pattern) that was being tested for matching an
2865 empty string contained only one non-empty item after a nested subpattern.
2866 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2867 incorrectly, because the \d was being skipped in the check.
2868
2869 22. The pcretest program now has a new pattern option /B and a command line
2870 option -b, which is equivalent to adding /B to every pattern. This causes
2871 it to show the compiled bytecode, without the additional information that
2872 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2873 is the same as /B/I).
2874
2875 23. A new optimization is now able automatically to treat some sequences such
2876 as a*b as a*+b. More specifically, if something simple (such as a character
2877 or a simple class like \d) has an unlimited quantifier, and is followed by
2878 something that cannot possibly match the quantified thing, the quantifier
2879 is automatically "possessified".
2880
2881 24. A recursive reference to a subpattern whose number was greater than 39
2882 went wrong under certain circumstances in UTF-8 mode. This bug could also
2883 have affected the operation of pcre_study().
2884
2885 25. Realized that a little bit of performance could be had by replacing
2886 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2887
2888 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2889
2890 27. Possessive quantifiers such as a++ were previously implemented by turning
2891 them into atomic groups such as ($>a+). Now they have their own opcodes,
2892 which improves performance. This includes the automatically created ones
2893 from 23 above.
2894
2895 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2896 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2897 the first matched character to be a colon. This applied both to named and
2898 numbered groups.
2899
2900 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2901
2902 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2903 building PCRE on Windows easier, so I have included it in the distribution.
2904
2905 31. There is now a check in pcretest against a ridiculously large number being
2906 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2907 loop, the loop is abandoned.
2908
2909 32. Forward references to subpatterns in conditions such as (?(2)...) where
2910 subpattern 2 is defined later cause pcre_compile() to search forwards in
2911 the pattern for the relevant set of parentheses. This search went wrong
2912 when there were unescaped parentheses in a character class, parentheses
2913 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2914
2915 33. "Subroutine" calls and backreferences were previously restricted to
2916 referencing subpatterns earlier in the regex. This restriction has now
2917 been removed.
2918
2919 34. Added a number of extra features that are going to be in Perl 5.10. On the
2920 whole, these are just syntactic alternatives for features that PCRE had
2921 previously implemented using the Python syntax or my own invention. The
2922 other formats are all retained for compatibility.
2923
2924 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2925 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2926 also .NET compatible.
2927
2928 (b) A recursion or subroutine call to a named group can now be defined as
2929 (?&name) as well as (?P>name).
2930
2931 (c) A backreference to a named group can now be defined as \k<name> or
2932 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2933 5.10, are also .NET compatible.
2934
2935 (d) A conditional reference to a named group can now use the syntax
2936 (?(<name>) or (?('name') as well as (?(name).
2937
2938 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2939 groups (named and numbered) that are never evaluated inline, but can be
2940 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2941 is always false. There may be only one alternative in such a group.
2942
2943 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2944 as the simple (?(R). The condition is true only if the most recent
2945 recursion is that of the given number or name. It does not search out
2946 through the entire recursion stack.
2947
2948 (g) The escape \gN or \g{N} has been added, where N is a positive or
2949 negative number, specifying an absolute or relative reference.
2950
2951 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2952 some "unreachable code" warnings.
2953
2954 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2955 things, this adds five new scripts.
2956
2957 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2958 There were also incompatibilities regarding the handling of \Q..\E inside
2959 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2960 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2961
2962 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2963 matches an empty string, and forcibly breaks the loop. There were bugs in
2964 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2965 against aaaa the result was just "a" rather than "aaaa", for example. Two
2966 separate and independent bugs (that affected different cases) have been
2967 fixed.
2968
2969 39. Refactored the code to abolish the use of different opcodes for small
2970 capturing bracket numbers. This is a tidy that I avoided doing when I
2971 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2972 The new approach is not only tidier, it makes it possible to reduce the
2973 memory needed to fix the previous bug (38).
2974
2975 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2976 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2977 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2978 mode.
2979
2980 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2981 report.
2982
2983 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2984 copy construction and assignment in the C++ wrapper.
2985
2986 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2987 couple of bugs that could have given wrong results in the "--newline=crlf"
2988 case.
2989
2990 44. Added a number of casts and did some reorganization of signed/unsigned int
2991 variables following suggestions from Dair Grant. Also renamed the variable
2992 "this" as "item" because it is a C++ keyword.
2993
2994 45. Arranged for dftables to add
2995
2996 #include "pcre_internal.h"
2997
2998 to pcre_chartables.c because without it, gcc 4.x may remove the array
2999 definition from the final binary if PCRE is built into a static library and
3000 dead code stripping is activated.
3001
3002 46. For an unanchored pattern, if a match attempt fails at the start of a
3003 newline sequence, and the newline setting is CRLF or ANY, and the next two
3004 characters are CRLF, advance by two characters instead of one.
3005
3006
3007 Version 6.7 04-Jul-06
3008 ---------------------
3009
3010 1. In order to handle tests when input lines are enormously long, pcretest has
3011 been re-factored so that it automatically extends its buffers when
3012 necessary. The code is crude, but this _is_ just a test program. The
3013 default size has been increased from 32K to 50K.
3014
3015 2. The code in pcre_study() was using the value of the re argument before
3016 testing it for NULL. (Of course, in any sensible call of the function, it
3017 won't be NULL.)
3018
3019 3. The memmove() emulation function in pcre_internal.h, which is used on
3020 systems that lack both memmove() and bcopy() - that is, hardly ever -
3021 was missing a "static" storage class specifier.
3022
3023 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3024 containing an extended class (one that cannot be represented by a bitmap
3025 because it contains high-valued characters or Unicode property items, e.g.
3026 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3027 pattern, but PCRE should not loop if you do not (it no longer does).
3028 [Detail: two cases were found: (a) a repeated subpattern containing an
3029 extended class; (b) a recursive reference to a subpattern that followed a
3030 previous extended class. It wasn't skipping over the extended class
3031 correctly when UTF-8 mode was not set.]
3032
3033 5. A negated single-character class was not being recognized as fixed-length
3034 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3035 compile error "lookbehind assertion is not fixed length".
3036
3037 6. The RunPerlTest auxiliary script was showing an unexpected difference
3038 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3039 write a Perl script that can interpret lines of an input file either as
3040 byte characters or as UTF-8, which is what "perltest" was being required to
3041 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3042 can't do is switch easily at run time between having the "use utf8;" pragma
3043 or not. In the end, I fudged it by using the RunPerlTest script to insert
3044 "use utf8;" explicitly for the UTF-8 tests.
3045
3046 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3047 the end of the subject string, contrary to the documentation and to what
3048 Perl does. This was true of both matching functions. Now it matches only at
3049 the start of the subject and immediately after *internal* newlines.
3050
3051 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3052 a pointer to an int instead of a pointer to an unsigned long int. This
3053 caused problems on 64-bit systems.
3054
3055 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3056 instance of the 'standard' template library not being so standard".
3057
3058 10. There was no check on the number of named subpatterns nor the maximum
3059 length of a subpattern name. The product of these values is used to compute
3060 the size of the memory block for a compiled pattern. By supplying a very
3061 long subpattern name and a large number of named subpatterns, the size
3062 computation could be caused to overflow. This is now prevented by limiting
3063 the length of names to 32 characters, and the number of named subpatterns
3064 to 10,000.
3065
3066 11. Subpatterns that are repeated with specific counts have to be replicated in
3067 the compiled pattern. The size of memory for this was computed from the
3068 length of the subpattern and the repeat count. The latter is limited to
3069 65535, but there was no limit on the former, meaning that integer overflow
3070 could in principle occur. The compiled length of a repeated subpattern is
3071 now limited to 30,000 bytes in order to prevent this.
3072
3073 12. Added the optional facility to have named substrings with the same name.
3074
3075 13. Added the ability to use a named substring as a condition, using the
3076 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3077 are numbers (not recommended). Forward references are permitted.
3078
3079 14. Added forward references in named backreferences (if you see what I mean).
3080
3081 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3082 pattern could run off the end of the subject. For example, the pattern
3083 "(?s)(.{1,5})"8 did this with the subject "ab".
3084
3085 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3086 PCRE_CASELESS was set when matching characters that were quantified with ?
3087 or *.
3088
3089 17. A character class other than a single negated character that had a minimum
3090 but no maximum quantifier - for example [ab]{6,} - was not handled
3091 correctly by pce_dfa_exec(). It would match only one character.
3092
3093 18. A valid (though odd) pattern that looked like a POSIX character
3094 class but used an invalid character after [ (for example [[,abc,]]) caused
3095 pcre_compile() to give the error "Failed: internal error: code overflow" or
3096 in some cases to crash with a glibc free() error. This could even happen if
3097 the pattern terminated after [[ but there just happened to be a sequence of
3098 letters, a binary zero, and a closing ] in the memory that followed.
3099
3100 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3101 over the years. Originally (before any Unicode support), just the bottom 8
3102 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3103 output from "man perlunicode" includes this:
3104
3105 The regular expression compiler produces polymorphic opcodes. That
3106 is, the pattern adapts to the data and automatically switches to
3107 the Unicode character scheme when presented with Unicode data--or
3108 instead uses a traditional byte scheme when presented with byte
3109 data.
3110
3111 Sadly, a wide octal escape does not cause a switch, and in a string with
3112 no other multibyte characters, these octal escapes are treated as before.
3113 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3114 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3115 Unicode string.
3116
3117 I have not perpetrated such confusion in PCRE. Up till now, it took just
3118 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3119 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3120 translate to the appropriate multibyte character.
3121
3122 29. Applied some refactoring to reduce the number of warnings from Microsoft
3123 and Borland compilers. This has included removing the fudge introduced
3124 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3125 a warning about an unused variable.
3126
3127 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3128 characters since release 4.0, because Perl (from release 5.004) does not.
3129 [Or at least, is documented not to: some releases seem to be in conflict
3130 with the documentation.] However, when a pattern was studied with
3131 pcre_study() and all its branches started with \s, PCRE still included VT
3132 as a possible starting character. Of course, this did no harm; it just
3133 caused an unnecessary match attempt.
3134
3135 22. Removed a now-redundant internal flag bit that recorded the fact that case
3136 dependency changed within the pattern. This was once needed for "required
3137 byte" processing, but is no longer used. This recovers a now-scarce options
3138 bit. Also moved the least significant internal flag bit to the most-
3139 significant bit of the word, which was not previously used (hangover from
3140 the days when it was an int rather than a uint) to free up another bit for
3141 the future.
3142
3143 23. Added support for CRLF line endings as well as CR and LF. As well as the
3144 default being selectable at build time, it can now be changed at runtime
3145 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3146 specify that it is scanning data with non-default line endings.
3147
3148 24. Changed the definition of CXXLINK to make it agree with the definition of
3149 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3150
3151 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3152 recursions. This makes a big different to stack usage for some patterns.
3153
3154 26. If a subpattern containing a named recursion or subroutine reference such
3155 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3156 the space required for the compiled pattern went wrong and gave too small a
3157 value. Depending on the environment, this could lead to "Failed: internal
3158 error: code overflow at offset 49" or "glibc detected double free or
3159 corruption" errors.
3160
3161 27. Applied patches from Google (a) to support the new newline modes and (b) to
3162 advance over multibyte UTF-8 characters in GlobalReplace.
3163
3164 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3165 difference for some implementation of PCRE in some Windows version.
3166
3167 29. Added some extra testing facilities to pcretest:
3168
3169 \q<number> in a data line sets the "match limit" value
3170 \Q<number> in a data line sets the "match recursion limt" value
3171 -S <number> sets the stack size, where <number> is in megabytes
3172
3173 The -S option isn't available for Windows.
3174
3175
3176 Version 6.6 06-Feb-06
3177 ---------------------
3178
3179 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3180 in pcreposix.h. I have copied the definition from pcre.h.
3181
3182 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3183 because pcre.h is no longer a built file.
3184
3185 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3186 not normally included in the compiled code.
3187
3188
3189 Version 6.5 01-Feb-06
3190 ---------------------
3191
3192 1. When using the partial match feature with pcre_dfa_exec(), it was not
3193 anchoring the second and subsequent partial matches at the new starting
3194 point. This could lead to incorrect results. For example, with the pattern
3195 /1234/, partially matching against "123" and then "a4" gave a match.
3196
3197 2. Changes to pcregrep:
3198
3199 (a) All non-match returns from pcre_exec() were being treated as failures
3200 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3201 error message is output. Some extra information is given for the
3202 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3203 probably the only errors that are likely to be caused by users (by
3204 specifying a regex that has nested indefinite repeats, for instance).
3205 If there are more than 20 of these errors, pcregrep is abandoned.
3206
3207 (b) A binary zero was treated as data while matching, but terminated the
3208 output line if it was written out. This has been fixed: binary zeroes
3209 are now no different to any other data bytes.
3210
3211 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3212 used to set a locale for matching. The --locale=xxxx long option has
3213 been added (no short equivalent) to specify a locale explicitly on the
3214 pcregrep command, overriding the environment variables.
3215
3216 (d) When -B was used with -n, some line numbers in the output were one less
3217 than they should have been.
3218
3219 (e) Added the -o (--only-matching) option.
3220
3221 (f) If -A or -C was used with -c (count only), some lines of context were
3222 accidentally printed for the final match.
3223
3224 (g) Added the -H (--with-filename) option.
3225
3226 (h) The combination of options -rh failed to suppress file names for files
3227 that were found from directory arguments.
3228
3229 (i) Added the -D (--devices) and -d (--directories) options.
3230
3231 (j) Added the -F (--fixed-strings) option.
3232
3233 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3234
3235 (l) Added the --colo(u)r option.
3236
3237 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3238 is not present by default.
3239
3240 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3241 items such as (?R) or (?1), when the recursion could match a number of
3242 alternatives. If it matched one of the alternatives, but subsequently,
3243 outside the recursion, there was a failure, the code tried to back up into
3244 the recursion. However, because of the way PCRE is implemented, this is not
3245 possible, and the result was an incorrect result from the match.
3246
3247 In order to prevent this happening, the specification of recursion has
3248 been changed so that all such subpatterns are automatically treated as
3249 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3250
3251 4. I had overlooked the fact that, in some locales, there are characters for
3252 which isalpha() is true but neither isupper() nor islower() are true. In
3253 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3254 and ordfeminine) are like this. This affected the treatment of \w and \W
3255 when they appeared in character classes, but not when they appeared outside
3256 a character class. The bit map for "word" characters is now created
3257 separately from the results of isalnum() instead of just taking it from the
3258 upper, lower, and digit maps. (Plus the underscore character, of course.)
3259
3260 5. The above bug also affected the handling of POSIX character classes such as
3261 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3262 permanent tables. Instead, the bit maps for such a class were previously
3263 created as the appropriate unions of the upper, lower, and digit bitmaps.
3264 Now they are created by subtraction from the [[:word:]] class, which has
3265 its own bitmap.
3266
3267 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3268 It is created by subtracting the vertical space characters (\x09, \x0a,
3269 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3270 subtraction was done in the overall bitmap for a character class, meaning
3271 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3272 be recognized. This bug has been fixed.
3273
3274 7. Patches from the folks at Google:
3275
3276 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3277 real life, but is still worth protecting against".
3278
3279 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3280 regular expressions".
3281
3282 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3283 have it.
3284
3285 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3286 "configure" and the latter not, in order to fix a problem somebody had
3287 with compiling the Arg class on HP-UX.
3288
3289 (e) Improve the error-handling of the C++ wrapper a little bit.
3290
3291 (f) New tests for checking recursion limiting.
3292
3293 8. The pcre_memmove() function, which is used only if the environment does not
3294 have a standard memmove() function (and is therefore rarely compiled),
3295 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3296 returning a result (though PCRE never actually uses the result).
3297
3298 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3299 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3300 returned instead of calling malloc() with an overflowing number that would
3301 most likely cause subsequent chaos.
3302
3303 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3304
3305 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3306 with this option is matched, the nmatch and pmatch options of regexec() are
3307 ignored.
3308
3309 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3310 provided in case anyone wants to the the POSIX interface with UTF-8
3311 strings.
3312
3313 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3314 C++ linking (needed for some HP-UX environments).
3315
3316 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3317 (unused parameter) and in the pcre_printint() function (omitted "default"
3318 switch label when the default is to do nothing).
3319
3320 15. Added some code to make it possible, when PCRE is compiled as a C++
3321 library, to replace subject pointers for pcre_exec() with a smart pointer
3322 class, thus making it possible to process discontinuous strings.
3323
3324 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3325 much the same function. They were added by different people who were trying
3326 to make PCRE easy to compile on non-Unix systems. It has been suggested
3327 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3328 for compiling on Windows systems. I have therefore replaced it with
3329 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3330 defaults to "extern" for C or "extern C" for C++, which works fine on
3331 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3332 SCOPE with something explicit in config.h. In addition:
3333
3334 (a) pcreposix.h still had just "extern" instead of either of these macros;
3335 I have replaced it with PCRE_DATA_SCOPE.
3336
3337 (b) Functions such as _pcre_xclass(), which are internal to the library,
3338 but external in the C sense, all had PCRE_EXPORT in their definitions.
3339 This is apparently wrong for the Windows case, so I have removed it.
3340 (It makes no difference on Unix-like systems.)
3341
3342 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3343 of recursive calls to match(). This is different to MATCH_LIMIT because
3344 that limits the total number of calls to match(), not all of which increase
3345 the depth of recursion. Limiting the recursion depth limits the amount of
3346 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3347 when PCRE is compiled, and changed at run time. A patch from Google adds
3348 this functionality to the C++ interface.
3349
3350 18. Changes to the handling of Unicode character properties:
3351
3352 (a) Updated the table to Unicode 4.1.0.
3353
3354 (b) Recognize characters that are not in the table as "Cn" (undefined).
3355
3356 (c) I revised the way the table is implemented to a much improved format
3357 which includes recognition of ranges. It now supports the ranges that
3358 are defined in UnicodeData.txt, and it also amalgamates other
3359 characters into ranges. This has reduced the number of entries in the
3360 table from around 16,000 to around 3,000, thus reducing its size
3361 considerably. I realized I did not need to use a tree structure after
3362 all - a binary chop search is just as efficient. Having reduced the
3363 number of entries, I extended their size from 6 bytes to 8 bytes to
3364 allow for more data.
3365
3366 (d) Added support for Unicode script names via properties such as \p{Han}.
3367
3368 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3369 matching that character.
3370
3371 20. When matching a repeated Unicode property with a minimum greater than zero,
3372 (for example \pL{2,}), PCRE could look past the end of the subject if it
3373 reached it while seeking the minimum number of characters. This could
3374 happen only if some of the characters were more than one byte long, because
3375 there is a check for at least the minimum number of bytes.
3376
3377 21. Refactored the implementation of \p and \P so as to be more general, to
3378 allow for more different types of property in future. This has changed the
3379 compiled form incompatibly. Anybody with saved compiled patterns that use
3380 \p or \P will have to recompile them.
3381
3382 22. Added "Any" and "L&" to the supported property types.
3383
3384 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3385 but give a compile time error if the value is greater than 0xff.
3386
3387 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3388 accidentally not being installed or uninstalled.
3389
3390 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3391 made were to insert the current release number. This seemed silly, because
3392 it made things harder for people building PCRE on systems that don't run
3393 "configure". I have turned pcre.h into a distributed file, no longer built
3394 by "configure", with the version identification directly included. There is
3395 no longer a pcre.h.in file.
3396
3397 However, this change necessitated a change to the pcre-config script as
3398 well. It is built from pcre-config.in, and one of the substitutions was the
3399 release number. I have updated configure.ac so that ./configure now finds
3400 the release number by grepping pcre.h.
3401
3402 26. Added the ability to run the tests under valgrind.
3403
3404
3405 Version 6.4 05-Sep-05
3406 ---------------------
3407
3408 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3409 "--" to be printed when multiple files were scanned, even when none of the
3410 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3411 consider it to be a bug, and have restored the previous behaviour.
3412
3413 2. A couple of code tidies to get rid of compiler warnings.
3414
3415 3. The pcretest program used to cheat by referring to symbols in the library
3416 whose names begin with _pcre_. These are internal symbols that are not
3417 really supposed to be visible externally, and in some environments it is
3418 possible to suppress them. The cheating is now confined to including
3419 certain files from the library's source, which is a bit cleaner.
3420
3421 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3422 file's purpose clearer.
3423
3424 5. Reorganized pcre_ucp_findchar().
3425
3426
3427 Version 6.3 15-Aug-05
3428 ---------------------
3429
3430 1. The file libpcre.pc.in did not have general read permission in the tarball.
3431
3432 2. There were some problems when building without C++ support:
3433
3434 (a) If C++ support was not built, "make install" and "make test" still
3435 tried to test it.
3436
3437 (b) There were problems when the value of CXX was explicitly set. Some
3438 changes have been made to try to fix these, and ...
3439
3440 (c) --disable-cpp can now be used to explicitly disable C++ support.
3441
3442 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3443 backslash in a target when C++ was disabled. This confuses some
3444 versions of "make", apparently. Using an intermediate variable solves
3445 this. (Same for CPP_LOBJ.)
3446
3447 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3448 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3449 necessary on certain architectures.
3450
3451 4. Added a setting of -export-symbols-regex to the link command to remove
3452 those symbols that are exported in the C sense, but actually are local
3453 within the library, and not documented. Their names all begin with
3454 "_pcre_". This is not a perfect job, because (a) we have to except some
3455 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3456 available (and never for static libraries). I have made a note to try to
3457 find a way round (a) in the future.
3458
3459
3460 Version 6.2 01-Aug-05
3461 ---------------------
3462
3463 1. There was no test for integer overflow of quantifier values. A construction
3464 such as {1111111111111111} would give undefined results. What is worse, if
3465 a minimum quantifier for a parenthesized subpattern overflowed and became
3466 negative, the calculation of the memory size went wrong. This could have
3467 led to memory overwriting.
3468
3469 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3470
3471 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3472 operating environments where this matters.
3473
3474 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3475 PCRE options from within the C++ wrapper.
3476
3477 5. Named capturing subpatterns were not being correctly counted when a pattern
3478 was compiled. This caused two problems: (a) If there were more than 100
3479 such subpatterns, the calculation of the memory needed for the whole
3480 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3481 back references of the form \12, where the number was greater than 9, were
3482 not recognized as back references, even though there were sufficient
3483 previous subpatterns.
3484
3485 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3486 versions of gcc, e.g. 2.95.4.
3487
3488
3489 Version 6.1 21-Jun-05
3490 ---------------------
3491
3492 1. There was one reference to the variable "posix" in pcretest.c that was not
3493 surrounded by "#if !defined NOPOSIX".
3494
3495 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3496 the cross-check on the old pcre_info() function, for the benefit of the
3497 cut-down version of PCRE that is currently imported into Exim.
3498
3499 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3500 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3501 patterns that start (?i) but I don't think that matters. The use of (?i) is
3502 just an example; this all applies to the other options as well.
3503
3504 4. Since libtool seems to echo the compile commands it is issuing, the output
3505 from "make" can be reduced a bit by putting "@" in front of each libtool
3506 compile command.
3507
3508 5. Patch from the folks at Google for configure.in to be a bit more thorough
3509 in checking for a suitable C++ installation before trying to compile the
3510 C++ stuff. This should fix a reported problem when a compiler was present,
3511 but no suitable headers.
3512
3513 6. The man pages all had just "PCRE" as their title. I have changed them to
3514 be the relevant file name. I have also arranged that these names are
3515 retained in the file doc/pcre.txt, which is a concatenation in text format
3516 of all the man pages except the little individual ones for each function.
3517
3518 7. The NON-UNIX-USE file had not been updated for the different set of source
3519 files that come with release 6. I also added a few comments about the C++
3520 wrapper.
3521
3522
3523 Version 6.0 07-Jun-05
3524 ---------------------
3525
3526 1. Some minor internal re-organization to help with my DFA experiments.
3527
3528 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3529 didn't matter for the library itself when fully configured, but did matter
3530 when compiling without UCP support, or within Exim, where the ucp files are
3531 not imported.
3532
3533 3. Refactoring of the library code to split up the various functions into
3534 different source modules. The addition of the new DFA matching code (see
3535 below) to a single monolithic source would have made it really too
3536 unwieldy, quite apart from causing all the code to be include in a
3537 statically linked application, when only some functions are used. This is
3538 relevant even without the DFA addition now that patterns can be compiled in
3539 one application and matched in another.
3540
3541 The downside of splitting up is that there have to be some external
3542 functions and data tables that are used internally in different modules of
3543 the library but which are not part of the API. These have all had their
3544 names changed to start with "_pcre_" so that they are unlikely to clash
3545 with other external names.
3546
3547 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3548 a different (DFA) algorithm. Although it is slower than the original
3549 function, it does have some advantages for certain types of matching
3550 problem.
3551
3552 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3553 including restarting after a partial match.
3554
3555 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3556 defined when compiling for Windows was sent to me. I have put it into the
3557 code, though I have no means of testing or verifying it.
3558
3559 7. Added the pcre_refcount() auxiliary function.
3560
3561 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3562 match before or at the first newline in the subject string. In pcretest,
3563 the /f option on a pattern can be used to set this.
3564
3565 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3566 would behave wrongly. This has been present in PCRE since release 4.0.
3567
3568 10. A number of changes to the pcregrep command:
3569
3570 (a) Refactored how -x works; insert ^(...)$ instead of setting
3571 PCRE_ANCHORED and checking the length, in preparation for adding
3572 something similar for -w.
3573
3574 (b) Added the -w (match as a word) option.
3575
3576 (c) Refactored the way lines are read and buffered so as to have more
3577 than one at a time available.
3578
3579 (d) Implemented a pcregrep test script.
3580
3581 (e) Added the -M (multiline match) option. This allows patterns to match
3582 over several lines of the subject. The buffering ensures that at least
3583 8K, or the rest of the document (whichever is the shorter) is available
3584 for matching (and similarly the previous 8K for lookbehind assertions).
3585
3586 (f) Changed the --help output so that it now says
3587
3588 -w, --word-regex(p)
3589
3590 instead of two lines, one with "regex" and the other with "regexp"
3591 because that confused at least one person since the short forms are the
3592 same. (This required a bit of code, as the output is generated
3593 automatically from a table. It wasn't just a text change.)
3594
3595 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3596 option but starts with a hyphen. Could be a pattern or a path name
3597 starting with a hyphen, for instance.
3598
3599 (h) "-" can be given as a file name to represent stdin.
3600
3601 (i) When file names are being printed, "(standard input)" is used for
3602 the standard input, for compatibility with GNU grep. Previously
3603 "<stdin>" was used.
3604
3605 (j) The option --label=xxx can be used to supply a name to be used for
3606 stdin when file names are being printed. There is no short form.
3607
3608 (k) Re-factored the options decoding logic because we are going to add
3609 two more options that take data. Such options can now be given in four
3610 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3611
3612 (l) Added the -A, -B, and -C options for requesting that lines of context
3613 around matches be printed.
3614
3615 (m) Added the -L option to print the names of files that do not contain
3616 any matching lines, that is, the complement of -l.
3617
3618 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3619 continue to scan other files.
3620
3621 (o) The -s option was incorrectly implemented. For compatibility with other
3622 greps, it now suppresses the error message for a non-existent or non-
3623 accessible file (but not the return code). There is a new option called
3624 -q that suppresses the output of matching lines, which was what -s was
3625 previously doing.
3626
3627 (p) Added --include and --exclude options to specify files for inclusion
3628 and exclusion when recursing.
3629
3630 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3631 Hopefully, it now does.
3632
3633 12. Missing cast in pcre_study().
3634
3635 13. Added an "uninstall" target to the makefile.
3636
3637 14. Replaced "extern" in the function prototypes in Makefile.in with
3638 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3639 world, but is set differently for Windows.
3640
3641 15. Added a second compiling function called pcre_compile2(). The only
3642 difference is that it has an extra argument, which is a pointer to an
3643 integer error code. When there is a compile-time failure, this is set
3644 non-zero, in addition to the error test pointer being set to point to an
3645 error message. The new argument may be NULL if no error number is required
3646 (but then you may as well call pcre_compile(), which is now just a
3647 wrapper). This facility is provided because some applications need a
3648 numeric error indication, but it has also enabled me to tidy up the way
3649 compile-time errors are handled in the POSIX wrapper.
3650
3651 16. Added VPATH=.libs to the makefile; this should help when building with one
3652 prefix path and installing with another. (Or so I'm told by someone who
3653 knows more about this stuff than I do.)
3654
3655 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3656 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3657 match everything, including newlines. This is not POSIX-compatible, but
3658 somebody wanted the feature. From pcretest it can be activated by using
3659 both the P and the s flags.
3660
3661 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3662
3663 19. libpcre.pc was being incorrectly installed as executable.
3664
3665 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3666 it now also looks for '\r' so that it will work unmodified on Windows.
3667
3668 21. Added Google's contributed C++ wrapper to the distribution.
3669
3670 22. Added some untidy missing memory free() calls in pcretest, to keep
3671 Electric Fence happy when testing.
3672
3673
3674
3675 Version 5.0 13-Sep-04
3676 ---------------------
3677
3678 1. Internal change: literal characters are no longer packed up into items
3679 containing multiple characters in a single byte-string. Each character
3680 is now matched using a separate opcode. However, there may be more than one
3681 byte in the character in UTF-8 mode.
3682
3683 2. The pcre_callout_block structure has two new fields: pattern_position and
3684 next_item_length. These contain the offset in the pattern to the next match
3685 item, and its length, respectively.
3686
3687 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3688 insertion of callouts before each pattern item. Added the /C option to
3689 pcretest to make use of this.
3690
3691 4. On the advice of a Windows user, the lines
3692
3693 #if defined(_WIN32) || defined(WIN32)
3694 _setmode( _fileno( stdout ), 0x8000 );
3695 #endif /* defined(_WIN32) || defined(WIN32) */
3696
3697 have been added to the source of pcretest. This apparently does useful
3698 magic in relation to line terminators.
3699
3700 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3701 for the benefit of those environments where the "b" makes a difference.
3702
3703 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3704 to know about it. I have put a hack into configure.in that adds in code
3705 to set GCC=yes if CC=icc. This seems to end up at a point in the
3706 generated configure script that is early enough to affect the setting of
3707 compiler options, which is what is needed, but I have no means of testing
3708 whether it really works. (The user who reported this had patched the
3709 generated configure script, which of course I cannot do.)
3710
3711 LATER: After change 22 below (new libtool files), the configure script
3712 seems to know about icc (and also ecc). Therefore, I have commented out
3713 this hack in configure.in.
3714
3715 7. Added support for pkg-config (2 patches were sent in).
3716
3717 8. Negated POSIX character classes that used a combination of internal tables
3718 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3719 [[:^ascii]]. Typically, they would match almost any characters. The other
3720 POSIX classes were not broken in this way.
3721
3722 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3723 to find the match, as PCRE was deluded into thinking that the match had to
3724 start at the start point or following a newline. The same bug applied to
3725 patterns with negative forward assertions or any backward assertions
3726 preceding ".*" at the start, unless the pattern required a fixed first
3727 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3728
3729 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3730 starting at the last subject character, bytes beyond the end of the subject
3731 string were read.
3732
3733 11. Renamed the variable "class" as "classbits" to make life easier for C++
3734 users. (Previously there was a macro definition, but it apparently wasn't
3735 enough.)
3736
3737 12. Added the new field "tables" to the extra data so that tables can be passed
3738 in at exec time, or the internal tables can be re-selected. This allows
3739 a compiled regex to be saved and re-used at a later time by a different
3740 program that might have everything at different addresses.
3741
3742 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3743 -R library as well as a -L library.
3744
3745 14. The debugging options of pcretest (-d on the command line or D on a
3746 pattern) showed incorrect output for anything following an extended class
3747 that contained multibyte characters and which was followed by a quantifier.
3748
3749 15. Added optional support for general category Unicode character properties
3750 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3751 support. It adds about 90K to the size of the library. The meanings of the
3752 inbuilt class escapes such as \d and \s have NOT been changed.
3753
3754 16. Updated pcredemo.c to include calls to free() to release the memory for the
3755 compiled pattern.
3756
3757 17. The generated file chartables.c was being created in the source directory
3758 instead of in the building directory. This caused the build to fail if the
3759 source directory was different from the building directory, and was
3760 read-only.
3761
3762 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3763 file. No doubt somebody will tell me if they don't make sense... Also added
3764 Dan Mooney's comments about building on OpenVMS.
3765
3766 19. Added support for partial matching via the PCRE_PARTIAL option for
3767 pcre_exec() and the \P data escape in pcretest.
3768
3769 20. Extended pcretest with 3 new pattern features:
3770
3771 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3772 write the compiled pattern to the file whose name is "rest-of-line".
3773 This is a straight binary dump of the data, with the saved pointer to
3774 the character tables forced to be NULL. The study data, if any, is
3775 written too. After writing, pcretest reads a new pattern.
3776
3777 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3778 compiled pattern from the given file. There must not be any
3779 occurrences of "<" in the file name (pretty unlikely); if there are,
3780 pcretest will instead treat the initial "<" as a pattern delimiter.
3781 After reading in the pattern, pcretest goes on to read data lines as
3782 usual.
3783
3784 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3785 and 16-bit fields in a compiled pattern, to simulate a pattern that
3786 was compiled on a host of opposite endianness.
3787
3788 21. The pcre-exec() function can now cope with patterns that were compiled on
3789 hosts of opposite endianness, with this restriction:
3790
3791 As for any compiled expression that is saved and used later, the tables
3792 pointer field cannot be preserved; the extra_data field in the arguments
3793 to pcre_exec() should be used to pass in a tables address if a value
3794 other than the default internal tables were used at compile time.
3795
3796 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3797 now diagnosed as an error. Previously, most of the time, a negative number
3798 would have been treated as zero, but if in addition "ovector" was passed as
3799 NULL, a crash could occur.
3800
3801 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3802 new versions from the libtool 1.5 distribution (the last one is a copy of
3803 a file called libtool.m4). This seems to have fixed the need to patch
3804 "configure" to support Darwin 1.3 (which I used to do). However, I still
3805 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3806 workstation).
3807
3808 24. Changed the PCRE licence to be the more standard "BSD" licence.
3809
3810
3811 Version 4.5 01-Dec-03
3812 ---------------------
3813
3814 1. There has been some re-arrangement of the code for the match() function so
3815 that it can be compiled in a version that does not call itself recursively.
3816 Instead, it keeps those local variables that need separate instances for
3817 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3818 needs to "recurse". Keeping track of where control must go is done by means
3819 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3820 hide most of the details from the main code, and operates only if
3821 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3822 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3823 operating.
3824
3825 To make it easier for callers to provide specially tailored get/free
3826 functions for this usage, two new functions, pcre_stack_malloc, and
3827 pcre_stack_free, are used. They are always called in strict stacking order,
3828 and the size of block requested is always the same.
3829
3830 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3831 PCRE has been compiled to use the stack or the heap for recursion. The
3832 -C option of pcretest uses this to show which version is compiled.
3833
3834 A new data escape \S, is added to pcretest; it causes the amounts of store
3835 obtained and freed by both kinds of malloc/free at match time to be added
3836 to the output.
3837
3838 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3839 what's available on my current Linux desktop machine.
3840
3841 3. When matching a UTF-8 string, the test for a valid string at the start has
3842 been extended. If start_offset is not zero, PCRE now checks that it points
3843 to a byte that is the start of a UTF-8 character. If not, it returns
3844 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3845 this is necessary because there may be backward assertions in the pattern.
3846 When matching the same subject several times, it may save resources to use
3847 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3848
3849 4. The code for checking the validity of UTF-8 strings has been tightened so
3850 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3851 containing "overlong sequences".
3852
3853 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3854 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3855 should have been "&", but it just so happened that all the cases this let
3856 through by mistake were picked up later in the function.
3857
3858 6. I had used a variable called "isblank" - this is a C99 function, causing
3859 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3860
3861 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3862 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3863 so the version is shown; (c) stop "make test" echoing "./RunTest".
3864
3865 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3866
3867 9. The prototype for memmove() for systems that don't have it was using
3868 size_t, but the inclusion of the header that defines size_t was later. I've
3869 moved the #includes for the C headers earlier to avoid this.
3870
3871 10. Added some adjustments to the code to make it easier to compiler on certain
3872 special systems:
3873
3874 (a) Some "const" qualifiers were missing.
3875 (b) Added the macro EXPORT before all exported functions; by default this
3876 is defined to be empty.
3877 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3878 that it reads its output file name as an argument instead of writing
3879 to the standard output and assuming this can be redirected.
3880
3881 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3882 class containing characters with values greater than 255, PCRE compilation
3883 went into a loop.
3884
3885 12. A recursive reference to a subpattern that was within another subpattern
3886 that had a minimum quantifier of zero caused PCRE to crash. For example,
3887 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3888 recursion. If the recursively-called subpattern itself had a zero repeat,
3889 that was OK.
3890
3891 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3892 buffer into which it was copied (for escape processing) was still set at
3893 1024, so long lines caused crashes.
3894
3895 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3896 "internal error: code overflow...". This applied to any character class
3897 that was followed by a possessive quantifier.
3898
3899 15. Modified the Makefile to add libpcre.la as a prerequisite for
3900 libpcreposix.la because I was told this is needed for a parallel build to
3901 work.
3902
3903 16. If a pattern that contained .* following optional items at the start was
3904 studied, the wrong optimizing data was generated, leading to matching
3905 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3906 matching string must start with a or b or c. The correct conclusion for
3907 this pattern is that a match can start with any character.
3908
3909
3910 Version 4.4 13-Aug-03
3911 ---------------------
3912
3913 1. In UTF-8 mode, a character class containing characters with values between
3914 127 and 255 was not handled correctly if the compiled pattern was studied.
3915 In fixing this, I have also improved the studying algorithm for such
3916 classes (slightly).
3917
3918 2. Three internal functions had redundant arguments passed to them. Removal
3919 might give a very teeny performance improvement.
3920
3921 3. Documentation bug: the value of the capture_top field in a callout is *one
3922 more than* the number of the hightest numbered captured substring.
3923
3924 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3925 in incorrectly linking with a previously installed version. They now link
3926 explicitly with libpcre.la.
3927
3928 5. configure.in no longer needs to recognize Cygwin specially.
3929
3930 6. A problem in pcre.in for Windows platforms is fixed.
3931
3932 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3933 pcretest, it used to include the size of the study block as part of its
3934 output. Unfortunately, the structure contains a field that has a different
3935 size on different hardware architectures. This meant that the tests that
3936 showed this size failed. As the block is currently always of a fixed size,
3937 this information isn't actually particularly useful in pcretest output, so
3938 I have just removed it.
3939
3940 8. Three pre-processor statements accidentally did not start in column 1.
3941 Sadly, there are *still* compilers around that complain, even though
3942 standard C has not required this for well over a decade. Sigh.
3943
3944 9. In pcretest, the code for checking callouts passed small integers in the
3945 callout_data field, which is a void * field. However, some picky compilers
3946 complained about the casts involved for this on 64-bit systems. Now
3947 pcretest passes the address of the small integer instead, which should get
3948 rid of the warnings.
3949
3950 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3951 both compile and run time, and gives an error if an invalid UTF-8 sequence
3952 is found. There is a option for disabling this check in cases where the
3953 string is known to be correct and/or the maximum performance is wanted.
3954
3955 11. In response to a bug report, I changed one line in Makefile.in from
3956
3957 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3958 to
3959 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3960
3961 to look similar to other lines, but I have no way of telling whether this
3962 is the right thing to do, as I do not use Windows. No doubt I'll get told
3963 if it's wrong...
3964
3965
3966 Version 4.3 21-May-03
3967 ---------------------
3968
3969 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3970 Makefile.
3971
3972 2. Some refactoring to improve the quality of the code:
3973
3974 (i) The utf8_table... variables are now declared "const".
3975
3976 (ii) The code for \cx, which used the "case flipping" table to upper case
3977 lower case letters, now just substracts 32. This is ASCII-specific,
3978 but the whole concept of \cx is ASCII-specific, so it seems
3979 reasonable.
3980
3981 (iii) PCRE was using its character types table to recognize decimal and
3982 hexadecimal digits in the pattern. This is silly, because it handles
3983 only 0-9, a-f, and A-F, but the character types table is locale-
3984 specific, which means strange things might happen. A private
3985 table is now used for this - though it costs 256 bytes, a table is
3986 much faster than multiple explicit tests. Of course, the standard
3987 character types table is still used for matching digits in subject
3988 strings against \d.
3989
3990 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3991 ending in _t are). So I've renamed it as ESC_tee.
3992
3993 3. The first argument for regexec() in the POSIX wrapper should have been
3994 defined as "const".
3995
3996 4. Changed pcretest to use malloc() for its buffers so that they can be
3997 Electric Fenced for debugging.
3998
3999 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4000 to read one or more bytes before the start of the subject string. Often this
4001 had no effect on PCRE's behaviour, but in some circumstances it could
4002 provoke a segmentation fault.
4003
4004 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4005 to try to read one or more bytes before the start of the subject string.
4006
4007 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4008 UTF-8 support could misbehave in various ways if the subject string
4009 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4010 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4011 back over UTF-8 characters.)
4012
4013
4014 Version 4.2 14-Apr-03
4015 ---------------------
4016
4017 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4018
4019 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4020 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4021 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4022 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4023 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4024 and BUILD_EXEEXT
4025 Note: automatic setting of the BUILD variables is not yet working
4026 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4027 compile-time but not at link-time
4028 [LINK]: use for linking executables only
4029 make different versions for Windows and non-Windows
4030 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4031 libraries
4032 [LINK_FOR_BUILD]: new variable
4033 [OBJEXT]: use throughout
4034 [EXEEXT]: use throughout
4035 <winshared>: new target
4036 <wininstall>: new target
4037 <dftables.o>: use native compiler
4038 <dftables>: use native linker
4039 <install>: handle Windows platform correctly
4040 <clean>: ditto
4041 <check>: ditto
4042 copy DLL to top builddir before testing
4043
4044 As part of these changes, -no-undefined was removed again. This was reported
4045 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4046 in any case.
4047
4048 3. Some tidies to get rid of compiler warnings:
4049
4050 . In the match_data structure, match_limit was an unsigned long int, whereas
4051 match_call_count was an int. I've made them both unsigned long ints.
4052
4053 . In pcretest the fact that a const uschar * doesn't automatically cast to
4054 a void * provoked a warning.
4055
4056 . Turning on some more compiler warnings threw up some "shadow" variables
4057 and a few more missing casts.
4058
4059 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4060 option, a class that contained a single character with a value between 128
4061 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4062
4063 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4064 option, a class that contained several characters, but with at least one
4065 whose value was between 128 and 255 caused PCRE to crash.
4066
4067
4068 Version 4.1 12-Mar-03
4069 ---------------------
4070
4071 1. Compiling with gcc -pedantic found a couple of places where casts were
4072 needed, and a string in dftables.c that was longer than standard compilers are
4073 required to support.
4074
4075 2. Compiling with Sun's compiler found a few more places where the code could
4076 be tidied up in order to avoid warnings.
4077
4078 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4079 first of these names is deprecated in the latest Autoconf in favour of the name
4080 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4081 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4082 analogy I have changed it to CFLAGS_FOR_BUILD.
4083
4084 4. Added -no-undefined to the linking command in the Makefile, because this is
4085 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4086 linking step for the pcreposix library.
4087
4088 5. PCRE was failing to diagnose the case of two named groups with the same
4089 name.
4090
4091 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4092 literal character that is needed in the subject for a match, and scans along to
4093 ensure that it is present before embarking on the full matching process. This
4094 saves time in cases of nested unlimited repeats that are never going to match.
4095 Problem: the scan can take a lot of time if the subject is very long (e.g.
4096 megabytes), thus penalizing straightforward matches. It is now done only if the
4097 amount of subject to be scanned is less than 1000 bytes.
4098
4099 7. A lesser problem with the same optimization is that it was recording the
4100 first character of an anchored pattern as "needed", thus provoking a search
4101 right along the subject, even when the first match of the pattern was going to
4102 fail. The "needed" character is now not set for anchored patterns, unless it
4103 follows something in the pattern that is of non-fixed length. Thus, it still
4104 fulfils its original purpose of finding quick non-matches in cases of nested
4105 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4106
4107
4108 Version 4.0 17-Feb-03
4109 ---------------------
4110
4111 1. If a comment in an extended regex that started immediately after a meta-item
4112 extended to the end of string, PCRE compiled incorrect data. This could lead to
4113 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4114
4115 2. Moved to autoconf 2.53 and libtool 1.4.2.
4116
4117 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4118 the special perltest8 script is no longer needed - all the tests can be run
4119 from a single perltest script.
4120
4121 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4122 by \s. It has now been removed in PCRE. This means it isn't recognized as
4123 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4124 class [:space:] *does* include VT, thereby creating a mess.
4125
4126 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4127 space and tab.
4128
4129 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4130 its new features into the main test script, reducing the number of scripts.
4131
4132 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4133 were backward compatible, and made the (?i) apply to the whole pattern, as if
4134 /i were given. Now it behaves more logically, and applies the option setting
4135 only to what follows. PCRE has been changed to follow suit. However, if it
4136 finds options settings right at the start of the pattern, it extracts them into
4137 the global options, as before. Thus, they show up in the info data.
4138
4139 8. Added support for the \Q...\E escape sequence. Characters in between are
4140 treated as literals. This is slightly different from Perl in that $ and @ are
4141 also handled as literals inside the quotes. In Perl, they will cause variable
4142 interpolation. Note the following examples:
4143
4144 Pattern PCRE matches Perl matches
4145
4146 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4147 \Qabc\$xyz\E abc\$xyz abc\$xyz
4148 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4149
4150 For compatibility with Perl, \Q...\E sequences are recognized inside character
4151 classes as well as outside them.
4152
4153 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4154 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4155 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4156 signed/unsigned warnings.
4157
4158 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4159 option for pcretest, so I've replaced it by a simple function that does just
4160 that job.
4161
4162 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4163 "pcregrep -".
4164
4165 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4166 Java package. This provides some syntactic sugar for simple cases of what my
4167 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4168 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4169 item, you can use this simplified notation. Note that only makes sense with
4170 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4171 greediness, whatever the setting of the PCRE_UNGREEDY option.
4172
4173 13. A change of greediness default within a pattern was not taking effect at
4174 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4175 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4176 was abstracted outside.
4177
4178 14. PCRE now supports the \G assertion. It is true when the current matching
4179 position is at the start point of the match. This differs from \A when the
4180 starting offset is non-zero. Used with the /g option of pcretest (or similar
4181 code), it works in the same way as it does for Perl's /g option. If all
4182 alternatives of a regex begin with \G, the expression is anchored to the start
4183 match position, and the "anchored" flag is set in the compiled expression.
4184
4185 15. Some bugs concerning the handling of certain option changes within patterns
4186 have been fixed. These applied to options other than (?ims). For example,
4187 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4188 been the other way round. Some of this was related to change 7 above.
4189
4190 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4191 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4192 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4193 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4194
4195 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4196 mode. Unlike ".", it always matches newline, whatever the setting of
4197 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4198 assertions. Perl allows it, but it doesn't (in general) work because it can't
4199 calculate the length of the lookbehind. At least, that's the case for Perl
4200 5.8.0 - I've been told they are going to document that it doesn't work in
4201 future.
4202
4203 18. Added an error diagnosis for escapes that PCRE does not support: these are
4204 \L, \l, \N, \P, \p, \U, \u, and \X.
4205
4206 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4207 reading past the end of the pattern in cases such as /[abcd/.
4208
4209 20. PCRE was getting more memory than necessary for patterns with classes that
4210 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4211
4212 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4213 compiling PCRE for use with Virtual Pascal.
4214
4215 22. Small fix to the Makefile to make it work properly if the build is done
4216 outside the source tree.
4217
4218 23. Added a new extension: a condition to go with recursion. If a conditional
4219 subpattern starts with (?(R) the "true" branch is used if recursion has
4220 happened, whereas the "false" branch is used only at the top level.
4221
4222 24. When there was a very long string of literal characters (over 255 bytes
4223 without UTF support, over 250 bytes with UTF support), the computation of how
4224 much memory was required could be incorrect, leading to segfaults or other
4225 strange effects.
4226
4227 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4228 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4229 there was a subsequent back reference to those brackets. This meant that, for
4230 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4231 possible to check for precisely this case. All we can do is abandon the
4232 optimization if .* occurs inside capturing brackets when there are any back
4233 references whatsoever. (See below for a better fix that came later.)
4234
4235 26. The handling of the optimization for finding the first character of a
4236 non-anchored pattern, and for finding a character that is required later in the
4237 match were failing in some cases. This didn't break the matching; it just
4238 failed to optimize when it could. The way this is done has been re-implemented.
4239
4240 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4241
4242 28. Added a new feature that provides some of the functionality that Perl
4243 provides with (?{...}). The facility is termed a "callout". The way it is done
4244 in PCRE is for the caller to provide an optional function, by setting
4245 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4246 global variable. By default it is unset, which disables all calling out. To get
4247 the function called, the regex must include (?C) at appropriate points. This
4248 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4249 This provides a means of identifying different callout points. When PCRE
4250 reaches such a point in the regex, if pcre_callout has been set, the external
4251 function is called. It is provided with data in a structure called
4252 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4253 matching continues; if it returns a non-zero value, the match at the current
4254 point fails. However, backtracking will occur if possible. [This was changed
4255 later and other features added - see item 49 below.]
4256
4257 29. pcretest is upgraded to test the callout functionality. It provides a
4258 callout function that displays information. By default, it shows the start of
4259 the match and the current position in the text. There are some new data escapes
4260 to vary what happens:
4261
4262 \C+ in addition, show current contents of captured substrings
4263 \C- do not supply a callout function
4264 \C!n return 1 when callout number n is reached
4265 \C!n!m return 1 when callout number n is reached for the mth time
4266
4267 30. If pcregrep was called with the -l option and just a single file name, it
4268 output "<stdin>" if a match was found, instead of the file name.
4269
4270 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4271 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4272 pcre_exec(). This saves a malloc/free per call. The default value of
4273 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4274 when configuring.
4275
4276 32. The default maximum size of a compiled pattern is 64K. There have been a
4277 few cases of people hitting this limit. The code now uses macros to handle the
4278 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4279 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4280 configuring. Tests 2 and 5 work only with 2-byte links because they output
4281 debugging information about compiled patterns.
4282
4283 33. Internal code re-arrangements:
4284
4285 (a) Moved the debugging function for printing out a compiled regex into
4286 its own source file (printint.c) and used #include to pull it into
4287 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4288 separate copies.
4289
4290 (b) Defined the list of op-code names for debugging as a macro in
4291 internal.h so that it is next to the definition of the opcodes.
4292
4293 (c) Defined a table of op-code lengths for simpler skipping along compiled
4294 code. This is again a macro in internal.h so that it is next to the
4295 definition of the opcodes.
4296
4297 34. Added support for recursive calls to individual subpatterns, along the
4298 lines of Robin Houston's patch (but implemented somewhat differently).
4299
4300 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4301 allow it to read and process whole directories in Win32. This code was
4302 contributed by Lionel Fourquaux; it has not been tested by me.
4303
4304 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4305 used to name a group. Names consist of alphanumerics and underscores, and must
4306 be unique. Back references use the syntax (?P=name) and recursive calls use
4307 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4308 numbers. The function pcre_fullinfo() can be used after compilation to extract
4309 a name/number map. There are three relevant calls:
4310
4311 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4312 PCRE_INFO_NAMECOUNT yields the number of entries
4313 PCRE_INFO_NAMETABLE yields a pointer to the map.
4314
4315 The map is a vector of fixed-size entries. The size of each entry depends on
4316 the length of the longest name used. The first two bytes of each entry are the
4317 group number, most significant byte first. There follows the corresponding
4318 name, zero terminated. The names are in alphabetical order.
4319
4320 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4321 case instead of 255. Making it the same both with and without UTF-8 support
4322 means that the same test output works with both.
4323
4324 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4325 calling malloc() with a zero argument.
4326
4327 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4328 optimization. I've improved things by keeping a bitmap of backreferences with
4329 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4330 fact referenced, the optimization can be applied. It is unlikely that a
4331 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4332 the match to follow \n) will appear inside brackets with a number greater than
4333 31, but if it does, any back reference > 31 suppresses the optimization.
4334
4335 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4336 of disabling numbered capturing parentheses. Any opening parenthesis that is
4337 not followed by ? behaves as if it were followed by ?: but named parentheses
4338 can still be used for capturing (and they will acquire numbers in the usual
4339 way).
4340
4341 41. Redesigned the return codes from the match() function into yes/no/error so
4342 that errors can be passed back from deep inside the nested calls. A malloc
4343 failure while inside a recursive subpattern call now causes the
4344 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4345
4346 42. It is now possible to set a limit on the number of times the match()
4347 function is called in a call to pcre_exec(). This facility makes it possible to
4348 limit the amount of recursion and backtracking, though not in a directly
4349 obvious way, because the match() function is used in a number of different
4350 circumstances. The count starts from zero for each position in the subject
4351 string (for non-anchored patterns). The default limit is, for compatibility, a
4352 large number, namely 10 000 000. You can change this in two ways:
4353
4354 (a) When configuring PCRE before making, you can use --with-match-limit=n
4355 to set a default value for the compiled library.
4356
4357 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4358 a different value is set. See 45 below.
4359
4360 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4361
4362 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4363 of things that can be changed at compile time. The first argument specifies
4364 what is wanted and the second points to where the information is to be placed.
4365 The current list of available information is:
4366
4367 PCRE_CONFIG_UTF8
4368
4369 The output is an integer that is set to one if UTF-8 support is available;
4370 otherwise it is set to zero.
4371
4372 PCRE_CONFIG_NEWLINE
4373
4374 The output is an integer that it set to the value of the code that is used for
4375 newline. It is either LF (10) or CR (13).
4376
4377 PCRE_CONFIG_LINK_SIZE
4378
4379 The output is an integer that contains the number of bytes used for internal
4380 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4381
4382 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4383
4384 The output is an integer that contains the threshold above which the POSIX
4385 interface uses malloc() for output vectors. See item 31 above.
4386
4387 PCRE_CONFIG_MATCH_LIMIT
4388
4389 The output is an unsigned integer that contains the default limit of the number
4390 of match() calls in a pcre_exec() execution. See 42 above.
4391
4392 44. pcretest has been upgraded by the addition of the -C option. This causes it
4393 to extract all the available output from the new pcre_config() function, and to
4394 output it. The program then exits immediately.
4395
4396 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4397 order to support additional features. One way would have been to define
4398 pcre_exec2() (for example) with extra arguments, but this would not have been
4399 extensible, and would also have required all calls to the original function to
4400 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4401 is used for passing in "extra" data from pcre_study().
4402
4403 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4404 contains the following fields:
4405
4406 flags a bitmap indicating which of the following fields are set
4407 study_data opaque data from pcre_study()
4408 match_limit a way of specifying a limit on match() calls for a specific
4409 call to pcre_exec()
4410 callout_data data for callouts (see 49 below)
4411
4412 The flag bits are also defined in pcre.h, and are
4413
4414 PCRE_EXTRA_STUDY_DATA
4415 PCRE_EXTRA_MATCH_LIMIT
4416 PCRE_EXTRA_CALLOUT_DATA
4417
4418 The pcre_study() function now returns one of these new pcre_extra blocks, with
4419 the actual study data pointed to by the study_data field, and the
4420 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4421 before. That is, this change is entirely upwards-compatible and requires no
4422 change to existing code.
4423
4424 If you want to pass in additional data to pcre_exec(), you can either place it
4425 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4426 block.
4427
4428 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4429 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4430 times with different match limits, until it finds the minimum value needed for
4431 pcre_exec() to complete. The value is then output. This can be instructive; for
4432 most simple matches the number is quite small, but for pathological cases it
4433 gets very large very quickly.
4434
4435 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4436 returns the size of the data block pointed to by the study_data field in a
4437 pcre_extra block, that is, the value that was passed as the argument to
4438 pcre_malloc() when PCRE was getting memory in which to place the information
4439 created by pcre_study(). The fourth argument should point to a size_t variable.
4440 pcretest has been extended so that this information is shown after a successful
4441 pcre_study() call when information about the compiled regex is being displayed.
4442
4443 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4444 because what follows is always an absolute path. (Later: it turns out that this
4445 is more than cosmetic for MinGW, because it doesn't like empty path
4446 components.)
4447
4448 49. Some changes have been made to the callout feature (see 28 above):
4449
4450 (i) A callout function now has three choices for what it returns:
4451
4452 0 => success, carry on matching
4453 > 0 => failure at this point, but backtrack if possible
4454 < 0 => serious error, return this value from pcre_exec()
4455
4456 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4457 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4458 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4459 use by callout functions. It will never be used by PCRE itself.
4460
4461 (ii) The pcre_extra structure (see 45 above) has a void * field called
4462 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4463 pcre_callout_block structure has a field of the same name. The contents of
4464 the field passed in the pcre_extra structure are passed to the callout
4465 function in the corresponding field in the callout block. This makes it
4466 easier to use the same callout-containing regex from multiple threads. For
4467 testing, the pcretest program has a new data escape
4468
4469 \C*n pass the number n (may be negative) as callout_data
4470
4471 If the callout function in pcretest receives a non-zero value as
4472 callout_data, it returns that value.
4473
4474 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4475 there were some redundant $(CFLAGS) in commands that are now specified as
4476 $(LINK), which already includes $(CFLAGS).
4477
4478 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4479 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4480 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4481 one-byte characters throughout. Note that case-insensitive matching applies
4482 only to characters whose values are less than 256. PCRE doesn't support the
4483 notion of cases for higher-valued characters.
4484
4485 (i) A character class whose characters are all within 0-255 is handled as
4486 a bit map, and the map is inverted for negative classes. Previously, a
4487 character > 255 always failed to match such a class; however it should
4488 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4489
4490 (ii) A negated character class with a single character < 255 is coded as
4491 "not this character" (OP_NOT). This wasn't working properly when the test
4492 character was multibyte, either singly or repeated.
4493
4494 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4495 mode, for example: \x{100}{2,3}.
4496
4497 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4498 singly or repeated) now correctly test multibyte characters. However,
4499 PCRE doesn't recognize any characters with values greater than 255 as
4500 digits, spaces, or word characters. Such characters always match \D, \S,
4501 and \W, and never match \d, \s, or \w.
4502
4503 (v) Classes may now contain characters and character ranges with values
4504 greater than 255. For example: [ab\x{100}-\x{400}].
4505
4506 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4507 PCRE in UTF-8 mode.
4508
4509 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4510 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4511 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4512 value.)
4513
4514 53. The single man page has become too large. I have therefore split it up into
4515 a number of separate man pages. These also give rise to individual HTML pages;
4516 these are now put in a separate directory, and there is an index.html page that
4517 lists them all. Some hyperlinking between the pages has been installed.
4518
4519 54. Added convenience functions for handling named capturing parentheses.
4520
4521 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4522 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4523 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4524 are faulted.
4525
4526 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4527 calling configure. These values are used when compiling the dftables.c program
4528 which is run to generate the source of the default character tables. They
4529 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4530 you will need to set these values.
4531
4532 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4533
4534
4535 Version 3.9 02-Jan-02
4536 ---------------------
4537
4538 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4539
4540 2. If --disable-static was given, the building process failed when trying to
4541 build pcretest and pcregrep. (For some reason it was using libtool to compile
4542 them, which is not right, as they aren't part of the library.)
4543
4544
4545 Version 3.8 18-Dec-01
4546 ---------------------
4547
4548 1. The experimental UTF-8 code was completely screwed up. It was packing the
4549 bytes in the wrong order. How dumb can you get?
4550
4551
4552 Version 3.7 29-Oct-01
4553 ---------------------
4554
4555 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4556 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4557 this didn't happen under Solaris 8, where I normally test things.
4558
4559 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4560 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4561 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4562 this fix an hour or so after the initial 3.7 release.)
4563
4564
4565 Version 3.6 23-Oct-01
4566 ---------------------
4567
4568 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4569 offsets passed as NULL with zero offset count.
4570
4571 2. The config.guess and config.sub files had not been updated when I moved to
4572 the latest autoconf.
4573
4574
4575 Version 3.5 15-Aug-01
4576 ---------------------
4577
4578 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4579 had been forgotten.
4580
4581 2. By using declared but undefined structures, we can avoid using "void"
4582 definitions in pcre.h while keeping the internal definitions of the structures
4583 private.
4584
4585 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4586 user point of view, this means that both static and shared libraries are built
4587 by default, but this can be individually controlled. More of the work of
4588 handling this static/shared cases is now inside libtool instead of PCRE's make
4589 file.
4590
4591 4. The pcretest utility is now installed along with pcregrep because it is
4592 useful for users (to test regexs) and by doing this, it automatically gets
4593 relinked by libtool. The documentation has been turned into a man page, so
4594 there are now .1, .txt, and .html versions in /doc.
4595
4596 5. Upgrades to pcregrep:
4597 (i) Added long-form option names like gnu grep.
4598 (ii) Added --help to list all options with an explanatory phrase.
4599 (iii) Added -r, --recursive to recurse into sub-directories.
4600 (iv) Added -f, --file to read patterns from a file.
4601
4602 6. pcre_exec() was referring to its "code" argument before testing that
4603 argument for NULL (and giving an error if it was NULL).
4604
4605 7. Upgraded Makefile.in to allow for compiling in a different directory from
4606 the source directory.
4607
4608 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4609 options bits, the pointer it was passed was to an int instead of to an unsigned
4610 long int. This mattered only on 64-bit systems.
4611
4612 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4613 generated) instead of pcre.in, which it its source. Also made the same change
4614 in several of the .c files.
4615
4616 10. A new release of gcc defines printf() as a macro, which broke pcretest
4617 because it had an ifdef in the middle of a string argument for printf(). Fixed
4618 by using separate calls to printf().
4619
4620 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4621 script, to force use of CR or LF instead of \n in the source. On non-Unix
4622 systems, the value can be set in config.h.
4623
4624 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4625 absolute limit. Changed the text of the error message to make this clear, and
4626 likewise updated the man page.
4627
4628 13. The limit of 99 on the number of capturing subpatterns has been removed.
4629 The new limit is 65535, which I hope will not be a "real" limit.
4630
4631
4632 Version 3.4 22-Aug-00
4633 ---------------------
4634
4635 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4636
4637 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4638
4639
4640 Version 3.3 01-Aug-00
4641 ---------------------
4642
4643 1. If an octal character was given, but the value was greater than \377, it
4644 was not getting masked to the least significant bits, as documented. This could
4645 lead to crashes in some systems.
4646
4647 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4648 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4649
4650 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4651 These just pass their arguments on to (pcre_free)(), but they are provided
4652 because some uses of PCRE bind it to non-C systems that can call its functions,
4653 but cannot call free() or pcre_free() directly.
4654
4655 4. Add "make test" as a synonym for "make check". Corrected some comments in
4656 the Makefile.
4657
4658 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4659 Makefile.
4660
4661 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4662 command called pgrep for grepping around the active processes.
4663
4664 7. Added the beginnings of support for UTF-8 character strings.
4665
4666 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4667 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4668 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4669 out for the ar command.)
4670
4671
4672 Version 3.2 12-May-00
4673 ---------------------
4674
4675 This is purely a bug fixing release.
4676
4677 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4678 of ZA. This was just one example of several cases that could provoke this bug,
4679 which was introduced by change 9 of version 2.00. The code for breaking
4680 infinite loops after an iteration that matches an empty string was't working
4681 correctly.
4682
4683 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4684 when matched against abbab (for example). After matching an empty string, it
4685 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4686 caused it to match further down the string than it should.
4687
4688 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4689 was there because it doesn't seem to be needed, and it causes trouble on some
4690 systems, as it is not a Standard C header. It has been removed.
4691
4692 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4693 were reported on the Macintosh. The changes were from
4694
4695 while ((c = *(++ptr)) != 0 && c != '\n');
4696 to
4697 while ((c = *(++ptr)) != 0 && c != '\n') ;
4698
4699 Totally extraordinary, but if that's what it takes...
4700
4701 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4702 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4703 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4704 assumes the way PCRE uses memmove() (always moving upwards).
4705
4706 6. PCRE is being used in one environment where strchr() is not available. There
4707 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4708 faster code anyway.
4709
4710
4711 Version 3.1 09-Feb-00
4712 ---------------------
4713
4714 The only change in this release is the fixing of some bugs in Makefile.in for
4715 the "install" target:
4716
4717 (1) It was failing to install pcreposix.h.
4718
4719 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4720
4721
4722 Version 3.0 01-Feb-00
4723 ---------------------
4724
4725 1. Add support for the /+ modifier to perltest (to output $` like it does in
4726 pcretest).
4727
4728 2. Add support for the /g modifier to perltest.
4729
4730 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4731 matches null strings.
4732
4733 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4734 pattern. Perl treats empty patterns specially - it reuses the most recent
4735 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4736 effect.
4737
4738 5. The POSIX interface was broken in that it was just handing over the POSIX
4739 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4740 required a bigger vector, with some working space on the end. This means that
4741 the POSIX wrapper now has to get and free some memory, and copy the results.
4742
4743 6. Added some simple autoconf support, placing the test data and the
4744 documentation in separate directories, re-organizing some of the
4745 information files, and making it build pcre-config (a GNU standard). Also added
4746 libtool support for building PCRE as a shared library, which is now the
4747 default.
4748
4749 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4750 09 are not valid octal constants. Single digits will be used for minor values
4751 less than 10.
4752
4753 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4754 existing programs that set these in the POSIX interface can use PCRE without
4755 modification.
4756
4757 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4758 return all that pcre_info() returns, plus additional data. The pcre_info()
4759 function is retained for compatibility, but is considered to be obsolete.
4760
4761 10. Added experimental recursion feature (?R) to handle one common case that
4762 Perl 5.6 will be able to do with (?p{...}).
4763
4764 11. Added support for POSIX character classes like [:alpha:], which Perl is
4765 adopting.
4766
4767
4768 Version 2.08 31-Aug-99
4769 ----------------------
4770
4771 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4772 trying to match at the startoffset position, but instead was moving forward to
4773 the next newline as if a previous match had failed.
4774
4775 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4776 and could get into a loop if a null string was matched other than at the start
4777 of the subject.
4778
4779 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4780 be distinguished at compile time, and for completeness also added PCRE_DATE.
4781
4782 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4783 in GnuWin32 environments.
4784
4785
4786 Version 2.07 29-Jul-99
4787 ----------------------
4788
4789 1. The documentation is now supplied in plain text form and HTML as well as in
4790 the form of man page sources.
4791
4792 2. C++ compilers don't like assigning (void *) values to other pointer types.
4793 In particular this affects malloc(). Although there is no problem in Standard
4794 C, I've put in casts to keep C++ compilers happy.
4795
4796 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4797 should be (const char *).
4798
4799 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4800 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4801 However, I haven't made this a standard facility. The documentation doesn't
4802 mention it, and the Makefile doesn't support it.
4803
4804 5. The Makefile now contains an "install" target, with editable destinations at
4805 the top of the file. The pcretest program is not installed.
4806
4807 6. pgrep -V now gives the PCRE version number and date.
4808
4809 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4810 causing the entire string to be ignored, instead of just the last character.
4811
4812 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4813 non-matching string, it can take a very, very long time, even for strings of
4814 quite modest length, because of the nested recursion. PCRE now does better in
4815 some of these cases. It does this by remembering the last required literal
4816 character in the pattern, and pre-searching the subject to ensure it is present
4817 before running the real match. In other words, it applies a heuristic to detect
4818 some types of certain failure quickly, and in the above example, if presented
4819 with a string that has no trailing " it gives "no match" very quickly.
4820
4821 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4822 other alternatives are tried instead.
4823
4824
4825 Version 2.06 09-Jun-99
4826 ----------------------
4827
4828 1. Change pcretest's output for amount of store used to show just the code
4829 space, because the remainder (the data block) varies in size between 32-bit and
4830 64-bit systems.
4831
4832 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4833 start matching at. This allows lookbehinds to work when searching for multiple
4834 occurrences in a string.
4835
4836 3. Added additional options to pcretest for testing multiple occurrences:
4837
4838 /+ outputs the rest of the string that follows a match
4839 /g loops for multiple occurrences, using the new startoffset argument
4840 /G loops for multiple occurrences by passing an incremented pointer
4841
4842 4. PCRE wasn't doing the "first character" optimization for patterns starting
4843 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4844 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4845 the letter 'x'. On long subject strings, this gives a significant speed-up.
4846
4847
4848 Version 2.05 21-Apr-99
4849 ----------------------
4850
4851 1. Changed the type of magic_number from int to long int so that it works
4852 properly on 16-bit systems.
4853
4854 2. Fixed a bug which caused patterns starting with .* not to work correctly
4855 when the subject string contained newline characters. PCRE was assuming
4856 anchoring for such patterns in all cases, which is not correct because .* will
4857 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4858 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4859 must be retried after every newline in the subject.
4860
4861
4862 Version 2.04 18-Feb-99
4863 ----------------------
4864
4865 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4866 computation of the store needed to hold the pattern was incorrect (too large).
4867 If such patterns were nested a few deep, this could multiply and become a real
4868 problem.
4869
4870 2. Added /M option to pcretest to show the memory requirement of a specific
4871 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4872
4873 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4874 compiled in such a way that the backtracking after subsequent failure was
4875 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4876 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4877
4878
4879 Version 2.03 02-Feb-99
4880 ----------------------
4881
4882 1. Fixed typo and small mistake in man page.
4883
4884 2. Added 4th condition (GPL supersedes if conflict) and created separate
4885 LICENCE file containing the conditions.
4886
4887 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4888 Perl, that is the internal \ allows the delimiter to be included in the
4889 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4890 the final delimiter, add \ to the end of the pattern (to test the error).
4891
4892 4. Added the convenience functions for extracting substrings after a successful
4893 match. Updated pcretest to make it able to test these functions.
4894
4895
4896 Version 2.02 14-Jan-99
4897 ----------------------
4898
4899 1. Initialized the working variables associated with each extraction so that
4900 their saving and restoring doesn't refer to uninitialized store.
4901
4902 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4903 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4904 fix the problem.
4905
4906 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4907 calls, and wasn't printing the correct value for compiling calls. Increased the
4908 default value of LOOPREPEAT, and the number of significant figures in the
4909 times.
4910
4911 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4912
4913 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4914 a building problem on Windows NT with a FAT file system.
4915
4916
4917 Version 2.01 21-Oct-98
4918 ----------------------
4919
4920 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4921 to character tables built by pcre_maketables() in the current locale. If NULL
4922 is passed, the default tables are used.
4923
4924
4925 Version 2.00 24-Sep-98
4926 ----------------------
4927
4928 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4929 it any more.
4930
4931 2. Allow quantification of (?>) groups, and make it work correctly.
4932
4933 3. The first character computation wasn't working for (?>) groups.
4934
4935 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4936 end of the subject) and add 5.005's \z, which really does match only at the
4937 very end of the subject.
4938
4939 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4940
4941 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4942 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4943 localized options. All options to pcre_study() were also removed.
4944
4945 7. Add other new features from 5.005:
4946
4947 $(?<= positive lookbehind
4948 $(?<! negative lookbehind
4949 (?imsx-imsx) added the unsetting capability
4950 such a setting is global if at outer level; local otherwise
4951 (?imsx-imsx:) non-capturing groups with option setting
4952 (?(cond)re|re) conditional pattern matching
4953
4954 A backreference to itself in a repeated group matches the previous
4955 captured string.
4956
4957 8. General tidying up of studying (both automatic and via "study")
4958 consequential on the addition of new assertions.
4959
4960 9. As in 5.005, unlimited repeated groups that could match an empty substring
4961 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4962 runtime if any iteration does actually match an empty substring.
4963
4964 10. Include the RunTest script in the distribution.
4965
4966 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4967 discrepancies, some of which were old and were also with respect to 5.004. They
4968 have now been fixed.
4969
4970
4971 Version 1.09 28-Apr-98
4972 ----------------------
4973
4974 1. A negated single character class followed by a quantifier with a minimum
4975 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4976 program crashes, or just wrong answers. This did not apply to negated classes
4977 containing more than one character, or to minima other than one.
4978
4979
4980 Version 1.08 27-Mar-98
4981 ----------------------
4982
4983 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4984
4985 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4986 latter must appear before anything that relies on it in the pattern.
4987
4988
4989 Version 1.07 16-Feb-98
4990 ----------------------
4991
4992 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4993 repeat of a potentially empty string).
4994
4995
4996 Version 1.06 23-Jan-98
4997 ----------------------
4998
4999 1. Added Markus Oberhumer's little patches for C++.
5000
5001 2. Literal strings longer than 255 characters were broken.
5002
5003
5004 Version 1.05 23-Dec-97
5005 ----------------------
5006
5007 1. Negated character classes containing more than one character were failing if
5008 PCRE_CASELESS was set at run time.
5009
5010
5011 Version 1.04 19-Dec-97
5012 ----------------------
5013
5014 1. Corrected the man page, where some "const" qualifiers had been omitted.
5015
5016 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5017 input syntax.
5018
5019 3. Fixed memory leak which occurred when a regex with back references was
5020 matched with an offsets vector that wasn't big enough. The temporary memory
5021 that is used in this case wasn't being freed if the match failed.
5022
5023 4. Tidied pcretest to ensure it frees memory that it gets.
5024
5025 5. Temporary memory was being obtained in the case where the passed offsets
5026 vector was exactly big enough.
5027
5028 6. Corrected definition of offsetof() from change 5 below.
5029
5030 7. I had screwed up change 6 below and broken the rules for the use of
5031 setjmp(). Now fixed.
5032
5033
5034 Version 1.03 18-Dec-97
5035 ----------------------
5036
5037 1. A erroneous regex with a missing opening parenthesis was correctly
5038 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5039 on some systems.
5040
5041 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5042 it was reported that one broken compiler failed on the former because "code" is
5043 also an independent variable.
5044
5045 3. The erroneous regex a[]b caused an array overrun reference.
5046
5047 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5048 fail on data ending with that character. (It was going on too far, and checking
5049 the next character, typically a binary zero.) This was specific to the
5050 optimized code for single-character negative classes.
5051
5052 5. Added a contributed patch from the TIN world which does the following:
5053
5054 + Add an undef for memmove, in case the the system defines a macro for it.
5055
5056 + Add a definition of offsetof(), in case there isn't one. (I don't know
5057 the reason behind this - offsetof() is part of the ANSI standard - but
5058 it does no harm).
5059
5060 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5061 most of the places where whitespace preceded '#'. I have given up and
5062 allowed the remaining 2 cases to be at the margin.
5063
5064 + Rename some variables in pcre to eliminate shadowing. This seems very
5065 pedantic, but does no harm, of course.
5066
5067 6. Moved the call to setjmp() into its own function, to get rid of warnings
5068 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5069
5070 7. Constructs such as \d{8,} were compiling into the equivalent of
5071 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5072 outcome, but in this particular case used more store than had been allocated,
5073 which caused the bug to be discovered because it threw up an internal error.
5074
5075 8. The debugging code in both pcre and pcretest for outputting the compiled
5076 form of a regex was going wrong in the case of back references followed by
5077 curly-bracketed repeats.
5078
5079
5080 Version 1.02 12-Dec-97
5081 ----------------------
5082
5083 1. Typos in pcre.3 and comments in the source fixed.
5084
5085 2. Applied a contributed patch to get rid of places where it used to remove
5086 'const' from variables, and fixed some signed/unsigned and uninitialized
5087 variable warnings.
5088
5089 3. Added the "runtest" target to Makefile.
5090
5091 4. Set default compiler flag to -O2 rather than just -O.
5092
5093
5094 Version 1.01 19-Nov-97
5095 ----------------------
5096
5097 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5098 like /([ab]*)*/, that is, for classes with more than one character in them.
5099
5100 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5101 as /((?>a*))*/ (a PCRE_EXTRA facility).
5102
5103
5104 Version 1.00 18-Nov-97
5105 ----------------------
5106
5107 1. Added compile-time macros to support systems such as SunOS4 which don't have
5108 memmove() or strerror() but have other things that can be used instead.
5109
5110 2. Arranged that "make clean" removes the executables.
5111
5112
5113 Version 0.99 27-Oct-97
5114 ----------------------
5115
5116 1. Fixed bug in code for optimizing classes with only one character. It was
5117 initializing a 32-byte map regardless, which could cause it to run off the end
5118 of the memory it had got.
5119
5120 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5121
5122
5123 Version 0.98 22-Oct-97
5124 ----------------------
5125
5126 1. Fixed bug in code for handling temporary memory usage when there are more
5127 back references than supplied space in the ovector. This could cause segfaults.
5128
5129
5130 Version 0.97 21-Oct-97
5131 ----------------------
5132
5133 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5134
5135 2. Optimized negated single characters not to use a bit map.
5136
5137 3. Brought error texts together as macro definitions; clarified some of them;
5138 fixed one that was wrong - it said "range out of order" when it meant "invalid
5139 escape sequence".
5140
5141 4. Changed some char * arguments to const char *.
5142
5143 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5144
5145 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5146 pcretest.
5147
5148
5149 Version 0.96 16-Oct-97
5150 ----------------------
5151
5152 1. Added a simple "pgrep" utility to the distribution.
5153
5154 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5155 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5156 where "ddd" means "one or more decimal digits".
5157
5158 3. Fixed serious bug. If a pattern had a back reference, but the call to
5159 pcre_exec() didn't supply a large enough ovector to record the related
5160 identifying subpattern, the match always failed. PCRE now remembers the number
5161 of the largest back reference, and gets some temporary memory in which to save
5162 the offsets during matching if necessary, in order to ensure that
5163 backreferences always work.
5164
5165 4. Increased the compatibility with Perl in a number of ways:
5166
5167 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5168 to request this handling. The option can be set at compile or exec time.
5169
5170 (b) $ matches before a terminating newline by default; an option
5171 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5172 mode). The option can be set at compile or exec time.
5173
5174 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5175 the same as Perl's. If the decimal number it represents is less than 10
5176 or there aren't that many previous left capturing parentheses, an octal
5177 escape is read. Inside a character class, it's always an octal escape,
5178 even if it is a single digit.
5179
5180 (d) An escaped but undefined alphabetic character is taken as a literal,
5181 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5182 escapes.
5183
5184 (e) {0} is now permitted. (The previous item is removed from the compiled
5185 pattern).
5186
5187 5. Changed all the names of code files so that the basic parts are no longer
5188 than 10 characters, and abolished the teeny "globals.c" file.
5189
5190 6. Changed the handling of character classes; they are now done with a 32-byte
5191 bit map always.
5192
5193 7. Added the -d and /D options to pcretest to make it possible to look at the
5194 internals of compilation without having to recompile pcre.
5195
5196
5197 Version 0.95 23-Sep-97
5198 ----------------------
5199
5200 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5201 \x20 at the start of a run of normal characters. These were being treated as
5202 real characters, instead of the source characters being re-checked.
5203
5204
5205 Version 0.94 18-Sep-97
5206 ----------------------
5207
5208 1. The functions are now thread-safe, with the caveat that the global variables
5209 containing pointers to malloc() and free() or alternative functions are the
5210 same for all threads.
5211
5212 2. Get pcre_study() to generate a bitmap of initial characters for non-
5213 anchored patterns when this is possible, and use it if passed to pcre_exec().
5214
5215
5216 Version 0.93 15-Sep-97
5217 ----------------------
5218
5219 1. /(b)|(:+)/ was computing an incorrect first character.
5220
5221 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5222 but not actually doing anything yet.
5223
5224 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5225 as Perl does (e.g. [-az] or [az-]).
5226
5227 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5228 all possible positions.
5229
5230 5. Split up into different modules to avoid including unneeded functions in a
5231 compiled binary. However, compile and exec are still in one module. The "study"
5232 function is split off.
5233
5234 6. The character tables are now in a separate module whose source is generated
5235 by an auxiliary program - but can then be edited by hand if required. There are
5236 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5237 toupper() in the code.
5238
5239 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5240 make them global. Abolish the function for setting them, as the caller can now
5241 set them directly.
5242
5243
5244 Version 0.92 11-Sep-97
5245 ----------------------
5246
5247 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5248 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5249
5250 2. Caseless matching was not working in character classes if the characters in
5251 the pattern were in upper case.
5252
5253 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5254
5255 4. Make PCRE_ANCHORED public and accept as a compile option.
5256
5257 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5258 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5259 pass them.
5260
5261 6. Give an error if bad option bits passed at compile or run time.
5262
5263 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5264 pcretest to cause it to pass that flag.
5265
5266 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5267 options, and the first character, if set.
5268
5269 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5270
5271
5272 Version 0.91 10-Sep-97
5273 ----------------------
5274
5275 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5276 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5277
5278 2. PCRE was looping on encountering an indefinitely repeated back reference to
5279 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5280 Perl does - treats the match as successful.
5281
5282 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5